WO1991014973A1

WO1991014973A1 - Mirrors

Info

Publication number: WO1991014973A1
Application number: PCT/GB1991/000440
Authority: WO
Inventors: Roy Jonathan Rosser
Original assignee: British Technology Group Ltd
Priority date: 1990-03-22
Filing date: 1991-03-22
Publication date: 1991-10-03
Also published as: GB9106138D0; EP0521083A1; GB2242995B; JPH05508909A; GB2242995A; GB9006493D0

Abstract

A mirror collimating arrangement for a beam of radiation comprises a plurality of saddle-toroid-shaped reflecting surfaces (1) wherein the dimensions of the major and minor radii (R, e) of said reflecting surfaces are selected so that the change in reflectivity with angle is substantially complementary to the inhomogeneity of the cross-sectional shape (12) of the beam.

Description

Mirrors This invention relates to mirrors and, in particular, to mirrors and optical arrangements for x-ray lithography using synchrotron sources. It embodies means for compensating for shape inhomogeneities when using a saddle toroid or saddle toroid array mirror for collimating and homogenising synchrotron radiation and may especially be applied to a method which utilises variation in grazing incidence reflectivity with angle. X-ray proximity lithography 1s a technique for high-yield manufacture of integrated circuits. The principle of synchrotron based lithography is simple. By placing a patterned mask 1n close proximity (about 50μm) to a photoresist-coated silicon wafer and then Illuminating through the mask, with l.Onm radiation (or similar), the pattern is shadowed on to the photoresist. For lithography, the ideal beam would uniformly Illuminate a patch about 50mm by 50mm - 1 e. matching the size and shape of the patterned mask. Although synchrotron radiation has many advantages for this application, principally intensity and stability, 1t does have one major shortcoming. The divergence in the vertical plane 1s small (about lmR at lnm for a 750MeV machine), so that at 10m from the bending magnets (the geometric source of the radiation), the vertical extent of the beam Is only lOmrn, compared with a horizontal spread that can be of the order of 200mm. Moreover it has a gaussian distribution in the vertical direction. The non-uniform, horizontal slit of radiation produced by the synchrotron can be converted to a useful shape and uniformity 1n a number of ways. The most successful to date has been the use of a cylindrical mirror that detects and colli ates the beam is scanned up and down. This has two shortcomings. The scanning mechanism has to be in an ultra high vacuum section, close to the synchrotron, so that if mechanical failure occurs, a repair would take a minimum of 1 to 2 days. This is necessary to restore the ultra high vacuum. Secondly, the Illuminated area must be over-scanned to ensure uniformity. Demonstrated systems achieve only a duty cycle of 25% (i.e. 75% of the light is wasted) and designers only expect to achieve a 50% duty cycle.

Another method is to use a specially designed, saddle toroid array mirror (STA). Based on preliminary calculations that compared the STA mirror's expected performance to that of existing systems, it can increase the flux throughput by about 180%. Even 1f the duty cycle of the scanning systems is improved to the 50% that is usually regarded as the achievable limit, the stationary STA mirror would represent a flux enhancement of about 40%. The STA mirror has the added advantage that it requires no moving parts in the ultra high vacuum of the synchrotron.

The invention will be particularly described with reference to the accompanying drawings 1n which:-

Figure 1 is a schematic drawing of a basic saddle toroid, Figure 2 is the corresponding view of an array of saddle toroids, Figure 3 shows the cross-sectional shape or footprint of a beam at the lithography station after reflection by a saddle toroid, Figure 4 1s a schematic plan of an x-ray lithography beamline; and Figure 5 is a plot of the reflectivity of gold. To understand STA mirrors, it 1s convenient to look at a single saddle-toroid. This is a toroid with a concave minor radius p and a convex major radius R, shown in Figure 1. This can be thought of as a cylindrical mirror which achieves dispersion in the vertical direction by being bent rather than scanned. The bending in the major axis is small for synchrotron lithography, requiring a radius of the order of 50μm. Such optics could be formed by bending cylindrical mirrors.

