Metrology of etched quartz and chrome embedded phase shift gratings using scatterometry

Scatterometry is presented as an optical metrology technique potentially capable of determining the critical parameters of phase shifting diffraction grating test structures (sidewall profile, linewidth, and etch depth where appropriate). The technique is noncontact, rapid and nondestructive. The test grating structure is illuminated by a laser beam and the intensities in the different diffracted orders are measured as the angle of incidence of the laser beam is varied over a certain range. Two different phase shifting photomasks were measured. The first mask consists of an array of chrome and chromeless phase etched gratings, fabricated at AT&T Bell Labs using c-beam techniques. The grating linewidths on this mask varied from nominal 0.5 ,am to 5.0 pm, while the etch depths varied from a nominal 190 urn to 400 nm depths. Both the chrome and the quartz gratings were measured. Linewidth and etch depth data obtained using scatterometry for the quartz gratings is presented and compared with AFM measurements of the same gratings. Each grating was measured using a Digital Instruments AFM located at AT&T. The absolute difference between the scatterometer and AFM measurements are calculated. The second photomask is an attenuating phase shift mask fabricated by DuPont Photomasks, which utilizes an attenuating chrome film stack to produce the desired phase shift. Therefore, this photomask is not phase etched; rather the phase shift is obtained through the chIome based absorber layer. Light transmitted through the chrome lines undergoes a 1800 phase shift at the exposure wavelength, relative to the light transmitted through the adjacent quartz spaces. The linewidth and line profile of the chrome grating was determined using scatterometry. The measurements of the diffracted orders were made using the 26 scatterometer located at the University of New Mexico. The shape of the diffraction curves obtained in this manner has been shown to be sensitive to the grating structure parameters (sidewall profile, etch depth, linewidth, etc.).1'2'3 An estimate of the quartz phase etched structure parameters and the attenuating O-8194-1787-4/95/$6.OO SPIE Vol. 2439 / 479 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 01/01/2013 Terms of Use: http://spiedl.org/terms chrome grating parameters was obtained through a combination of rigorous coupled wave theory (RCWT) and minimum mean square error (MMSE) ana1ysis Finally, the long term and short term repeatabilities of the scatterometer measurements are shown to be excellent. 2.0 BACKGROUND The lithographic process is the technical driver of the semiconductor industry. In a presentation given by K. Brown of SEMATECH at the 1994 BACUS conference, she comments that "Mask technology will continue to be a key enabler of the total lithographic system... independent of other technology alternatives". With so much emphasis being placed on mask technology, there is also additional emphasis being placed on the development of inspection tools for the photomask industry. One active area of research is in the development of inspection tools for characterizing photomasks. The theory of PSMs has been covered widely in the literature.6 Basically, phase shifting photomasks utilize interference of adjacent diffracted fields from closely spaced apertures which are optically 1800 out of phase. The phase shift of 1800 can be obtained by either depositing onto one aperture a transmissive layer of thickness d, given by, d= , 1 2(n—1) or etching one of the apertures to a depth, d. The wavelength, , is that used during wafer exposure and n is the index of the transmissive shifter material. The performance of the phase shifted area depends upon the actual etch depth (or thickness), the lmewidth, the sidewall angles, the line profile, and the alignment of the shifter to the apertures (in the case of a deposited layer). Photomask metrology tools must therefore address the measurement of these critical parameters. 3.0 SAMPLE DESCRIPTION 3.1 AT&T Photomask The first phase shifting photomask used in this work was fabricated by AT&T Bell Labs. The photomask is a standard Hoya five-inch square, 90 mils thick photomask with a chrome oxide on chrome film stack deposited on a quartz substrate. The photomask layout consists of a matrix of 48 e-beam written gratings. Figure 1 is a schematic of the mask layout as described below. There are four columns of different linewidths, with nominal 50/50 line to space ratios, which are repeated yielding eight columns in total. The smallest nominal linewidth was 0.5 pm, followed by a 1 .0 m lmewidth, a 2.0 pm linewidth, and finally a 5 .0 m linewidth. The six rows represent different levels of etch depth from a nominal 190 nm etch to a nominal 400 nm etch. At each row and column position, there are four 5mm by 5mm size gratings, two of which are etched chrome on etched quartz and two are simply etched quartz. Reactive ion etch (RIE) processing of the photomask helped to ensure vertical sidewall structure. Further, at each location one pair of quartz and chrome gratings were written with vertical lines while the other pair of gratings have horizontally written lines. The data and results presented in this paper are for the gratings with vertical lines exclusively. 3.2 DUPONT EMBEDDED PHASE SHIFT PHOTOMASK The second photomask is a chrome embedded phase shift photomask produced by DuPont Photomasks. The structure of interest on the photomask is a large area (2 cm by 2 cm) composed of a series of diffraction gratings (78 in total) each ofwhich is 2 cm wide by 150 pm high. The grating lines are 150 ,am long. The specific grating structure and material parameters are discussed fully in later sections. 4.0 THEORETICAL MODEL FOR SCATTEROMETRY The scatterometry technique for measuring grating parameters is a two fold process: first, measuring the scatter from the grating, and second, finding a grating model which fits the experimental data. The solution to the second part, finding the appropriate grating model, requires one to be able to compute the scatter from a hypothetical structure. The accuracy in modeling the physical structure with a hypothetical structure is contingent upon how well the grating material parameters are known at the probe wavelength of the scatterometer. In the case of a chrome grating, knowledge of the real and imaginary parts of the index of refraction and the thickness of the chrome oxide and chrome layers are essential for an accurate model. Details of how the material parameters for the chrome gratings are determined is addressed in a separate section. For the simpler case of an etched quartz (fused silica) grating, only the complex index of refraction of the quartz material is necessary and is easily obtained. In both cases, the scatter from the hypothetical structure is computed using rigorous coupled wave theory8'11 (RCWT), which is discussed