THIN FILM DEPOSITION AND CHARACTERIZATION
The act of depositing a thin film to the surface is thin-film deposition and it is any technique for depositing a thin film of material onto a substrate or onto previously deposited layers. Most deposition techniques control layer thickness within a few tens of nanometers molecular beam epitaxy allows a single layer of atoms to be deposited at a time. It is useful in the manufacture of optics (for reflective, anti- reflective coatings or self-cleaning), electronics (layers of semi-conductors, and conductors from integrated circuits etc.
The synthesis and characterization of copper oxide Cu2O thin films via different techniques have attracted considerable attention due to their potential application prospects in solar cells.
Deposition techniques fall into two broad categories, depending on whether the process is primarily chemical or physical.
Physical deposition uses mechanical, electromechanical or thermodynamic means to produce a thin film of solid. A typical example is the formation of frost. It is categorized into;
ü Thermal evaporator uses an electric resistance heater to melt the material.
ü Electron beam evaporator; fires a high-energy beam from electron gun to boil a small spot of material; since the heating is not uniform , lower vapor pressure material can be deposited.
ü Molecular beam epitaxy (MBE); slow streams of an element can be directed at the substrate, so that material deposits one atomic layer at a time.
ü Sputtering relies on a plasma (usually gases like argon) to knock material from a “target” a few atoms at a time.
ü Pulsed laser deposition systems work by an ablation process.
ü Electro hydrodynamic deposition is a relatively new process of thin film deposition.
ü Cathodic arc deposition (arc-PVD) which is a kind of ion beam deposition where an electric arc is created that literally blasts ions from the cathode.
In chemical deposition, a fluid precursor undergoes a chemical change at a solid surface, leaving a solid layer. It is further categorized by the phase of the precursor:
ü Plating relies on liquid precursors, often a solution of water with a salt of the metal to be deposited.
ü Chemical solution deposition (CSD) uses a liquid precursor, usually a solution of organ metallic powders dissolved in an organic solvent.
ü Spin coating uses liquid precursor or sol-gel precursor deposited onto a smooth flat substrate which is subsequently spun at a high velocity to centrifugally spread the solution over the substrate.
ü Chemical vapor deposition (CVD) generally uses a gas-phase precursor, often a halide or hydride of the element to be deposited.
ü Atomic layer deposition (ALD) uses gaseous precursor to deposit conformal thin films one layer at a time.
ü SILAR method; the successive ionic layer absorption and reaction method is used to create thin films from a variety of different coatings for applications including solar panels and semi conductors.
Cu2O is p-type semiconductor with a band gap of approximately 2.0 eV [1-2] and a cubic structure. Cupric oxide (CuO) is also a p-type semiconductor having a band gap of 1.21–1.51 eV and monoclinic crystal structure. The different physical and chemical techniques that have been utilized to grow Cu2O thin film on glass  include reactive sputtering, chemical vapor deposition, spraying, thermal oxidation, electrodeposition etc. In SILAR, thin films are obtained by immersing the substrate into separately placed cationic and anionic precursors. Between every immersion it is rinsed in distilled water or ion exchanged water. For Cu2O deposition, the method consists of alternate immersion of a glass substrate in a complex of copper ion and hot NaOH solution. Fixed concentration of copper complex solution (0.5 M) and NaOH solution (1.0 M) was used in these.
A thin film is a layer of material ranging from fraction of a nanometer (monolayer) to several micrometers in thickness.
Electronic semiconductors devices and optical coating are the main applications benefitting from thin films construction.
A familiar application of thin films is the household mirror which typically has a thin metal coating on the back of a sheet of glass to form a reflective interface.
The process of silvering was once commonly used to produce mirrors. A very – thin film coating (less than a nanometer thick) is used to produce two-way mirror.
The performance of optical coatings (e.g. entire reflective or AR coating) are typically enhanced when the thin film coating consists of multiple layers having varying thickness and refractive indices. Similarly, a periodic structure of alternating thin films of different materials may collectively form a so-called super lattice which explains the phenomenon of quantum confinement by restricting electron phenomenon by two-dimensions.
