Reading Cource Report-2
Reading Cource Report-2
Reading Cource Report-2
Reading course -2 Report submitted to the faculty of Homi Bhabha National Institute In partial fulfillment of the requirement for the degree of Doctor of philosophy
Indus Synchrotron Utilization Division RAJA RAMANNA CENTER FOR ADVANCED TECHNOLOGY INDORE, M.P.- 452013
August 2011
CONTENTS
1- Introduction 2- Scattering geometry 3- Basic principle 4- Layered system 4.1- Single layer film on substrate 4.2- Bilayer film on substrate 4.3- Multilayer structure 5- Effect of interfacial width on reflectivity 6- Types of roughness 7- Effect of roughness on reflectivity of a surface 7.1-From a single surface 7.2- From a thin film (A single layer on a substrate) 8- Experimental techniques 9- Books referred
ii
1-Introduction:Precise characterization of thin film/multilayer structures in terms of their structural properties is of utmost importance for their performance evaluation. These artificial structures are deposited under non-equilibrium conditions; there exists considerable deviation of the parameters from the designed values. The structural parameters of importance are layer thicknesses, density, roughness, interdiffusion, crystal structure, and crystallite size. The techniques used to extract information about these parameters are grazing incidence x-ray reflectivity, wide angle x-ray diffraction, atomic force microscopy, and transmission electron microscopy etc. The present reading course covers the details of x-ray reflectivity technique and its application for characterizing thin film/ multilayer structures. Structural characterisation with x-rays is a non-destructive way of studying the surfaces and interfaces at atomic scale. Moreover, the x-ray techniques can be made surface sensitive by working at grazing angles of incidence since the penetration depth is restricted to the surface layers. In the x-ray region, the refractive index of all materials being less than unity, total external reflection occurs when the radiation with wave vector medium from vacuum for grazing angles less than the critical angle classical electron radius and is incident on a , where is
is the electron density. Below the critical angle, the wave vector
inside the medium is purely imaginary and hence the wave cannot propagate. Thus only an evanescent wave which decays exponentially inside the medium exists within the penetration depth shown in figure-1.
Fig-1 Evanescent wave (exponentially decaying with increasing distance from the interface).
bends away from the normal as given by the Snells law. A schematic of reflection and refraction taking place at the boundary between vacuum and medium of refractive index n is shown in figure-2.
2-Scattering geometry:In a coplanar geometry of x-ray scattering the incident wave, reflected wave and surface normal vector lie in the same plane (figure-2) and wave vector defines the direction of
perfectly monochromatic wave. Distribution of scattered intensity in a reciprocal space gives the information about structural parameters of a sample.
Fig-2 (a) Reflection and transmission of plane electromagnetic wave, represented by wave vector of a medium optically characterized by refractive index
In the following section the mapping of Cartesian component of wave vector transfer in a sample is discussed along with the formalism of a multilayer reflectivity for smooth and rough interfaces. Geometry of x-ray scattering from a medium of refractive index is shown in figure-
2. In specular reflection condition, both incident and reflected waves, defined by wave vector and respectively, make equal angles with the boundary. Since both .
2
and
propagate
1 2
The parallel component of momentum transfer vector lying in xy plane is zero. The
perpendicular component (along z-direction) given by above equation is changed either by varying incident angle at the fixed wavelength or by varying wavelength at the constant angle . A general case of scattering is depicted in figure-3.
and
and and
respectively then momentum transfer along the x and z directions is given by plane while xz is a scattering plane.
and
transfer takes place in x direction also. Such scans where the two angles are kept unequal are subject of diffuse scattering experiments. These experiments yield the information about the lateral surface structure.
3 4
These x and z components can be varied experimentally in four different ways by coupled and independent motion of the sample and detector as listed below:
(a) Specular scan:If both sample and detector are rotated such that the incident and reflected angles are equal i.e. specular scan or then it results in scan. while . This scan is known as
(b) Offset scan:If both sample and detector are rotated such that a constant offset angle between them is maintained i.e. This scan is called offset scan.
(c) Sample rock scan:If detector is kept fixed at a constant angle rocked around a central value and only sample is , then this scan is called sample rock scan.
(d) Detector scan:When sample is kept fixed at a constant incidence angle and only detector is scanned i.e. fixed and varied, then this scan is called detector scan.
