COATINGS CHARACTERIZATIONS BY THE MIRAGE ...

HAL Id: jpa-00229560https://hal.archives-ouvertes.fr/jpa-00229560

Submitted on 1 Jan 1989

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

COATINGS CHARACTERIZATIONS BY THEMIRAGE EFFECT AND THE PHOTOTHERMAL

MICROSCOPEJ. Roger, D. Fournier, A. Boccara, F. Lepoutre

To cite this version:J. Roger, D. Fournier, A. Boccara, F. Lepoutre. COATINGS CHARACTERIZATIONS BY THEMIRAGE EFFECT AND THE PHOTOTHERMAL MICROSCOPE. Journal de Physique Colloques,1989, 50 (C5), pp.C5-295-C5-310. �10.1051/jphyscol:1989537�. �jpa-00229560�

JOURNAL DE PHYSIQUE Colloque C5, suppl6ment au n05, Tome 50, mai 1989

COATINGS CHARACTERIZATIONS BY THE MIRAGE EFFECT AND THE PHOTOTHERMAL MICROSCOPE

J.P. ROGER, D. FOURNIER, A.C. BOCCARA and F. LEPOUTRE

Laboratoire d'optique Physique, CNRS ER-5, Ecole Superieure de Physique et de Chirnie, 10, rue Vauquelin, F-75231 Paris Cedex 05, France

Resume - L'effet mirage et le microscope photothermique sont des methodes optiques capables de mesurer sans contact les proprietes optiques et thermiques des revetements. Trhs simples dans leur principe, ces methodes sont decrites avec precision par des modhles theoriques et sont ainsi quantitatives. Des mesures de propriet6s moyennes ou locales (la resolution peut atteindre le micromhtre) illustrent les principales applications : detection de trhs faibles absorptions, s,pectres de films semiconducteurs, contr8le non destructif.

Abstract - The mirage effect and the photothermal microscope are totally optical methods able to measure without any contact optical and thermal properties of coatings. Very simple in their principle the methods are described precisely by theoretical models and thus are quantitative. Bulk as well as localized measurements (spatial resolutions about 1 pm can be achieved) illustrate the main applications : very weak absorption detection, spectra of semiconductors films, non destructive testing.

Photothermal methods have achieved a great success since a few years, because they have allowed to solve "difficult" cases in the area of solids state characterization. The combination ' of thermal and optical investigations of samples leads to the determination of optical absorption coefficients and thermal properties of opaque, or diffusing samples, of thin films or multilayered structures. The photothermal methods have been successfully applied to powder samples, non polished samples, crystalline, polycrystalline or amorphous thin films, multilayered structures and to detection of subsurface defects.

We will describe in a first part the general principles of a photothermal experiment and will give the outline plane of the calculations which lead to absolute measurements. In a second part we will give some characteristic examples to describe the different aplications optical absorption measurements, thermal characterization, localization of defects.

1 - EXPERIMENTAL SET-UP When a sample is irradiated by a light flux, there is, most of the time, conversion of the absorbed light energy into thermal energy and the sample is heated. If the impinging beam is modulated, the sample exhibits a modulated temperature. This periodic temperature looks like a wave (thermal wave) which investigates the material. There are several solutions to run directly or indirectly this measurement : photoacoustic scheme (with a microphonic cell), optical scheme (I.R. detection or "mirage" detection). Most of the times these different methods measure the periodic surface temperature. The aim of this section is to describe in details a mirage experiment which was found to be very suitable for thermal waves probing and the most sensitive, and we send the reader

Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jphyscol:1989537

C5-296 JOURNAL DE PHYSIQUE

towards other references in order to find experrsental details of the others detection schemes 1 A short description of the photothermal microscope will be given in section 5.

Modu 1 ated

Pump Beam

F i g . 1 - "Mirage" d e t e c t i o n p r i n c i p l e .