Work is being done with ferromagnetic and ferroelectric thin films for use as computer memory. It is also being applied to pharmaceuticals via thin film drug delivery. High hardness and inertness of ceramic materials make this type of thin coating of interest for protection of substrate materials against corrosion, oxidation and wear. In particular, the use of such coatings and cutting tools can extend the life of these items by several orders of magnitude. Research is being done on a new class of thin-film inorganic oxide materials called amorphous heavy-metal cation multicomponent oxides which could be used to make transparent transistors that are inexpensive, stable and environmentally benign.
The act of applying a thin film to a surface is thin film deposition any technique for depositing a thin film of material onto a substrate or onto previously deposited layers. “thin” is a relative term, but most depositing techniques control layer thickness with a few lens of nanometers molecular beam epitaxy allows a single layer of atoms to be deposited at a time.
DEPOSITION OF THIN FILMS
Here a fluid precursor undergoes a chemical change at a solid surface, leaving a solid layer. An everyday example is the formation of soot on a cool object when it is place inside a flame. Since the fluid surrounds the solid object, deposition happens on every surface, with little regard to direction. Thin film from chemical deposition techniques tend to be conformal rather than directional.
Chemical deposition is further categorized by the phase of the precursor.
PLATING relies on liquid precursor often a solution of water with a salt of the metal to be deposited some plating processes are driven entirely by reagents in the solution usually for noble metals but by for the most commercially important process is electroplating. It was not commonly used in semiconductor processing for many years, but has seen a resurgence with more under spread use of chemical-mechanical polishing techniques.
CHEMICAL SOLUTION DEPOSITION (CSD) OR CHEMICAL BATH DEPOSITION (CBD) Uses a liquid precursor, usually a solution of organ metallic powders dissolved in an organic solvent. This is a relatively inexpensive, simple thin film process that is able to produce stoichiometrically accurate crystalline phase. This technique is also known as the sol-gel method because the ‘sol’ (or solution) gradually evolves towards the formation of gel-like diphasic system.
SPM coating or spin casting uses a liquid precursor, or sol-gel precursor deposited onto a smooth, flat substrate which is subsequently spin at a high velocity to centrifugally spread the solution is spin and the viscosity of the sol determine the ultimate thickness of the deposited film. Repeated deposition can be carried out to increase the thickness of films as desired. Thermal treatment is often carried out in order to crystallize the amorphous spin coated film. Such crystalline films can exhibit certain crystalline preferred orientation after crystallization on single crystal substrates.
CHEMICAL VAPOUR DEPOSITION (CVD) Generally uses a gas-phase precursor often a halide or hydride of the element to be deposited. In the case of MoCVD, an organ metallic gas is used. Commercial techniques use low pressure of precursor gas.
PLASMA ENHANCED CVD (PECVD) Uses an ionized vapour, or plasma as a precursor, unlike the soot example above, commercial PECVD relies on electromagnetic means (electric current, microwave excitation) rather than a chemical reaction to produce a plasma.
ATOMIC LAYER DEPOSITION (ALD); Uses gaseous precursor to deposit conformal thin films at a time. The process is split up into two half reactions, runs in sequence and repeated for each layer, in order to ensure total larger saturation before beginning the next layer. Therefore, one reactant is deposited first, and then the second reactant is deposited during which a chemical reaction occurs on the substrate, forming the desired composition.
As a result of the stepwise, the process is slower than CVD, however it can be run at low temperatures, unlike CVD.
Uses mechanical, electromechanical or thermodynamic means to produce a thin film of solid. An everyday example is the formation of frost. Since most engineering materials are held together by relatively high energies, and chemical reactions are not used to store these energies, commercial physical deposition systems to require a low-pressure vapour environment to function properly. Most can be classified as physical vapour deposition (PVD).
The material to be deposited is placed in an energetic entropic environment, so that particles of material escape its surface. Facing this source is a cooler surface which draws energy from these particles as they arrive, allowing them to form a solid layer. The whole system is kept in a vacuum deposition chamber to allow the particles to travel as freely as possible since particles tend to follow a straight, films deposited by physical means are commonly directional rather than conformal. Examples of physical deposition include;
A THERMAL EVAPORATOR uses an electric resistance heater to melt and raise its vapour pressure to a useful range. This is done in a high vacuum, both to allow the vapour without reacting with or scattering against other gases. Phase atoms in the chamber, and reduce the incorporation of impurities from the residual gas in the vacuum chamber. Obviously, only materials with higher vapour pressure than the heating element can be deposited without contamination of the film. Molecular beam epitaxy is a particularly sophiscated form of thermal evaporation.