The disadvantage of using a simple saddle-toroid is that it merely expands the gausslan-profiled synchrotron light. The resultant vertical nonuniformity can be overcome by using a saddle-toroid array (STA) mirror, shown in Figure 2. It is a large number (up to 1000) of micro saddle-torolds 1 all on the same substrate 2. Such devices can be made by standard techniques used in the microelectronics industry, as demonstrated by the production of arrays of micro lenses. The STA mirror works because each micro-toroid sees a small sector of incoming beam and spreads it over the entire lithography patch. The incoming beam is smoothed out using spatial averaging rather than the temporal averaging that occurs with a scanning mirror.

For lower homogeneity applications, the number of icrotorbids may be relatively low - as few as five - in which case the optics may be made by conventional grinding.

However, all these saddle toroid array mirrors, used as the only or principal collimating and homogenising element, suffer from the fact that the output patch-shaped like a truncated segment 12 of a pie, as illustrated in Figure 3. This results 1n an inhomogenelty in the vertical direction, even if the incoming beam's gaussian profile has been smoothed out. This Invention solves this shape inhomogenelty, by choosing a material and a set of conjugates, so that the change 1n reflectivity with angle compensates, to first order, the shape inhomogeneity.

According to the present invention there 1s provided a mirror collimating arrangement for a beam of radiation comprising a plurality of saddle-toroid-shaped reflecting surfaces wherein the dimensions of the major and minor radii of said reflecting surfaces are selected so that the change in reflectivity with angle is substantially complementary to the inhomogeneity of the cross-sectional shape of the beam.

The minor radius of the saddle toroid or base of the saddle toroid array is determined by geometrical optics and the need to colli ate the beam, using the formula:

p = 2d sin(α) 1 where p is the minor radius, d is the distance from the source point (a bending magnet) 14 to the centre of the mirror 10, and α is the grazing angle of incidence. The slope of the sides of the patch of illumination 12, shown in Figure 3, is determined by the minor radius and the required width of illumination at the lithography station. For an illumination width w, and a minor radius p, the angle θ (in radians) of the sloping sides is given, to first order, by:-

θ = w/p 2

The shape inhomogeneity, caused by the expansion at the bottom and the contraction at the top of the footprint 12 as shown in Figure 3, can be characterised by the maximum width, w(max) and the minimum width, w(min).

The ratio w(max)/w(min) is given, to first approximation by either of the related formulae:

w(max)/w(min) I + tan(θ/2) 3

or

w(max)/w(min) = 1 + w/(2p)

The crucial feature of this method is to now compensate for this shape compression, by choosing the appropriate angles of reflectivity. Figure 5 shows the reflectivity of gold at x-ray wavelengths for a number of different angles. Because of the nature of the saddle toroid, all the rays reflected to the upper, smaller, crescent of the lithography footprint 12 have undergone the maximum angle of reflection and all the rays reflected to the bottom, larger, crescent of the footprint 12 have undergone the minimum angle of reflection. The rays in between have all experienced an angle of reflection between these two extremes, proportional to their distance from the upper crescent. By choosing a material in which larger angles of grazing incidence have a lower coefficient of reflecti ity, and where the variation of reflectivity is roughly linear with angle over the angular range required, it is possible to chose angles so that the change in reflectivity compensates for the shape factor i.e.

r(max)/r(min) = w(max)/w(min) 5

where r(max) is the maximum reflectivity and r(min) is the minimum reflectivity required.