These scans are useful to probe the sample structure along the x and z direction both. In scan, wave vector transfer restricts in the z-direction. Thus, this scan yields the structural
information of a sample in the perpendicular direction, where it is assumed that the sample is laterally homogeneous and scattering process will not change the degree of coherence of the xrays. Nevertheless, the laterally inhomogeneous or rough structure destroys the degree of coherence of the waves. Because of the non-homogeneity arising from, surface roughness, the surface normal is not uniquely defined, and it increases the diffuse scattered background. The
scan modes described above are used to record the diffuse scattering components, which yield the information of roughness correlation.
3-Basic Principle:Reflection and refraction of electromagnetic radiation at the interfaces between two materials depend on the optical index contrast of the materials. Since the frequency of x-ray radiation is greater than the resonance frequency of the medium (in medium, electron undergoes harmonic oscillations in presence of the electric field of radiation), the refractive index is found close to unity for all the materials and is given as imaginary parts of refractive index, If interface, the law of refraction gives where and are real and
is exit angle of refracted radiation for vacuum/medium . For , the wave cannot propagate in
the medium. The corresponding incidence angle is called the critical angle, which is defined as , where is electron density of the medium. Below the critical angle, the is purely imaginary. Thus, wave . Except for small losses due in the critical
wave vector inside the medium cannot propagate inside the medium for all incident angle
to absorption; all other incoming radiation will be reflected. Penetration depth angle region is restricted to few tens of angstrom and defined as
5 6
Therefore the x-ray reflectivity technique becomes highly surface sensitive at extreme grazing angles. A typical penetration depth curve for two different wavelengths, x-rays) and (soft x-ray), in molybdenum material is shown in figure-4. (hard
Fig-4 X-ray penetration depth inside molybdenum film at two different wavelengths (soft x-ray)as a function of normalized angle of incidence
Below critical angle the penetration depth is few tens of angstrom whereas for increases drastically. As soft x-rays are highly absorbing in all materials penetration depth is only few hundreds of angstrom even in
it the
in micrometer range for harder x-rays where absorption term is very small
Reflection from a single smooth boundary can be calculated using the well-known Fresnel formulas written below. For a single vacuum/medium interface, the Fresnel reflection and transmission coefficients for s polarized light can be written as
8
where and are glancing angles of incidence and refraction respectively. For p
polarized radiation (electric field vector in the scattering plane), reflection and transmission coefficients gets modified; but for grazing incidence, either state of polarization yields identical results for . The above coefficients are complex since the refractive indices are complex. In
6
an experiment, the measured parameter is the Fresnel reflectivity R F that is given as the modulus square of the amplitude reflection coefficient r. The equation for Fresnel reflectivity can be simplified in the small angle approximation for as
9
An example of s-polarized reflectivity from a single smooth surface as a function of normalized incidence angle almost 100%, if is given in figure-5. Below the critical angle the reflectivity is , the reflectivity drops drastically
and the reflectivity fall is inversely proportional to the fourth power of momentum transfer vector for all . The total external reflection region below the critical angle is very ratio increases the shape of the curve
Fig-5-X-ray reflectivity from a single surface as a function of normalized angle of incidence angle, the total external reflection yields the 100% reflectivity if =0. As gets modified.[Ref-2]
Below critical
approximation for a silicon surface. The solid line corresponds to the Porod law.
Fig-6- Fresnel reflectivity of silicon surface at shown as open circles. The solid line is the
with
and
is
angle x-ray scattering. The inset shows the Fresnel transmitted intensity for the same set of parameters. [Ref-3]
Porod law is applicable to small angle x-ray scattering which has a similar functional dependence. The deviation of the Porod law at small q-values is due to the approximations involved viz. neglect of refraction and multiple scattering effects. This regime is also known as the Born approximation (BA) or the kinematical approximation. In the Born approximation, the scatterer is assumed to be illuminated by a plane wave and the scattering is calculated as if the scatterer is absent. Each element of the sample scatters the incident wave and the resulting scattered field by all the elements at a distant point are summed up to calculate the total scattering. This neglects the multiple scattering effects and is thus valid for weak scattering, where reflectivity is approximately less than 10%. It can be observed that the reflectivity drops by five orders of magnitude over a range of 10 times the critical wave vector transfer. Hence a very intense x-ray source is required for reflectivity measurements to collect the data over a sufficient q-range so as to enable the determination of the electron density variation with good accuracy. The Fresnel transmitted reflectivity is shown in the inset. For , the incident and
8
specular wave fields are in counter-phase at the surface which leads to zero intensity in the transmitted (evanescent) wave field. For , the phases of the two waves match at the
surface which gives rise to a two-fold enhancement in the transmitted (evanescent) wave field in the medium. This leads to a four-fold enhancement in the transmitted intensity in the absence of absorption.
