Surface roughness determination using the acousto-optic technique: Theory and experiment R. Briers and O. Leroy Interdisciplinary Research Center, Katholieke Universiteit Leuven, Campus Kortrijk, B-8500 Kortrijk, Belgium S. Devolder, M. Wevers, and P. De Meester Department of Metallurgy and Materials Engineering, Katholieke Universiteit Leuven, B-3001 Leuven, Belgium Received 14 February 1997; accepted for publication 3 June 1997 With the acousto-optic nondestructive testing technique, phase shifts between the incident and the reflected ultrasound are measured at critical angles of the investigated material. This letter is focusing theoretically as well as experimentally on the influence of a small surface roughness some micrometersof a thick material on the phase information. © 1997 American Institute of Physics. S0003-69519704631-7 In most of the theoretical and practical works on surface roughness only the amplitude of the diffracted ultrasonic field is used for analysis. A detailed overview of all methods developed to study diffraction gratings and a nearly complete list of the most important references on this subject are given in the book of Maystre. 1 A new nondestructive testing tech- nique based on phase measurements in a critical angle of the substrate material was developed at the Department of Met- als and Material Engineering, K. U. Leuven, in close col- laboration with the Interdisciplinary Research Center at the K. U. Leuven, Campus Kortrijk. In the experimental setup this phase information is obtained using two 2 or only one 3 laser beam depending on the kind of information one wants to obtain. A theoretical model 4 was elaborated to determine the influence of a corrugated material on the phase information. Let the coordinate system be chosen as in Fig. 1 such that the periodically rough liquid–solid boundary is described by x , z = f x -z =0 with f x += f x , 1 in which is the profile periodicity. The peak-to-peak am- plitude of the roughness is denoted by h . The time-harmonic incident ultrasonic wave is a Gaussian beam of half-width w with a displacement u i ( x , z ) represented by the Fourier integral 5 u i x , z =1/2 - + V k x expi k x x +k z z  dk x , 2 with V k x = w 0 exp-k i -k x w 0 /2 2 , 3 and whereby k z =k 2 -k x 2 , 4 k i =k sin i , 5 w 0 =k /cos w , 6 k being the longitudinal wave number and i the angle of incidence. The displacement u t r ( x , z ) of the total reflected field can be calculated as u t r x , z = m=- m=+ 1/2 - + R m k x V k x expi k m x -k zm z  dk x = m=- m=+ | u m r x , z | expi m r , 7 or u t r x , z =| u t r x , z | expi t r , 8 with | u t r | , t r , respectively, the amplitude and the phase of the total reflected field and R m ( k x ) the reflection coefficient of the m -th diffraction order. The value of k m can be deduced from the diffraction equation 6 k m =k x +m2 / , 9 and k zm = k 2 -k m 2 . 10 To determine the reflection coefficient R m ( k x ) in Eq. 7the total reflected ultrasonic field u 1 ( x , z ) in the liquid medium 1and the total transmitted ultrasonic field u 2 ( x , z ) in the solid medium 2are decomposed in an infinite sum of plane homogeneous and inhomogeneous plane waves 7,8 represent- ing the different diffraction orders. The amplitudes of these waves are found implementing the generalized boundary conditions 9 along the corrugated surface, being the continu- ity of the normal displacements u 1 =u 2 , 11 and the continuity of stresses j =1 3 T ij 1 j = j =1 3 T ij 2 j ; i =1,3. 12 Since these boundary conditions are periodic functions in x , they can be expanded in Fourier series. A sufficient condition for the final solution is to demand that the Fourier coefficients are equal over one period . Equating these Fou- rier coefficients results in an infinite system of coupled linear 599 Appl. Phys. Lett. 71 (5), 4 August 1997 0003-6951/97/71(5)/599/3/$10.00 © 1997 American Institute of Physics Copyright ©2001. All Rights Reserved.