In order to run a photothermal deflection experiment, we have to join a light source such as a dye laser, a xenon arc with a monochromator, a FTIR spectrometer or a visible FT spectrometer /2/ with a "mirage" detection 3 The periodically (at frequency f = w / 2 n ) heated sample surface is an heat source for the surrounding medium which exhibits a periodic temperature gradient near the sample surface. This periodic temperature gradient gives rise to a ' refractive index gradient suitable --for periodically deflecting a probe beam propagating along the surface of the solid. The amplitude of this periodic deflection is given by :

where 4 is the interaction pathway between the probe beam and the temperature gradient dT/dx and n is the refractive index of the medium. We will show below how dT/dx can be calculated and related to the optical properties of the sample. The average in (1) is a spatial average over both the pathway P and finite waist size of the probe beam. This method is particularly sensitive if the sample is immersed in a transparent liquid (in CC14 for instance dn/dT = 5.10-4) and the smallest measured sample surface temperature variations correspond to :

TS = O / J H z in air and

F i g . 2 - Experimental arrangement of t h e compact mirage c e l l .

The figure no 2 describes a new geometry for the "mirage" set-up / 4 / . This cell includes three blocks : the laser compartment, the focusing optics, the position sensor with its filter in order to avoid the stray scattered light. This structure is sufficiently small for setting in the sample compartment of a commercial Fourier transform I .R. spectrometer. Let us underline that a semiconductcr laser can be used lnstead of an He-Me

laser, for probing the index gradient. This lasc source presents two advantages : its small size and a strong reduction of the pointing and intensity noises above 50 kHz in comparison with He-Ne laser ohes.

2 - CALCULATION OF THE PERIODIC TEMPERATURE VARIATION OF THE SAMPLE SURFACE

In order to reach the surface temperature of the sample, we have to solve the heat diffusion equation in all the regions where the heat can propagate. These calculations run by Rosencwaig and Gersho in 1976 /5/ for the case of a photoacoustic detection have been extended to the case of "mirage" detection by several authors /6/ - We will limit here our discussion to the case of a sample deposited on a transparent substrate.

Fluid

(b) (s) <f >

Fig. 3 - Geometry of the 1.D system.

Figure 3 describes the ID geometry of the system. This geometry is the most appropriate for spectroscopic application but it is sometimes more efficient to use a tight focusing of the pump beam to increase the signal.. In this case a 3D calculation is necessary /7/.

The sample is irradiated by a modulated light beam whose amplitude at the frequency &/2n is 8 0 . Thus heat diffusion equations can be written in the three regions shown in fig. 3, transparent fluid (f) sample (s) and transparent backing (b) :

aL Ts I a Ts - = - - - A exp (ax) (exp jut) a x Ds a t

for -1 < x < 0

3% 1 - = - - a T b fo r - 1 - 1 , < x < -1 a x Db a ~

where Ti (i = f, s, b) is the modulated temperature in the region i and Dt the thermal diffusivity. The thermal diffusivity Di = ki/ ftct where kt is the thermal conductivity, cl the specific heat and pi the density.

JOURNAL DE PHYSIQUE

The term A exp(ax) exp(jo t) represents the modulated heat source due to the light absorbed by the sample, a being the optical absorption coefficient.

where q is the light-heat conversion efficiency by non radiative deexcitation. We shall take q = 1 in the following calculations. The general steady state solutions of these equations ( 2 1 , ( 3 ) , ( 4 ) are given by :

region I : ~ t ( x , t) = Ts exp(-or x + jot) ( 5 )

region I1 : Ts(x, t) = [Uexp(osx) +Vexp(-as x)

region 111 : Tb (x, t) = W exp [ub ( x + P ) + j o t] (7)

where Ts, U, V, E and W are complex constants.

a = 1 + p i where pi is the thermal diffusion length

These constants are found when applying the boundary conditions for heat flux and temperature at interfaces.

We can thus write the modulated sample surface temperature : (at x = 0) :

kb U b kb Ds with - b = - - - ( -

ks 0 s ks Db

This quantity b characterizes the thermal impedance adaptation between the sample (coating) and its backing.

This complex temperature can be written as : Ts = ITS I (exp - j 9) . We will now examine two following cases of practical interest for spectrescopic measurement on coatings.

2.1 - Hiah frequency measurements

When the modulation frequency is large enough to have the thermal diffusion length ps much smaller than the sample thickness &' (i.e. thermally thick sample) equation 8 becomes :

where ITs 1 and 9 can be written as an explicit function of the product (a US).

and ? = Atn (2/a2 ps ) + Atn lap/ (2-a@ (11)

We can see that ITs 1 and ? are independent of b (i. e. of the thermal contact between the sample and the backing). The phase variation between very small absorption coefficients and large ones is n/4. We will show in the next section that a careful analysis of the phase signal will allow an absolute defermination of a.