4- Layered System:Fresnels reflectivity formalism for a single smooth surface can be extended to explain the x-ray reflectivity from a layered media of different materials. Adding the wave field reflected from different boundaries in a recursive manner and applying the proper boundary conditions yield the reflected intensity from a multilayer structure.
4.1-Single layer film on a substrate:In case of a thin film of a material deposited on the substrate, there are two interfaces viz. vacuum/film and film/substrate that are separated by the film thicknessd. The interference between the radiation scattered from the two interfaces results in the net reflectivity of the structure. Let the Fresnel reflected amplitude of the top and bottom interfaces be and respectively and be the normal component of the wave vector transfer in
vacuum, film and substrate respectively. The reflected amplitude of the thin film structure is given as,
10
with
11
Where
is the phase difference between the scattered amplitude from the two interfaces
which give rise to the oscillatory behaviour in the reflectivity pattern measured as a function of angle or the wave vector transfer. The oscillation frequency determines the thickness of the film; if q is the difference in wave vector transfer (in the film taking refraction into account) between
9
amplitude of the oscillations is proportional to the contrast in electron density between the film and the substrate. If only the single scattering events from the two interfaces are considered, the resultant reflectivity can be described by the numerator of equation-10 alone and the denominator term can be neglected. The multiple scattering effects give rise to higher order terms and the denominator term in equation-10 becomes significant. Another important aspect is the relation between the film thickness and the oscillation frequency. As the thickness of the film increases, the angular separation between two successive peaks decreases. This can be understood by the fact that the reflectivity is the Fourier transform of the electron density variation in the real space. This is illustrated in figure-7 which shows the simulated curve of a 200 and 800 thick Pt film on a silicon substrate.
B
10
0
10
-1
Reflectivity
10
-2
10
-3
10
-4
10
-5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
10
4.2-Bilayer film on a substrate:The formula for single layer film on the substrate can be extended to that of a two layer film, A and B, on the substrate. The reflectivity would be due to the interference of radiation scattered from three interfaces viz. vacuum/A, A/B and B/substrate. Due to the presence of two layers, the oscillation frequency of the thicker layer is modulated by that of the thinner layer. Using the analogy given in the above section for single layer films, the difference in wave vector transfer (in the medium) between the successive maxima of the high (low) frequency modulation gives the thickness of the thicker (thinner) layer. Figure-8 shows x-ray reflectivity spectra corresponding to different thicknesses of Mo/C bilayer structures. Curve corresponding to Mo and C layer thickness of 50 is marked as a where two frequency modulation viz. 50 and 100 are shown. Similarly curve marked as b is corresponding to CMo bilayer of 100 and 800 thicknesses respectively. This clearly shows the high frequency oscillations due to the thick Mo layer modulated by the thinner C layer.
10 10 10
-1
Reflectivity
10 10 10 10 10
-2
-3
(a)
-4
(b)
-5
-6
0.0
0.5
1.0
1.5
2.0
2.5
3.0
on Si substrate showing the frequency of the total film thickness modulated by the individual layer thicknesses. The curves are vertically displaced for clarity.
11
4.3-Multilayer structure:A multilayer structure consists of alternate layers of two materials A and B repeated over N periods. Reflection and transmission of the incident wave field inside a layered media is illustrated in figure-9. Interfaces formed at the boundaries between two materials are assumed to be sharp. The material and its structure both are homogeneous along the surface that is parallel to xy plane. The change in material occurs only in the z-direction. N layers of two materials between vacuum and substrate having the thicknesses recognized by the refractive indices sample at an angle are optically
inside the sample. Final reflection amplitude from a multilayer structure is calculated with the boundary condition that the tangential component of electric and magnetic field vectors.
Fig-9- Representation of incident reflected and transmitted waves inside a periodic multilayer system. Corresponding vector diagram for incident and reflected wave is also shown.
12
The multilayer structure is characterised by its period (distance between two layers of the same composition) and its ratio period. A multilayer with a between the thickness of the high electron density layer to the value equal to an integer j, suppresses the jth-order
harmonic and its multiples. The periodicityd in the real space gives rise to Bragg diffraction peaks in the reflectivity pattern at wave vector values of . In addition to the Bragg peaks,
N-2 subsidiary peaks are observed between the Bragg peaks due to the finite total thickness of the film. As N is increased, the intensity of the Bragg peaks increases and that of the subsidiary oscillations decreases. The reflected intensity distribution from a periodic multilayer stack can be represented by the following equation
12
where is the envelope function which depends on the composition profile within the
period. The relative thicknesses of the constituent layers in the period modulate the reflected intensity distributed into different orders of Bragg peaks. The simulated reflectivity pattern of Mo/Si multilayer having a period of 50 for 10 layer pairs (i.e.20 layers) is plotted in figure-10 for three different values of .