This required divergence, δ(req), given by

δ(req) = α(max) - α(min) + δ(inc) 6

where α(max) and α(min) are the maximum and minimum grazing angles of reflection, at which the reflectivity is r(min) and r(max) respectively and δ(inc) Is the divergence of the incident beam (I.e. the original vertical divergence of the synchrotron). The major radius, R, of a single saddle toroid 10 would be found, by first finding v, the second conjugate of the toroidal reflection.

v = m/δ(req)) 7

where δ(req) is measured in radians and m is the length of the saddle toroid. given by:

= dδ(inc)/(sin(α)) 8

where d is the source to mid-mirror distance and α is the angle of grazing incidence at the centre of the mirror 10.

Then, from the standard formula for the conjugates of an imagi ng system:

1 /u + 1 /v = 1 /f 9

5 In the case of this toroid, u = d and f = R sin(α)/2, so that the major radius R is given by:

R = 2dδ(inc)/{(δ(req)sinα + δ(inc))sinα} 10

10 For a saddle-toroid array, the major radius R(sta) is given by:

R(sta) = R/n 11

15 where n is the number of saddle toroids required

The distance, D, from the mid-mirror 10 to the lithography plane 16 can be determined using the formulae:

W(vert) - dδ(inc) = Dδ(req) 12

20 or D = W(vert)δ(req) 13

where W(vert) is the required spread of the synchrotron radiation in the vertical direction at the lithography station

25 16.

All these formulae would, in practice, be used to give a first approximation, which would then be refined by ray tracing for a final design.

Typical values for these parameters, calculated by these

30 formula are, assuming d, the distance from the source to the middle of the first mirror is 2700mm, that α, the angle of grazing incidence at the centre of the mirror, is 1.5° (26mR) and that the illuminated patch required w is 50mm wide, then:-

35 p, the minor radius of the mirror, is 141mm γ, the angle of the footprint, is 21° (approx). w(max)/w(min) , the ratio of geometric compression of the footprint, is 1.18 r(max) and r(min), the required maximum and minimum reflectivities are 75.6% and 64.5%, given respectively by the minimum and maximum angles of grazing incidence, α(min) and α(max), 1.3° and 1.8° (21 and 29mR) δ(req) is then 0.52° or 9mR m, the length of the mirror is 103mm R, the major radius for a single toroid, is 166955mm

R(sta), the major radius for the icro-toroids of a saddle-toroid array with 20 elements, is 8347 mm and D, the distance from the middle of the mirror to the lithography plane is 5255mm

Claims

1. A mirror collimating arrangement for a beam of radiation comprising a plurality of saddle-toroid-shaped reflecting surfaces (1) characterised in that the dimensions of the major and minor radii (p,R) of said reflecting surfaces are selected so that the change in reflectivity with angle is substantially complementary to the inhomogeneity of the cross-sectional shape of the beam.

2. A mirror collimating arrangement for a beam of radiation as claimed in claim 1 characterised in that the reflecting surfaces

(1) are of a material in which larger angles of grazing incidence exhibit a lower coefficient of reflectivity.

3. A mirror collimating arrangement for a beam of radiation as claimed in claim 2 character sed in that the variation of reflectivity with angle is approximately linear over the angular range required.

4. A mirror collimating arrangement for a beam of radiation as claimed in claim 2 characterised in that the material is gold.

5. A mirror collimating arrangement for a beam of radiation as claimed in any one of the preceding claims characterised in that the minor radius of the saddle toroid or base of the saddle toroid array is determined by the formula

p = 2d sin(α)

where p is the minor radius, d is the distance from a source point to the centre of the mirror 10, and α is the grazing angle of incidence.

6. A mirror collimating arrangement for a beam of radiation as claimed in any one of the preceding claims characterised in that the major radius R is given by

R = 2dδ(inc)/{(δ(req)sinα + δ(inc))sinα}

where d is the source to mid-mirror distance, α is the angle of grazing incidence at the centre of the mirror, δ(req⁾ is the required divergence of the beam and δ(inc) is the divergence of the incident beam.

7. Apparatus for lithography characterised in that 1t incorporates a mirror collimating arrangement (10) in accordance with any one of the preceding claims.