Gaussian beam parameter transformationat a refracting

P. BAUES

surface

The Fourier transform method is used to calculate the electric field of the astigmatic gaussian beam which appears when an incident stigmatic gaussian beam is refracted at the interface between media with different refractive indices. A beam parameter transformation law is derived which shows that the beam parameters of the refracted astigmatic beam can be calculated from the beam parameters of the incident stigmatic beam and the angles of incidence and refraction. An heuristic derivation of the transformation law based on geometrical optics is also given. The application of the law is illustrated by determining the plane parallel plate astigmatism.

Numerous optical devices contain optical components which change the direction of propagation of a beam of light by refraction. Such devices are, for example, electrooptic light deflectors l and optical resonators, which for Q-switching applications are provided with internal Brewster plates. In addition such resonators often consist of laser rods and modulators cut at Brewsters angle.2 The beam parameters of a gaussian beam which traverses a refracting surface undergo a transformation. Assuming an incident gaussian beam which is stigmatic, this transformation can be derived as a function of the angles of incidence and refraction. The incident gaussian beam is expanded into plane waves to which the laws of refraction and transmission are applied.3 The summation of the refracted partial plane waves leads approximately to a refracted gaussian beam with transformed beam parameters. A similar calculation was performed for normal incidence in reference 4. An heuristic derivation of the beam parameter transformation law based on geometrical optics is also given. As an application of the transformation law, the known astigmatism of a plane parallel plate 5*6 is determined. The assumed approximations do not hold in the vicinity of the total reflection angle. The reflection of a gaussian beam for this case has been examined in reference 7. The more general case of an incident gaussian beam with general astigmatism s can be treated along the lines described in this paper. This problem is omitted so that the basic ideas can be followed more easily. Plane wave expansion of a stigmatic gaussian beam The following considerations assume an incident gaussian beam which is stigmatic, propagates in the z direction, and has an electric field normal to the plane of incidence (Fig. 1). A treatment of the electric field component parallel to the plane of incidence is not necessary. All
The author is with Siemens AG, Research Laboratories, Germany. Received 15 December 1975. Munich, Fig.1

Medium 2 (Index of refractlon n2

Medium I (Index of refroctlon n,)

Refrochng surface

Refraction

of plane waves

results for the normal component can be applied analogously to the parallel component. The electric field of a paraxial gaussian beam is approximately normal to the direction of propagation. Under these assumptions the incident beam in medium 1 has an electric field given by:9 . Er,l(x)=Qexp
41

jcI(x2

ty2)

-jkrz

(1)

where a point in space is denoted by x = (x, y, z) and E,,, is a constant complex electric field vector amplitude which is normal to the direction of propagation, ie to the z direction, and to the plane of incidence. h is the wavelength of light in free space and kt = 2n n 1/A is the propagation constant in medium 1, which has the index of refraction

OPTICS AND LASER TECHNOLOGY.

APRIL

1976

85

IZr

The complex beam parameter: +(f+ k,)* ql i

% f (7)

11 is given by the spot size W, at the beam waist and the location zr of the beam waist on the z axis. The plane wave expansion of the electric field El,1 (x) of the gaussian beam is obtained by using the Fourier transform in three dimensions:

The vector f = (sin 0, , 0, cos 0 I) is a unit vector normal to refracting surface between mediums 1 and 2. The summation of the refracted partial plane waves yields the electric field Ez,, (x) of the refracted beam:

E2,1(~)

= s

F2,ltk2)

wb

j k2

- x1 dk,

(8)

K F, , , (k,
) = __i.___

(%I)3
K

(3)

where F,,, (k,) denotes the complex amplitude of a plant partial wave which propagates in medium I in the direction of the wave vector k, = (kr,,, k,,,,, kr,,). The length of the wave vector in medium 1 is Ik, I = k, = 2nA/n,. dx is the volume element in x space R. The scalar product of k, and x in the argument of the exponential function is denoted by a dot. Insertion of (1) in (3) results in:

This integral is the inverse Fourier transform and dkr = dk,, dkr, dkr, is the volume element in Fourier space A. Under assumptions which are now discussed the integral (8) can be solved approximately. Insertion of (6) in (8) results in :

E2,1@)

Tl(kl,oJ

F,,

L(kl

x F,,l (kl)=.i -- El,lw

4n

Y,,O

(k:., + k:,J
1

x exp[-j

k,(kr)

* x] dk,

(,I

k, +

2k,

- (k:

k:,J
1

(4)