2.2 - Low frequency measurements

We examine now the case of an absorbing layer deposited on a transparent substrate. If the modulation frequency is small to have h < ps i.e., Qa e < 1 one has (exp(+ osP ) I 1 2 ose 1 , and eq. 8 reduces to :

Fig. 4 - Phase var ia t ion o f the surface temperature o f GaAs 610 pm th i ck sample with the op t i ca l absorption c o e f f i c i e n t a for d i f f e r e n t values o f b . -

For very thin coatings (a. ! < < 1) such as micronic semiconductors films the phase of eq. (12) reduces to n/4 independently of a /I/. Fig. 4 shows the influence on the phase of TS of the thermal mismatch between the film and the substrate for a thermally hin GaAs samp e . One can see that the phase variation when going from a 2 < < 1 to a > > 1 is indeed strongly dependent of the parameter b.

2.3 - Multilayered structures

Fernelius /8/ has extended the Rosencwaig and Gersho model to the case of an absorbing layer (thickness h, absorption coefficient ac) deposited on an absorbing substrate (thickness e , absorption coefficient as ) (Fig. 5 ) .

JOURNAL DE PHYSIQUE

-1-1, - 1 0 h

F i g . 5 - Geometry of the 1 .D one layered system.

The sample surface temperature TS can be obtained by solving heat equations in the four media with the suitable boundary conditions. The expression is rather complicated :

where :

l + j and with oi = (i = f, c, s, b)

Pi

This formula (13) is very useful for coatings characterization. We will use it below in the case of a two layers semiconductor system (GaAlAs/GaAs/SiOz). It shows that the surface temperature is sensitive to optical absorption when a lot of thermal parameters (diffusivities Di via pi = JDi/nf, effusivities via c, b, g) are determined. A priori it seems that such a complex process represents a disadvantage. Actually with the help of a microcomputer it is a big advantage which allows localization of optical absorbtion and non destructive testing.

3 - OPTICAL MEASUREMENTS The simplest measurements carried out by the mirage effect are optical absorption (a) determinations at low

lrequenspy . For instance with these

coatings weakly absorbing (exp(-a! ) - 1 - a expression (12) reduces to :

The surface temperature and the mirage deflection are proportional to a. The two following examples demonstrate the extreme sensitivity which can be reached in such measurements.

3.1 - Optical losses determination in mirror coatinqs

Among the optical components which take place in the laser manufacturing, the mirrors play a major part. The dielectric coatings characteristics are especially important for laser-gyro performances, a typical absorption losses level being of the order of a few tens per million / 9 / .

For these reasons it is particularly useful to have, close to a production line, a set-up allowing a fast, precise and sensitive. measurement of absorption losses.

The compact mirage cell designed for these kinds of measurement uses a probe beam which is about colinear with the pump beam. Colinear detection /lo/ is possible as shown in figure 6 because the multidielectric selective mirrors under test, highly reflective at He-Ne pump beam wavelength of 633 nm, are transparent at the wavelength of the laser diode probe beam (780 nm). The heat deposited in the coating diffuses into the substrate (i.e. silica), and the radial temperature gradient is probed within the bulk of this transparent substrate.

If the modulation frequency is chosen low enough in order to neglect the thermal diffusion in the coating, a simple calibration is obtained by using a coating deposited on the same substrate but with an important absorption level so that it can be measured by classical reflection transmission photoelectric experiments.

On the one hand, in this mirage experiment, the signal increases linearly with the incident power of the pump beam ; on the other hand, the noise level, with a good laser diode can reach the photon noise of the position sensor at modulation frequency higher than 200 Hz. So that in our set-up, using a 20 mW He-Ne as pump beam, the signal equivalent to. noise corresponds to an absorption level of one per a million.

JOURNAL DE PHYSIQUE

Fig . 6 - Experimental s e t U D f o r o p t i c a l l o s s e s de terminat ion i n Coat ings usinu t h e c o l i n e a r miraue d e t e c t i o n . -1- 20 mW He-Ne pump l a s e r . -2- Mechanical chopper. -3- Compact mirage c e l l . -4- Lock-in and microcomputer a c q u i s i t i o n d e v i c e s . -5- Coating on s u b s t r a t e under t e s t i n g . -6- Focused He-Ne pump beam absorbed and r e f l e c t e d by t h e c o a t i n g . -7- Laser d i o d e probe beam t ransparent f o r t h e coa t ing and d e v i a t e d i n t o the s u b s t r a t b u l k . - The c o a t i n g t h i c k n e s s ( e l i s much smal ler than the thermal d i f f u s i o n l e n g t h f p s ) .