10
-1
gamma 0.4
Reflectivity
10
-3
10
-5
(a)
10
-1
gamma 0.5
Reflectivity
10
-3
10
-5
(b)
10
-1
gamma 0.33
Reflectivity
10
-3
10
-5
(c)
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Fig-10- Simulated reflectivity spectra of Mo/Si multilayer with a period of 50 at values (a) =0.40,(b) =0.50, and(c) =0.33.
using different
13
For =0.40, since 1/ is not an integer, there is no preferential attenuation in the intensity of the Bragg peaks and the higher orders decrease monotonically in intensity. For =0.50, the intensity at the even orders are suppressed. For =0.33, the third order and its multiples are suppressed.
5- Effect of interfacial width on reflectivity:The amplitude reflectance can be written within the Born approximation as:
13
Equation-13 implies that the amplitude reflectance is the Fourier transformation of the
gradient of electron density variation as a function of depth. When the electron density variation (z) in a medium is known, the calculation of reflectance of the medium is defined as the direct problem of reflection. On the contrary, in practice, of the medium has to be evaluated from which is thereby
defined as the inverse problem of reflection. The above equation is analogous to the definition of structure factor in conventional x-ray diffraction. The reflectivity is given as
14
is defined as a step
is thereby given by a delta function whose Fourier transform is unity which . Thus an abrupt interface gives rise to a reflectivity that is given by the
Fresnel formulae. In practice, the electron density variation across the interface between two media will not be abrupt and will have a gradual transition over a finite width. Figure-11 shows a typical variation of of having the functional form of an error function and the corresponding gradient is also shown which has a Gaussian distribution of width
14
Fig-11- Schematic diagram of density variation at the interfaces (a) For ideal case where density varies as step function at interface, (b) For the real system where density varies gradually at the interfaces. (c) Variation of .
15
which causes damping of the reflected intensity by the factor , termed as the static Debye-Waller factor. This term is analogous to the Debye-Waller factor used to model the thermal disorder in conventional x-ray diffraction, the basic difference being that the interfacial width is a frozen disorder. The interfacial width and the profile are the vital parameters to be determined from the measured reflectivity data.
6-Types of roughness:For a single layer film, the interfacial width scales with the film thickness. The same may not be valid for a multilayer structure due to the different growth behaviour of the two materials forming the multilayer structure. Roughness in multilayer structures can be divided into four limiting cases that are schematically shown in figure-12.
15
Fig-12- Schematic model of four different types of interface roughness in multilayer structures: (a) uncorrelated, (b) cumulative, (c) partially correlated, and (d) correlated.
(a) Uncorrelated type:- In this case, no replication occurs towards the free surface. The
roughness profiles of different interfaces are independent.
(b) Cumulative type:- The roughness increases towards the free surface, indicating
roughening during growth.
(c) Partially correlated type:- The roughness of a layer is partially transmitted towards
free surfaces.
16
The assumption made in the derivation of equation-16 is that the momentum transfer q is constant within the transition into the material, but the change in refractive index changes the propagation angle and the value of q within the transition layer from a value q in medium-1 to a value in medium-2, where the refraction can be ignored (Born Approximation). So by (geometric average), where q and are q values in the two media forming
replacing
the boundary, gives a very good approximation to a regourous theory (Novot Croce Model) and yields for the reflectivity reduction in the amplitude and in intensity.
16
17 7-Effect of roughness on reflectivity of a surface:7.1- From a single surface:Figure-13 shows the effect of interfacial roughness on the reflectivity of a single surface. As an example, the simulated reflectivity (for a wavelength of 1.54 ) of a silicon wafer is shown using the Gaussian interface model, for; (a)-zero roughness and (b)-7 roughness; It can be observed that the detrimental effect of roughness on the reflectivity is higher at larger angles (larger q-values).