6(z) is Diracs delta function and qr,o is the beam parameter at the location z = 0. ql,o is given from (2) by:

ql,o

=j

wi -

1
1

The gaussian beam (I) travels along the z axis, ie in the direction of the wave vector k, = kl,,O = (O,O, k,). The amplitudes F,,l(k,) of (4) have significant magnitudes only in the vicinity of kr = k,,o. Furthermore the transmission coefficient Tl(kr) is only a weak function of kr . Therefore the transmission coefficient TL(k,) has been placed, with value T1(k,, o), before the integral (9). After insertion of (4), this integral can be solved in the Fresnel approximation. The 6 function in (4) makes it possible to integrate (9) over the k,,, coordinate. Thus the argument:

b(kl) - x = k, @1,x,kl,,) * x
Refracted beam The gaussian beam (1) travels along the z axis. It is composed of the partial plane waves given by (4) refracted at the interface between mediums 1 and 2. A partial plane wave which propagates in medium 1 with amplitude F,,r in the direction of the wave vector kl propagates, after crossing the refracting surface, with amplitude Fz,r in the direction of the wave vector k, in medium 2 (Fig.1). are normal to the plane of incidence. The Fz,.L and FI,I amplitudes F,,, and F,,L are associated by:
F2.1 (k2)=7l(k,)F,,l(k,) (6)

of the exponential function in (9) is only a function of the This function is expanded up to coordinates kl,, and k,,,. the second-order terms in k,,, and k,,, around k,,, = 0 and k,,, = 0. The position vector x is measured in the coordinate system x, y, z (Fig.1); the positive z axis has the direction of k,(k,,,) with kl,o = (O,O, kl) in the x, ~1, z system. A plane wave travelling in medium 1 in the positive z direction propagates, after crossing the refracting surface, in the positive z direction in medium 2. The .v axis is normal to the plane of incidence and coincides with the y axis. The expansion of (10) is:

where TL (k,) is the amplitude transmission coefficient of an electric field component normal to the plane of incidence. T, (k,) and the wave vector k2 depend on the wave vector kl. The function k,(k,) can be obtained from Snells law of refraction and is: k2(k,)=kl -(f.k,)f

k2 - x = k2,o - x + a, kl,,
+ b,

+aykl,,
(I 1)

k:,, + &,

h-l,, k,,,, + b,. :.,,

The vector k,, o is parallel to the z axis,

- x = k,z

(12)

where ikz,ol = k2 = 2m2/h and n2 is the index ofrefraction of medium 2. The function (10) can be obtained from (7) and the condition kl,, = kl - (k: x + k: ,)/2 kl, which was imposed in (4), by integratjng eve; kl,, in (9).

86

OPTICS AND LASER TECHNOLOGY.

APRIL

1976

Differentiation of (10) with respect to kr,, second-order terms yields the coefficients: coser , (I,=----X cos 82
Uy

and kr,,

up to

The beam parameters of the refracted beam :

4x =

(4
c0se2 c0s e1

2
41,o +;

(15)

=y
-+p-k% -k: kl k2 sine, x

for the xz plane, and:

b, = -

(16)
for the yz plane are given by (14). In (15) and (16) ql,o is the beam parameter (5) of the incident stigmatic gaussian beam in medium 1 immediately before the refracting surface. The beam parameters qx, 0 and qy,o of the refracted astigmatic beam immediately behind the refracting surface result from (15) and (16) on inserting z = 0

cos3e22 k2

b,

=0

+ k;

~0s~;~

i2k: cos2tl 1 sin 0 r cos

X 2 k2 (13) 4x,o =

( >
cos e1
41,0

cos e2

41,o

U7a) U7b)

where kl,, = 0 and kl,, = 0 are used. The refracted partial plane waves in (9) can now be integrated taking account the expansion (11) and the coefficients in (13) to yield the electric field:

qy,o =

E2,1(4

The equations (17) constitute the law of transformation for the beam parameters of an incident stigmatic gaussian beam crossing a refracting surface. The beam parameter in the plane of incidence is transformed through (17a), whereas the beam parameter in the yz plane is, as can be seen from (17b), unchanged. Heuristic derivation of the beam parameter transformation law In geometric optics a beam refracted at the interface between mediums 1 and 2 (Fig.2) has a diameter dr in medium 1 and, parallel to the plane of incidence, a diameter d2 in medium 2. Fig.2 shows that: d2 = __
cos

x exp

c0s2e2 c0s2el

i;x2 , t

jh RY~
, -jk2z (14)

e,

of the refracted beam, which is an astigmatic gaussian beam. The result (14) is obtained by neglecting the second terms in the sums for b, and by in (13). This approximation is justified for paraxial beams, ie for small x, and for a sufficiently long distance z from the refracting surface. The approximation is better the smaller the angle of incidence 0r because the function sin 0 1 appears as a coefficient in the second term of the sums for b, and by. For total reflection the angle of refraction is e2 = 7r/2 and the expansion (11) breaks down because, since cos e2 = cos n/2 = 0, the coefficients a,, b, and by of (13) are no longer defined. This means that the result (14) for the refracted beam fails to be valid in the vicinity of the total reflection angle. Beam parameter transformation law

cOsel

d,

(18)

The beam diameter normal to the plane of incidence is not changed. The imaginary part of the beam parameter ql,o is proportional to the square of the spot size wr as can be seen from (5). The imaginary part of the beam parameter qx, o is similarly proportional to the square of the spot size w, . Thus the following heuristic substitutions can be performed:

d:+w:+q,,
d; +w;, -+qxl o law: (19)

Insertion of (19) in (1 S) yields the transformation

The stigmatic gaussian beam (1) generates the refracted beam (14) at the interface between mediums 1 and 2 (Fig.1). This generated beam has a simple astigmatism, ie the ellipses of constant intensity and the ellipses of constant phase have identical orientation. This orientation remains unchanged along the z axis. One of the main axes of these ellipses coincides with they axis and is thus normal to the plane of incidence, whereas the second main axis coincides with the x axis and is thus parallel to the plane of incidence.