The set-up is reliable and the experiments can be reproduced within a few percents.

In practice, the sample holder can be easily removed from the mirage cell and its design defines a reference plane for the coating under test with a very good precision without mechanical adjustments.

It is important to note that it takes few seconds for the non specialist operator to put the sample in its holder and to get a numerical result from the lock-in and the microcomputer acquisition devices. Actually, after the coating .operation each sample is systematically tested by this method which constitues the first in-line industrial application of the mirage detection.

Among all the methods that exist to measure optical losses with a high sensitivity, let us recall three principal classes :

The first ones are flying-time decrease measurements methods. These methods allow global losses measurement (Absorption + Transmission + Diffusion) and are limited in practice for high reflectivity mirrors ( > 99 % ) . But it remains delicate to make use of these methods without precise and difficult optical adjustments. This situation is often incompatible with in-line sample tests after manufacturing.

The second ones are calorimetric methods. They allow absorption losses measurement but they take too long time to be compatible with industrial production requirements.

The third group of methods uses photothermal phenomenons. In this group, we have chosen the mirage detection for its simplicity and high sensitivity : the same absorption losses measurements are possible with a

photoacoustic cell, for example, but in that case, sensitivity is smaller by one or two orders of magnitude than the one we have obtained.

3.2 - Optical spectra of amorphous and pol~cr~stalline semiconductors films

Photothermal deflection spectroscopy of amorphous semiconductors deposited on non absorbing substrate has been very successful since it has been proposed /7/, /11/ as a sensitive technique for low absorptance associated to gap and pseudo-gap states. Indeed nowadays more than twenty laboratories are uslng these techniques for material characterization 1121 -

I I I , ov

I C

0.7 a9 11 13 15 1.7 19

Fig. 7 - A-silicon spectrum recorded in C C I J .

The a-silicon spectrum shown on figure 7 has been recorded with a 0.7 pm thick sample deposed on fused silica and immersed in CC16 in order to increase the deflection signal. The sample illumination has been carried out by using a visible and near infrared Fourier 'transform spectrometer /2/ built in our laboratory. With such set-up a! as low as 10-5 can be reached, but a reasonable sensitivity has been also achieved by using xenon arc or quartz halogen tungsten sources associated with conventional monochromators. Let us recat1 that the best classical transmission measurements reach values of a around lo-=.

The amplitude of the deflection signal shows two characteristic features :

- In the visible region for strong a! the deflection signal varies as 1 - exp (-at (eq. 12 in the limit of very small ) . In order to get a , as shown on figure 7, the signal has to be "desaturated".

In the near infrared, for art < 1 we can observe a signal modulation due to optical interferences effect in the film /13/. This modulation, which is due to the variation of optical energy "stored" in the film as a function of the wavelength, is much less important than the large modulation of the transmission obtained with semiconductor samples whose refraction index is large. One can get rid of this modulation by a computation which accounts for the optical properties of the film /14/. As mentioned in section 2.2, the phase of the signal for such thin films ( e < 1 m is constant a 1 the spectrum in such low frequency (10 - 100 Hz) experiments and thus does not carry any useful information.

JOURNAL DE PHYSIQUE

F i g . 8 - O p t i c a l absorp t ion of a 1100 A t h i c k a-germanium b e f o r e and a f t m anneal ing.

Figure 8 shows the thermal deflection spectra of a 1100 A thick amorphous germanium sample before and after annealing. The shoulder at 0.7-0.8 eV for the annealed sample is related to pseudo-gap,state due to dangling bonds /15/. Such an information is difficult to get with other techniques.

4 - THERMAL MEASUREMENTS AS it appears in expression ( 8 ) , the surface temperature and the mirage effect are able to measure two kinds of thermal parameters of coatings : the thermal diffusivity DS and the ratio of the thermal effusivities of the coating and its backing es/ea. The thermal diffusivity characterizes the "velocity" of heat diffusion. This parameter is necessary to know the ability of the coating to evacuate or on the contrary to stop heat. The ratio of the thermal effusivities characterizes the quality of the thermal contact between the coating and its backing.