10
0
10
-1
= 0.0 A = 7.0 A
10
-2
Reflectivity
10
-3
10
-4
10
-5
10
-6
10
-7
0.0
0.5
1.0
1.5
2.0
17
7.2 From a thin film (A single layer on a substrate):As explained above, the amplitude of the interference oscillations from a single layer film depends on the electron density contrast between the film and the substrate. The presence of finite roughness at the vacuum/film interface and at the film/substrate interface
causes further reduction in the overall reflectivity and amplitude of the oscillations. The two roughnesses affect the reflectivity pattern differently thereby allowing their independent determination. The roughness of the top surface affects the overall rate of fall in the reduces the amplitude of
reflectivity with angle whereas the roughness of the bottom interface the oscillations with increase in angle. The amplitudes and
a smooth interface are modified for a rough interface. Thus by fitting the measured reflectivity data of a thin film to equation-10, the following parameters can be precisely determined. (a)- Film thickness (b)-Electron density of the film (c)-Width of the bottom and top interfaces. The above aspects are explained with the help of simulated curves in figure-14 where the following four cases are presented.
18
(i)(ii)-
and
and
with increase in angle. (iii)and : The bottom interfacial roughness reduces the amplitude of the
oscillations with increase in angle without affecting the envelope. This can be understood by examining equations (10) and (17) where the roughness term containing multiplies the
oscillatory terms in the numerator as well as in the denominator whereas the term containing (iv)multiplies only the less significant term in the denominator. and : The reduction in reflectivity is more severe than the above cases. It
can be observed that the amplitude of the oscillations is still preserved even at higher angles and is because the two waves are reduced approximately by the same amount and hence their coherence is preserved.
8-Experimental Techniques:The experimental requirements for x-ray reflectivity technique are similar to those found in a standard goniometer, in which a monochromatic x-ray beam is used and the scattering
vector is scanned by scanning the scattering angle. Other experimental considerations are source, detector, data collection, resolution, etc. which are similar to those for grazing incidence x-ray scattering measurements. A schematic of the typical angle scanning configuration is shown in Figure-15.
19
The incident and reflected angle are equal so that the scattering vector is aligned normal to the sample surface (i.e., in the specular reflection orientation). The critical wavevector for total external reflection is very small, typically on the order of angle on the sample must be about . This means that the incident
angular divergence significantly less than this value, with either slits or a monochromator crystal. The strict limitation on angular divergence typically precludes the use of focusing optics to increase the beam intensity. In addition, at this grazing angle, a small error in the position of the sample will make the beam miss the sample. This problem of the beam footprint aligning the beam to be completely contained on the sample is one of the most significant challenges of x-ray reflectivity. Further, the strong dependence of the Fresnel reflectivity on the scattering vector means that a large dynamic range, typically five to six orders of magnitude, is required to measure a reasonable range of scattering vectors. The scattering geometry is usually chosen to have the scattering plane vertical to take advantage of the small opening angle of the synchrotron radiation perpendicular to the plane of the ring. Resolution is determined usually by the rocking curve width of the monochromator or analyzer crystals. Furhter energy dispersive x-ray raflectivity has been devloped as an alternative to the conventional angle-scanning approach described above. In this type of reflectometer,a broadband source of radiation is employed in a fixed angle-scattering geometry, using an energydispersive detector to measure the reflectivity as a function of x-ray energy. A typical in situ experimental arrangement is shown in figure-16.
20
In this type of reflectometer x-ray energies up to 40 keV are typically used to obtain a useful range of scattering vectors. The characteristic lines of the source provide a convenient energy calibration for the detector, The detector is an energy-dispersive intrinsic Ge detector. The entire energy spectrum is collected simultaneously and transferred to computer for analysis. The resolution of the detector is typically about 1%, which limits the resolution of the
reflectivity spectrum. The optics required are extremely simple. Only slits are used to define the incident and exit angles.
21
9- Books referred:1- Jens Als- Nielsen Des McMorrow, Elements of modern x-ray physics, (John Wiley & Sons, 2001). 2- Mohammad Hussein Modi, Surfaces and interfaces studies in multilayers. (PhD Thesis 2005). 3- N Suresh, PhD Thesis 1998. 4- E. Spiller, Soft X-ray Optics (SPIE, Bellingham, USA, 1994). 5- Ullrich Pietsch, Vaclav Holy and Tilo Baumbach, High- Resolution X-Ray scattering From Thin Films to Lateral Nanostructures, (Springer,2004). 6- M. Tolan, X-Ray scattering from soft-matter thin films, Springer Tracts in Modern Physics, (Springer1999). 7- D. Attwood, Soft x-rays and extreme ultraviolet radiation: Principles and applications (Cambridge University Press, 1999). 8- Ron Jenkins, X-ray Diffraction, (John Wiley & Sons, 1996). 9- E.chason and T.M.Mayer, Thin film and surface characterization by specular x-ray reflectivity,(Critical Review in Solid State and Material sciences,22(1):1-67 (1997).
22