4x,o =

( )
c0s e1
~

c0s e2

2
41,0

which is the same as (17a). The beam diameter normal to the plane of incidence does not change. Therefore the beam parameter ql,o in this direction is preserved, leading to (17b).

OPTICS AND LASER TECHNOLOGY.

APRIL

1976

87

again be applied, hence:

4x =

4,. =4,,0

+ j Ip

(12)

The beam parameters q, and q,, are now different, signifying that the beam is astigmatic. The astigmatism ,4 introduced by the plate is equal to the distance between the beam waists parallel and normal to the plane of incidence. The astigmatism is given by:
Beam transformation Fig.2 geometrical optics at a refracting surface using

A = II Re (y,,

4X )

(73)

Medun wth Index of refroctlon

where Re(z) denotes the real part of a conlpiex number 2. Insertion of (22) in (23) yields an expression for the astigmatism:

for a plane-parallel plate, where 1 = d/cos ti2 and d is the thickness of the plate. The result (14) has been derived differently in references 5 and 6. Conclusions
Medium wth Index of refractlon n Plate i wth Index of refractlon no

Fig.3

Derivation

of the plane-parallel

plate astigmatism

Example: astigmatism of a plane-parallel

plate

The astigmatism of the plane-parallel pl;ite is determined by application of the transformation law (17a and 17b) and by taking account of the beam parameter transformation due to beam propagation.9 The plate has a refractive index of np and is embedded in a medium with an index of refraction n (Fig.3). A stigmatic gaussian beam incident from the left may be characterized by the beam parameter CJ,,o. After entering the plate, the beam parameters are:

An incoming stigmatic gaussian beam was decomposed in IO plane waves. After applying the known laws of refraction and transmission at an interface between two media with different indices of refraction, and after summation of the refracted plane waves, an astigmatic gaussian beam results. Thus the beam parameter transformation law (I 7a. 171)) is obtained which forms a basis for the calculation of astigmatism in optical arrangements with refracting surfaces, eg electro-optic light deflectors. 1 The transformation is also useful for computing the modes of optical resonators containing optical elements with Brewster plates. The basic concepts presented in this paper can also be used to calculate the beam parameter transformation, at a refracting surface, of a beam with general astigmatism.* References
Meyer, H., Riekmann, D., Schmidt, K. P., Schmidt, U..J., Rahlff, M., Schroder, E., Thust, W. Design and performance of a 20-stage digital light beam deflector Appl0pt 11 (1972) 1732.. 1736 Baues, I., Hundelshausen, U. v., MGckel, P. A concept for the generation of reproducible and controllable giant pulses Appl Phys Left 21 (1972) 135-137 Born, M., Wolf, E. Principles of Optics (Pergamun Press. Oxford, 1965) Leminger, D. Theorctischc untersuchung der susbrcltung van gaussschen strahlen in optisch isotropen und optisch einachsigen medien. (Diplomarbeit am lnstitut fiir Hochfrequenztechnik und Quantenelektronik Universitat Karlsruhe, 22 September 1970) Habegger, M. A. Astigmatism in light-deflector elements .I Opr SocAv~ 60 (1970) 326-331 Fliigge, J. Praxis der peomctrischcn optik (Vantlcnhoeck & Ruprecht, Gottingen, 1962) Horowitz, B. R., Tamir, T. Lateral displacement 01 a light beans at a dielectric interface J opt Sot, .dInl 61 (197 I) 586 594 Arnaud, J. A., Kogelnik, H. Gaussian light beam? with pcncral astigmatism Appl Opt 8 (1969) 1687 1693 Kogelnik, H., Li, T. Laser beams and resonators ,4ppl Opr 5 (1966) 1550-1567

c0s*e
4x = ___

cos*e2

ql,o
(20)

4y = 41.0

which can be obtained from (17a) and (17b), where qs and q,, are the beam parameters parallel and normal to the plane of incidence respectively. Within the plate the beam travels a distance I, whereupon qx and q,, become: c0s*e,
4x = __
CO82 e 1 41,0 + L nP

I
Yy =

91,o

+ -~ nP

(21)

When the beam leaves the plate the law ( 17a, 17b) must

88

OPTICS AND LASER TECHNOLOGY

APRIL

1976