Furthermore the knowledge of these two quantities is often necessary to determine the absolute absorption coefficient of the coating.

4.1 - Thermal diffusivity measurement of a micron thick semiconductor film

Fig . 9 - Phase l a q between r e a r s u r f a c e and f ront s u r f a c e e x c i t a t i o n s

vs f f / f c )"* . f c o f t e n c a l l e d c h a r a c t e r i s t i c frequency i s equal t o D s / e s Z

1 t o 5 : b = 0 , b = 0 . 3 3 , b = 1, b = 3 , b- The diagram shows the p r i n c i p l e o f the f r o n t s u r f a c e e x c i t a t i o n fF) and the r e a r s u r f a c e e x c i t a t i o n f R )

The sample is a polycrystalline CuInSe2 film, 3.5 pm thick, deposited on glass by evaporation from three sources (copger, indium, selenium). The mirage detection is achieved by a probe beam in air close to the film surface. The excitation comes from an Ar+ laser at 0.514 pm. At this wavelength the absorption length is smaller than 0.1 pm in the film. The phase lag n(e between rear surface and front surface excitations signals is recorded as function of the square root of the modulation frequency ff (see figure 10). The theoretical calculations of A ' P deduced from expressions similar to ( 8 ) show that at. high frequencies, the slope of

vs Jf leads to the thermal diffusivity of the film Ds, while at low frequencies one obtains the ratio ea/es of the effusivities of glass (b) and of the film (s). In figure 10 the experimental points are fitted by a curve which gives eb /es # 1 and Ds = 6.7 10 -' m2 s- . From such a measurement it can be deduced that the contact is good between the coating and the backing (a bad contact would have lead to ea/e~ = 0) and that the thermal properties of the coating are very different from the ones of the single crystal as it can be seen on figure 10 /16/.

AYk

100-

50-

- - * .- -

I I

100 200 q/nzh F i g . 20 - Experimental phase l a g a s a f u n c t i o n f 1 / 2 f o r t h e 3 . 5 urn t h i c k CuInSez film. Continuous curve c a l c u l a t e d w i t h DS = 6 . 7 10- nP s - I

Dashed c u r v e c a l c u l a t e d w i t h Ds = 3 . 1 lo-' JJP s - I ( v a l u e o f t h e d i f f u s i v i t y o f a s i n g l e c r y s t a l ) .

4.2 - Measurements on multila~ered materials

The sample is made of a coating (thickness 370 pm) of GaAs deposited on glass and covered by a film (thickness 9 pm) of GaAlAs. To find the quantitative spectra of all these layers, it is necessary to know their thermal properties. The method used by Yacoubi /17/ in this particular example can be understood by coming back to figure 4. In this figure the phase shift 0 between the signals measured at wavelengths of strong and weak absorptions can be quantitatively related to the ratio of thermal effusivities b.Yacoubi has recorded the phase variations versus the pump wavelength. With the help of the complete theoretical model of Fernelius (see section 2.3) he deduced the thermal parameters of the two layers and reconstituted the spectrum of the GaAlAs film (see figure 11).

C5-306 JOURNAL DE PHYSIQUE

Fig. 11 - a) ~xperimental phase spectrum of GaAlAs deposited on ) f l

5 - LOCAL MEASUREMENTS

4

40

r\

0 , 30:

0, 0 73 " w 2 I ' = l o - , '

0

The measurements described in the previous sections cab be considered as spatially averaged. Actually the samples are always unperfect. The two following examples demonstrate that the photothermal method localizes anomalies in the optical absorption or in the thermal properties. Furthermore with the help of theoretical models these measurements can be quantitative.

a b

5.1 - Local o~tical absorption

500 600 700 800 900 I000 I 1.10 LiS 1.20 2 .U 1.s

h (nm) E(ev)

1 ---

111 i I V j v ' . 1 subrlrall( I i'.. 1

I '--- j I I

t I

--J ........,.........,..-.. :.. * ; , . I I - I . I 1

I I . I

I I I . ..!:::: I I I I ..' I , . . 9

I I I I .. I

I . I I ; 0 . 1

I I . I Ar! & t A l A , :