AN ELECTRON BEAM EVAPURATOR fires a high-energy beam from an electron gun to boil a small spot of material, since the heating is hot uniform; lower vapour pressure materials can be deposited. The beam is usually bent through an angle of 2700cin order to ensure that the gun filament is not directly exposed to the evaporant flux. Typical deposition rates for electron beam evaporation range from 1 to 10 nanometer per second.
IN MOLECULAR BEAM EPITAXY (MBE) slow streams of an element can be directed at the substrate so that material deposits are atomic layer at a time. Compounds such as gallium arsenide are usually deposited by repeatedly applying a layer of one element (i.e. gallium) then a layer of the other (i.e. arsenic) so that the process is chemical, as well as physical. The beam of material can be generated by other physical means (that is by a furnace) or by a chemical reaction (chemical beam epitaxy).
SPUTTERING relies on a plasma (usually a noble gas such as argon) to knock material from a target a few atoms at a time. The target can be kept at a relatively low temperature since the process is not one evaporation, making this one of the most flexible deposition techniques. It is especially useful for compounds and mixtures, where different components would otherwise tend to evaporate at different rates.
NOTE: Sputtering step coverage is more or less conformal. It is widely used in the optical media. The manufacturing of all formats of CD, DVD and BD are done with the help of this technique.
It is a fast technique and also it provides a good thickness control.
PULSED LASER DEPOSITION; this system works by an ablation process. Pulses of focused laser light vaporize the surface of the target material and convert it to plasma. This plasma usually reverts to a gas before it reaches the substrate.
CATHODIC ARC DEPOSITION (ARC-PVD) which is a kind of ion beam deposition where an electrical arc is created that literally blasts ions from the cathode. The arc has an extremely high power density resulting in a high level of ionization (30-100%), multiply charged ions, neutral particles, clusters and macro-particles (droplets). If a reactive gas is introduced during the evaporation process, dissociation, ionization and excitation can occur during interaction with the flux and a compound film will be deposited.
ELECTROHYDRODYNAMIC DEPOSITION (ELECTRO-SPRAY) is a relatively new process of thin film deposition. The liquid to be deposited either in the form of nano-particle solution or simply a solution, is fed to a small capillary nozzle (usually metallic) which is connected to a high power source. The substrate on which the film has to be deposited is connected to the ground terminal of the power source. Through the influence of electrical field, the liquid coming out of the nozzle takes a conical shape (Taylor cone) and at the apex of the cone a thin film jet emanates which disintegrates into very fine and small positively charged droplets under the influence of Rayleigh charge limit. The droplets keep getting smaller and smaller and ultimately get deposited on the substrate as a uniform thin layer.
THIN-FILM PHOTOVOLTALIC CELLS
Thin film solar cells, third generation solar cell, and low cost photovoltaic cell.
Thin film technologies are also being developed as a means of substantially reducing the cost of photovoltaic (PV) systems. The rationale for this is that thin-film modules are cheaper to manufacture owing to their reduced material cost, energy costs. This is especially represented in the use of printed electronics (roll-to-roll) processes. Thin films belong to the second and third generation photovoltaic cell generation.
Thin-film printing technology is being used to apply solid state lithium polymers to a variety of substrates to create unique batteries for specialized applications. Thin-film batteries can be deposited directed onto chips or chip packages in any shape or size. Flexible batteries can be made by printing onto plastic, thin metal foils or papers.
Properties and Characterization of Thin Films
To make sure that coatings which were produced by a given process satisfy the specified technological demands a wide field of characterization, measurement and testing methods is available. The physical properties of a thin film are highly dependent on their thickness. The determination of the film thickness and of the deposition rate therefore is a fundamental task in thin film technology.
In many applications it is necessary to have a good knowledge about the current film thickness even during the deposition process, as e. g. in the case of optical coatings. Therefore one distinguishes between thickness measurement methods which are applied during deposition (“in situ”) and methods by which the thickness can be determined after finishing a coating run (“ex situ”).
These are methods which are based on the determination of a mass. The film
thickness d can be calculated from the mass of the coating m if the density ρ and the area A on which the material is deposited are known:
d = m/ ( Aρ) ………………………………………………………………………..(1.1)
For this method one has to bear in mind that the density of a coating may deviate
significantly from that of the bulk (e. g. due to porosity or implanted interstitial atoms). For exact measurements calibration is necessary.
