C. V. Kent and J. Lawson, J. Opt. Soc. Am. 27, 117 (1937).
W. Budde, Appl. Opt. 1, 201 (1962).
S. R. Rajagopalan and S. Ramaseshan, Proc. Indian Acad. Sci. A 60, 297 (1964).
S. R. Rajagopalan and S. Ramaseshan, Proc. Indian Acad. Sci. A 60, 379 (1964).
D. Clarke and J. F. Grainger, Polarized Light and Optical Measurement (Pergamon, New York, 1971).
J. I. Bohnert, Proc. IRE 39, 549 (1951).
B. D. Cahan and R. F. Spainer, Surf. Sci. 16, 166 (1969); also, in Proceedings of the Symposium on Recent Developments in Ellipsometry edited by N. M. Bashara, A. B. Buckman, and A. C. Hall (North-Holland, Amsterdam, 1969).
R. Greef, Rev. Sci. Instrum. 41, 532 (1970).
J. C. Suits, Rev. Sci. Instrum. 42, 19 (1971).
D. J. Scholtens, J. F. Kleibeuker, and J. Kommandeur, Rev. Sci. Instrum. 44, 153 (1973).
D. E. Aspnes, Opt. Commun. 8, 222 (1973).
P. S. Hauge and F. H. Dill, IBM J. Res. Dev. 17, 472 (1973).
D. E. Aspnes, J. Opt. Soc. Am. 64, 639 (1974).
D. E. Aspnes, J. Opt. Soc. Am. 64, 812 (1974).
R. M. A. Azzam and N. M. Bashara, J. Opt. Soc. Am. 62, 222 (1972).
D. A. Holmes and D. E. Feucht, J. Opt. Soc. Am. 57, 466 (1967).
R. M. A. Azzam and N. M. Bashara, J. Opt. Soc. Am. 62, 336 (1972).
R. M. A. Azzam and N. M. Bashara, J. Opt. Soc. Am. 61, 600 (1971).
W. R. Hunter, D. H. Eaton, and C. T. Sah, Surf. Sci. 20, 355 (1970).
R. M. A. Azzam and N. M. Bashara, J. Opt. Soc. Am. 62, 700 (1972). For the compensator, Tc0 is diagonal; the ratio of the 2,2 to the 1,1 matrix elements is equal to ρc [Eq. (11)]. For the entrance (and exit) window, Tw0 is the product of a complex constant times the 2×2 identity matrix.
R. M. A. Azzam and N. M. Bashara, J. Opt. Soc. Am. 61, 773 (1971); J. Opt. Soc. Am. 61, 1236 (1971).
Notice that, because of the inequality (28), the inequality (31) restricts Xr to lie inside, on, or not too far outside the unit circle |Xr| = 1 in the complex plane of polarization. In other words, the reflected polarization should be more p-like than s-like.
R. M. A. Azzam and N. M. Bashara, Appl. Phys. 1, 203 (1973); Appl. Phys. 2, 59 (1973). 94
Xr is controlled by the optical elements P and C of the polarizing arm of the ellipsometer, whereas XA is adjusted by rotating the analyzer A around the beam axis, Eq. (41).
See Eq. (5), Ref. 15.
Equations (46a) and (46b) can be simplified, without loss of generality, if we choose the reference position of the analyzer so that the major axis of the transmitted elliptical vibration XAO is parallel to the plane of incidence. Thus, in this case, XAO is pure imaginary, Re(xAO) = 0, and the second terms in the numerators of the right-hand sides of Eqs. (46a) and (46b) become zero.
See Fig. 4 and Table II of Ref. 15.
If an elliptical, instead of linear, analyzer is used, the two polarizations xr and xr* that differ only in handedness lead to different normalized Fourier coefficients α and β, hence can be distinguished. This can be seen from Eqs. (46a) and (46b) by noting that, for an elliptical analyzer, the second terms in the denominators of the right-hand sides of these equations are nonzero and switch sign as xr* is substituted instead of xr.
R. J. Archer and C. V. Shank, J. Opt. Soc. Am. 57, 191 (1967).
T. Yolken, R. Waxler, and J. Kruger, J. Opt. Soc. Am. 57, 283 (1967).
W. G. Oldham, J. Opt. Soc. Am. 57, 617 (1967).
F. L. McCrackin, J. Opt. Soc. Am. 60, 57 (1970).
J. A. Johnson and N. M. Bashara, J. Opt. Soc. Am. 60, 221 (1970).
D. E. Aspnes, J. Opt. Soc. Am. 61, 1077 (1971).
Equation (62) is the same as Eq. (52) except that, in the latter, xr is explicitly expressed in terms of the normalized Fourier coefficients α and β using Eq. (51).
When C= 0, Eqs. (62a) and (62b) become identical and equivalent to one equation, ρ/ρc= tanP/xr, and the subsequent discussion does not apply.
Alternatively, starting with Eq. (62b) and repeating steps similar to those that led to Eq. (63a), we can obtain a quadratic in ρonly.
Ways to achieve this are mentioned in Refs. 12 and 14.
This assumes that |ρ.c| = Tc is very close to unity, which is true for most compensators. From Eq. (51), note that Xr is real; hence it represents a linear vibration, if α2+β2= 1. This leads to an amplitude of the ac component of the photoelectric current equal to its dc component, as may be seen from Eq. (44) after the cosine and sine terms are combined into a single sine or cosine term. Thus, a linear state can be detected by a rotating analyzer from the condition of maximum (unity) modulation depth in the photoelectric current. See Ref. 14.
This may be chosen on the basis of optimum-precision considerations, as discussed in Refs. 12 and 13.
The azimuth of the compensator can be changed 90° electronically, e.g., by use of a KDP crystal mounted to rotate with the (quarter-wave) compensator as one unit, to which a half-wave voltage that can be regulated by a corrective electro-optic feedback loop is applied along the fast axis of the compensator.
The method of Kent and Lawson (Ref. 1) relies on producing a circular reflected state (detected by a rotating analyzer) by varying the azimuth of a linear polarizer in the incident beam and the angle of incidence. In the present discussion, we assume that the angle of incidence is fixed, but that a compensator is used in the incident beam, which, together with the polarizer, can be adjusted to make the reflected polarization circular.
R. M. A. Azzam and N. M. Bashara, J. Opt. Soc. Am. 62, 1375A (1972); J. Opt. Soc. Am. 62, 1521 (1972); J. Opt. Soc. Am. 64, 128 (1974).