In the case of the CuInSez film thermally characterized as described in section 4.1., the spectra recorded in amplitude and phase (see figure 12) across the gap allow to distinguish an anomalous absorption located at the surface of the film. The modulation frequency used in this experiment is such that !, > 11, (see section 2.2) . At energies larger than 1.3 eV all the optical absorption is located at the surface of the film (as &>> 1). The phase is equal to zero. Between 1.2 and 1 eV, the film becomes optically transparent (the gap is crossed) and the phase decreases due to a bulk absorption in the film. Below 1 eV, the phase would have to reach a constant value (dashed curve b in figure 12) which can be calculated assuming that the absorption occurs rather uniformely in the thickness of the film. On the contrary, a strong increase of the phase is observed : the phase seems to reach the value corresponding to the surface absorption. This clearly means that centers of absorption are detected on the surface. They are probably due to a surface pollution /18/.

7 to'

E - a" I- : S ; ~

0 Y

Y . . W .

U .

:

; - . .

* I

,.a*..

w'

18

10'

10'

4 ... '. - c : ,,f I '"'

m .

. .- .. . .. .. ,:..: . . .

F i g . 12 - Amplitude f a ) and phase experimental s p e c t r a o f a 3 . 5 p t h i c k CuInSez f i l m .

5.2 - Subsurface air films

It can be shown theoretically that the thermal waves are able to detect very thin slabs of air below the sample surface /19/. More precisely the ultimate thickness of air which can be detected is about 10-4 p, (up = JDs/nf) if the defect is located at less than ps below the surface. This result has been checked experimentally at low frequency /20/ (with ps = I mm, i.e. f = some Hz, 0.1 pm thick air slabs have been detected below 1 mm of metal). In the case of coatings (thickness e s about 1000 A) such a method would normally lead to detection of air slabs as small as .1 A when the modulation frequency is equal to some MHz.

F i g . 13 - T h e o r e t i c a l v a r i a t i o n s o f the mirage phase a s func t ion o f an a i r s l a b t h i c k n e s s l o c a t e d b e l o w a 1000 A t h i c k Cu c o a t i n g on SiOz.

C5-308 JOURNAL DE PHYSIQUE

Figure 13 illustrates the theoretical variations of the phase at f = 10 MHz versus the thickness of an air slab located below a 1000 A copper coating on Si02. Actually it is not possible to keep a diffusive model to thermally describe such a thin del,amination, but this example gives an idea of the noticeable sensitivity of the photothermal methods to thermal defects between a coating and its backing. To realize such an experiment the detection can no longer be achieved as described in figure 1. The two beams (pump and probe) are highly focused (spot diameters about 1 pm) by an optical microscope. The surface temperature can be measured either by the variations of the thermally induced optical reflection coefficient or by the deflection of the probe beam in the thermal lens above the sample surface (figure 14) /21/.

Fig. 14 - Schematic experimental set up of the photothermal microscope. The sam~le (below the microsco~e ob-jectivel is heated by the focused (1 pnP ) - beam of the Art laser.- This pump beam is modulated at frequencies large enough (1 to 20 MHz) to produce a temperature distribution of a micronic spatial extension /US = I to .10 pml . The periodic temperature distribution is probed by the He-Ne laser beam. The two beams colinear in the microscope are separated by a polarizing splitter. The periodic displacements or the periodic amplitude variations of the probe beam are detected on a silicon detector.

The system is just operating in our laboratory, experimental results on coatings will be presented at the meeting ; a qualitative example of micro thermal resistance (grain boundary) is shown in figure 15.

F i g . 15 - Phase s h i f t s and amplitude v a r i a t i o n s o f t h e p e r i o d i c s i q n a l s ob ta ined on two k i n d s o f m i c r o d e f e c t s . - 1 - f a t i g u e crack i n aluminium. The phase s h i f t o f n on the crack i s probably c h a r a c t e r i s t i c o f thermal expansion o f the crack e d g e s . - 2 - gra in boundary i n a ceramic. T h i s experimental behaviour i s c h a r a c t e r i s t i c o f a thermal r e s i s t a n c e .