The simplest method for film thickness determination is most probably the
determination of the mass gain of a coated substrate with an exact balance. Although,
together with the problem of film density mentioned above, also other obstacles exist (e. g. condensation of water vapor from the ambient) it is possible to determine film thickness with sufficient accuracy for several practical applications.
Quartz Oscillator Method
This set-up, which is commonly called “quartz oscillator microbalance (QMB)”, is
generally used for the in-situ determination and control of the film thickness and deposition rate in the case of PVD methods. In commercially available designs film thicknesses in the
range from 0,1 nm – 100 μm and deposition rates in the range from 0,01 – 100 nms-1 are
The method of mass determination was developed in 1959 by Sauerbrey and is
based on the change of the resonance frequency of an oscillating quartz crystal
f = N/ dq ………………………………………………………………………………………….(1.2)
if the crystal is coated by a film with the mass
∆m = ρAd ………………………………………………………………………………………… (1.3)
dq is the thickness of the quartz crystal and N is the spring constant of the crystal, A is the area, ρ the density and d the film thickness in the coated region of the quartz crystal.
The mass amount ∆m acts similar to a thickness change of the quartz crystal by
∆d = ∆m/ρqAq …………………………………………………………………………………. (1.4)
where Aq and ρq are the area and the density of the quartz plate, respectively. In this case the resonance frequency decreases proportionally to d if ∆f<<f is valid. With
For commercially available equipment good linearity is only ensured for mass gains
of 10% relative to the initial mass of the oscillator crystal. Recently developed devices can sustain much higher mass gains because non linearities are mathematically implemented in the algorithm of measurement. Benes (Institute of General Physics, Vienna University of Technology) has developed a method of measuring acoustic impedances (z-Values). By exciting different higher harmonics and by their exact measurement it is possible to determine the z-values which are necessary to experimentally tackle with the determination of high mass additions.
This very exact method is, unfortunately, not suitable for practical applications. It is
therefore mostly used for calibrating other measurement processes. All microbalances used for film thickness measurement basically work by compensating of the coating weight by a counteracting force. The compensation can be accomplished by optical or electrical (turning coil) systems. It is possible to measure the mass thicknesses m/ A as well as the coating rates m& / A
Dosed Mass Supply
Many deposition processes are executed with dosed mass supply, i. e. by using a
mass flow (given in kgs-1) which is held constant due to defined geometry of the plant and due to temporarily unchanging process parameters. This method demands calibration e. g. by weighing or chemical microanalysis to find the connection between the mass thickness m/A at the substrate and the mass flow.
After calibration it is sufficient in most cases to keep the relevant process parameters
constant to obtain the same thickness values in similar deposition times within certain
If the mass supplied per unit time, m, or equally the increase of film thickness per unit
time, d, is known and the deposition rate a is constant:
Because of its simplicity the method of dosed mass supply is widely used: for thermal
spraying, CVD processes, for galvanic and electroless deposition, for build up welding and in many applications based on sputtering and ion plating.
Optical coatings, unlike other applications, require the measurement of the film thickness as exact as possible during deposition. Therefore film thickness monitors are used which, especially in the case of multi coated optics (interference filters etc) are incorporated into closed loop controls by (in some cases quite complex) software components.
This method is mostly used in PVD processes for the production of single layer and multilayer coatings for optical applications. It measures the optical thickness and so that it allows for a compensation of changes in the refractive index n by corresponding changes of the film thickness d.
With a photometer, the intensity of light reflected on both interfaces of the sample or transmitted through the sample is measured. These intensities, related to the initial intensity, determine the reflectivity R and the transmittivity T of the film. The impinging light passes an interference filter which acts as a monocromator and is modulated by a chopper so that disturbances by stray light from the surrounding are avoided. The light ray is directed towards the substrate or a reference glass which is located in a test glass exchange unit. The light intensities are measured by photomultipliers. Computers allow for the automated control of the coating process which is especially important for multilayer coatings.
Although it is possible to calculate the film thickness d from R and T by using given values of the optical constants n and k (i.e. real part and imaginary part of the refractive index, respectively) it is in most cases easier to deduce d from measured curves.
In the case of absorption free or weakly absorbing films R and T change periodically with increasing d due to interference effects if coating and substrate have different refractive indices n.
The basic principle of this multiple beam interferometer is as follows: a thin transparent glass slide is put onto a highly reflecting surface. The glass slide is tilted against the surface by an angle α. If this set-up is illuminated by monochromatic light, multiple beam interferences lead to the formation of equidistant interference lines. The more partial beams interfere, the sharper the lines. Their distance depends on α. A film thickness can be determined by this method as follows: First a scratch is made into the film which reaches down to the substrate (it is also possible to mask a part of the substrate during deposition which leads to the formation of a step). Then the sample is coated by a highly reflective layer. Because of the scratch the distance reflector/interference slide is changed, which leads to an offset of the interference lines relative to each other.
The FECO (Fringes of Equal Chromatic Order) method has an even higher resolution than the Tolansky interferometer. Given a careful measurement a resolution of 0,1 nm can be achieved. Parallel white light is illuminating the combination of sample and reference slide. The reflected light is focused into the entrance slit of a spectrograph via a semitransparent mirror. The image of the step has to be normal to the entrance slit. The obtained spectra (interferograms) are displayed .Dark interference lines are observed at the wavelengths.
Direct methods allow the determination of film thickness either by mechanical
profiling or by observation in microscopes.
The film has to exhibit a step on a plane substrate. A diamond stylus (tip curvature
approx. 10 μm) is pulled along the surface at constant velocity. The step height is measured by a pick-up system. Prerequisite for an exact measurement are a suitable hardness of the film and a plane substrate. To prevent destruction of the sample the load on the tip can be reduced to approx. 10 μN.
The metering range ranges from some 5 nm to some 10 μm at a resolution of some Å in the most sensitive range. An additional application of these commercially available devices is the imaging of surface profiles and the acquisition of roughness values.
Optical and Electron Microscopy for Thickness Measurement
For the determination of the film thickness with the optical microscope e. g. metallographic cross sections are used. In the Transmission Electron Microscope (TEM) replica of a step in the film or cross sectional preparates can be investigated. In the Scanning Electron Microscope (SEM) the step itself or a fracture surface of the film can be imaged.
The resolution and the metric range depends on the instrument and on the magnification. In the optical microscope film thicknesses up to some mm with a resolution of 0.1 μm can be determined. In the SEM a resolution of 5 nm can be achieved and in the High Resolution TEM (HRTEM) 0,1 nm are possible.
Film Thickness Measurement by Electrical or Magnetic Quantities
This method is used in the case of PVD processes for the determination of the film thickness of metallic coatings. The monitoring element is an insulating plate with two parallel line contacts between which a film is deposited through a mask. The resistance R as a measure of film thickness is controlled via a bridge circuit. The deposition rate is determined by electronic differentiation.
With the aid of a zero point indicator the deposition process is stopped if the setpoint of the film thickness is reached. The metric range lies between 1 nm and 10 μm.
Applications are metal films for integrated circuits, resistance films made from NiCr,
metallized foils etc.
In analogy to the previous method the film thickness of insulating coatings can be determined by a monitoring element consisting of comb shaped, interlocking plane electrodes which allow the measurement of the capacity change during deposition.
Eddy Current Method
The thickness of insulating coatings on non ferrous metal or of non ferrous coatings on insulating substrates can be measured by this method. The measurable quantity is e. g. the voltage applied to a RF coil which is modified by eddy currents in the non ferrous metal.
Since this quantity is also dependent on the conductivity of the non ferrous metal a calibration is necessary. The method is mostly applied in polymer metallization.
This method is applied to films which are deposited on a substrate of plane ferritic steel. It is based on the measurement of the adhesive force of a magnet put on the coating (non ferrous metal, lacquer, polymer) which depends on the film thickness. Since this force also depends on the permeability of the steel a calibration is necessary. Also Ni as coating material is accessible to thickness measurement after calibration.
Thickness Measurement by Interaction with Particles
Evaporation Rate Monitor
Film thickness or deposition rate monitors were developed especially for applications based on evaporation technology. To control the vapor density in the vicinity of the
substarate the vapor is ionized at this position by collisions with electrons emitted from a glow filament. The ion current is measured.
More recent devices analyze the ion current by a quadrupole mass spectrometer. By this method the evaporation rates of simultaneously evaporated materials can be determined.