CONCLUSION

We have shown that the mirage effect is a unique technique for the determination of absolute very weak optical absorption coefficient in cases of thin films deposited on transparent backings. Moreover we have illustrated the ability of this method for the determination of optical parameters of each layer of a multilayered structure such as GaAlAs/GaAs. It will be soon possible to extend these quantitative studies on one layer structures to multilayers ones, to detect thermal mismatch between the layers, and to determine the optical coefficient of each individual layer of a large stack.

Unfortunately, this method needs the simultaneous determination of all the thermal parameters of the system to reach its absolute optical characterization. Thus it is often necessary to proceed both thermal measurements and spectroscopic experiment. Nevertheless, photothermal deflection spectroscopy seems a very fruitful method for the characterization of the microelectronic materials (amorphous silicon transistors - metal oxyde semiconductor transistors...), and for many electrooptical devices whose layered structure is relevant of this technique. We strongly believe that future techniques such as the photothermal microscope will extend this kind of control at a submicronic scale.

REFERENCES

/ 1/ Rosencwaig, A., Photoacoustics and photoacoustic spectroscopy, John Wiley and Sons, New York (1980).

/ 2 / Dkbarre, D., Boccara, A.C. and Fournier, D., Applied Optics, 20 (1981) 4281.

/ 3 / Boccara, A.C., Fournier, D. and Badoz, J., Appl. Phys. Lett. 36 (1980) 130.

C5-310 JOURNAL DE PHYSIQUE

/ 4/ Charbonnier, F. and Fournier, D., Rev. Sci. Instrum. 57 (1986) 1126.

/ 5/ Rosencwaig, A. and Gersho, A., J. Appl. Phys. 47 (1976) 64.

/ 6/ Murphy, J. and Aamodt, L., J. Appl. Phys. 51 (1980) 4580.

/ 7/ Jackson, W., Amer, N., Boccara, A.C. and Fournier, D., Appl. Opt. 20 (1981) 1333.

/ 8/ Fernelius, N.C., J. Appl. Phys. 51 (1980) 1.

/ 9/ Elleaume, P., Velghe, M., Billardon, M. and Ortega, J.M., Appl. Opt. 24 (1985) 2762. -

/lo/ Roger, J.P., Charbonnier, F., Fournier, D. and Boccara, A.C., Proceedings of LASERS'88 edited by the Society for Optical & Quantum

Electronics (1989).

/11/ Fournier, D., Boccara, A.C. and Badoz, J., Applied Optics 21 (1982) 74.

/12/ Amer, N. and Jackson, W., 1984, in "Semiconductors and Semimetals", Vol. 21 B, Pankove, J.I., Academic Press, New-York, 1984.

/13/ Buckley, R. and Beaglehole, D., Appl. Optics 16 (1977) 2495.

/14/ Cody, G.B. "Semiconductors and Semimetals", Vol. 21B, edited by Pankove, J.I., Academic Press, New York (1984).

/15/ Theye, M.L., Georghiu, A., Driss-Khodja, K. and Boccara, A.C., Proceedings of the 11th Int. Conf. on amorphous and liquid semiconductors. Rome, 1985.

/16/ Roger, J.P., Lepoutre, F., Fournier, D. and Boccara A.C., Thin Solid Films 155 (1987) 165.

/17/ Yacoubi, 1986. Ph. D. Thesis, University of Montpellier (France). Yasa, Z., Jackson, W. and Amer, N., Appl. Opt. 21 (1982) 21.

/18/ Roger, J.P., Lepoutre, F., Fournier, D. and Boccara, A.C.. In "Photoacoustic and Photothermal phenomena" edited by P. Hess and J. Pelzl. Springer Series in Optical Sciences, Vol. 58, p. 245 (1988).

/19/ Lepoutre, F., Fournier. D. and Boccara, A.C.. in "Non destructive characterization of materials", edited by J.F. BussiBres, J.P. Monchalin, C.O. Ruud and R.E. Green Jr., Plenum Press (1987).

/20/ Lepoutre, F., Charbonnier, F., Le Liboux, M., Nahoum, R., Fournier, D. and Boccara, A.C., "In Photoacoustic and Photothermal Phenomena", edited by P. Hess and J. Pelzl, Springer Verlag Series in Optical Sc,iences, Vol. 58, p. 412 (1988).

/21/ Opsal, J., Taylor, M.W., Lee Smith, W. and Rosencwaig, A., J. Appl. Phys. 61 (1987) 240.