With decreasing thickness the surface structure of a coating gains more and more importance. In the extreme case of ultrathin films the surface roughness may be in the order of the film thickness and can influence all film properties such as mechanical, electrical, magnetical or optical properties. Also film morphology, inner structure, texture and crystallinity are strongly connected to roughness evolution. This section shall briefly discuss basic roughness types, the mechanisms of their origin, roughness measurement and roughness quantification.
Types of Roughness
Generally, the reason for the development of roughness during the deposition process is the finite extension of the film forming particles and their random, temporally and spatially uncorrelated impingement at the growth front. The “building blocks” of the film do not necessarily have to be single atoms as it is the case for PVD coatings. They can also be complex molecules (e. g. for organic coatings) multi-particle aggregates as e. g. for cluster deposition or macroscopic aggregates like the ceramic or metallic droplets in the case of thermal spraying.
The simplest possible model of roughness development is the perpendicular impingement of not nearer specified particles with a finite extension a on random positions of a quadratic lattice at random times on an initially completely flat surface. A particle is added to the material ensemble (from now on called “aggregate”) as soon as it has a nearest neighbor beneath itself. Particles which are members of the aggregate but have no nearest neighbor above form the so-called “active surface” and constitute the growth front of the film. Particles can only be incorporated into the aggregate upon attachment to the acitve surface.
Self Affine Surfaces
Even small modifications of the previously described model of stochastic growth have significant consequences for the detailed shape of the active surface. At first, the SOS character of the growth model shall be conserved. A first approximation to realistic conditions can be implemented if one allows a particle to find a site with as many nearest neighbors as possible within a certain distance from the impingement site. This is a model of particle migration along the surface by surface diffusion and particle attachment to energetically favorable sites with many nearest neighbors, such as steps, kinks or point defects.
These relaxation mechanisms lead to the formation of lateral correlations in the active surface. This means that height values cannot change abruptly within the vicinity of a given point at the surface. Within the zone where a particle may find the energetically most favorable site, which enters the model as a free paremeter one can assume that height values change only gradually.
A big difference to the model of stochastic growth is that the roughness of the surface is not only dependent on the deposition time but also on the length of the interval on which the roughness (by which parameter ever it is defined) is determined.
Pore Formation and non “Solid-on-Solid”-Surfaces
Even the model of stochastic growth can lose its Solid-on-Solid character by a simple modification: If a particle is added to the aggregate when it just has a nearest neighbor, regardless if this neighbor is below or beside the deposited particle, overhangs and pore scan develop.
This growth model is called “ballistic aggregation” and shows lateral correlations and self affinity in the active surface as opposed to stochastic growth. The function which describes the active surface is still single valued as Fig. 4.14. shows. Only if all particles which are connected to the vacuum above the highest point of the aggregate via a continuous path are considered to be path of the surface the surface becomes a multi-valued function and the non SOS character of the model becomes obvious.
Also loosening the constraint of perpendicular particle incidence leads to the formation of long deep pores in the growing film because of the mechanism of shadowing if the particle mobility is sufficiently low. If the film forming particles hit a structured growth front from a wide range of impingement angles the normal growth velocity vn at a peak is higher
than vn in a valley.
Broad distributions of particle incidence are observed for sputtering processes or plasma supported PVD processes where particles have to cross rather large regions of elevated gas pressures and therefore are scattered often. Generally one can say that the higher the gas pressure the broader is the distribution of impingement angles at constant distance between source and substrate. Therefore shadowing is present for sputtering at high gas pressures which is also known from the structure zone diagrams of film growth.
Waviness and defined Structures
Material surfaces and coating surfaces can of course exhibit roughness or, in general, structured surfaces, due to a countless number of other processes. Even an ideal single crystalline surface shows corrugations because of the periodic arrangement of atoms.
Islanding in the first phases of growth may lead to the formation of more or less regular three dimensional features.
Material removing processes such as turning, milling or honing often generate very complex wavy structures which are hard to characterize quantitatively. Finally, surfaces can be willingly structured with very well defined microscopic structures as e. g. in microelectronics. This can be accomplished by photolithographic methods which mostly consist of selective material removal (“Top-Down-methods”). In recent time, however, defined micro and nanostructures are often manufactured by skilledly exploiting basic film growth mechanisms like islanding and stress triggered self organization (“Bottom-Up methods”).
The basic principle of each kind of roughness measurement is, in the case of a 2-dimensional surface, the acquisition of each height value h(x,y). Theoretically the assumption of a SOS surface (i. e. h(x,y) single valued) does not have to be made, but practically all methods yield only the uppermost height values. Regions of the surface which are located beneath overhangs are not imaged. In addition the intrinsic roughness (i. e. the roughness resulting from the deposition process only) is always convoluted with the roughness, tilt or waviness of the substrate. The separation of these two morphologies is not trivial and has to be done by the observer in most cases. Modern measurement systems offer a wide choice of data manipulation features (e. g. corrector of the acquired values for arbitrary tilts or filtering of components with specified wavelengths by Fourier algorithms) but these should be applied only by skilled persons because otherwise they could lead to massively corrupted results. Basically, data can be acquired in real space and in Fourier space.
Scanning Tunneling Microscope: The Scanning Tunneling Microscope (STM) was developed by Binning und Rohrer in 1986 and allowed, for the first time, to image two dimensional surfaces with atomic resolution. The measurement principle is based on the quantum mechanical tunneling effect: if a conducting tip (mostly W) is very closely approached to a conducting surface a tunneling current begins to flow. The magnitude of the current exponentially depends on the distance of the tip to the surface. If the tip is scanned across the surface the surface morphology can be extracted from the variations in the tunneling current. The vertical resolution is within the pm (10-12 m) region, the lateral resolution lies
well within the Å region. The exact vertical and lateral positioning of the tip is made possible by the use of piezoelectric positioning elements (which were also used in the prototype of the instrument).
The STM also provides a good example for the so-called feedback principle which is widely employed in measurement technique.
As previously mentioned the movement of the STM tip over the surface at a height which is kept constant relative to a reference height Href leads to a variation in the tunneling current Itunnel. This principle has several disadvantages: if e. g. the height of the surface exceeds Href the tip will contact the surface which, in most cases, will lead to the destruction or modification of the tip. Also Itunnel is a very small quantity with a low signal to noise ratio. It is therefore more useful to keep Itunnel and therefore the distance tip/surface constant and to measure the piezo voltage UPiezo of the vertical positioning element which is necessary to accomplish this (Fig. 4.19b.). The contact between tip and surface is avoided and the signal gains intensity since typical piezo voltages range between 1 – 1000 V which can easily be measured with high resolution.
The advantages of the STM are its robust set-up and the high imaging accuracy.
Disadvantages are the limitation to conductive surfaces and the fact that the tunneling current is not only dependent on the distance tip/surface but also on the tip geometry, on the chemical constitution of the tip which cannot accurately be controlled and on the chemical constitution of the surface. Nonetheless, especially the last point may also be a big advantage since it allows for a chemical distinction between single atoms on a surface.
Atomic Force Microscope: The Atomic Force Microscope (AFM) is the extension of the STM to insulating surfaces. A micromechanically manufactured cantilever (mostly made of Si or Si3N4) carries a tip with an opening angle of 20 – 50° on its end. The tip curvature is approximately 20 – 50 nm. If the tip is brought into contact with the surface the cantilever is deflected. The deflection is measured and kept constant by using the feedback principle.
From the signal applied to the vertical piezo the surface morphology can be reconstructed.
The most common method to measure the cantilever deflection today is to detect the displacement of a laser beam reflected from the backside of the cantilever by a four quadrant photodiode. This allows the detection of vertical and lateral cantilever deformations.
Another method can be the application of a conductive tip onto the backside of the cantilever which transforms the deflection to a tunneling current signal. This principle is common for AFMs used in Ultra High Vacuum (UHV).
One advantage of the AFM is, as previously mentioned, the independence from the electrical conductivity of the sample. In addition the direct contact between tip and sample surface also yields information about mechanical properties of the surface as e. g. the local coefficient of friction and the local hardness. Unfortunately this direct contact also brings disadvantages: atomic resolution cannot be achieved due to the high tip curvature. The surface is loaded mechanically and there is a high probability of tip contamination. All these points can be (at least partially) suppressed if the instrument is operated in “non-contact” mode. In this mode the free cantilever is excited in its resonance frequency. If the tip is approached to the surface the intermittent contact with the surface leads to a reduction in amplitude and to a phase shift compared to the free oscillation. The changes in amplitude or phase are used as a signal which enters the feedback loop and can again be used to reconstruct the surface morphology.
In recent time many probe techniques have evolved from the principles of STM and AFM. They allow the determination of a great variety of surface properties on different materials. As examples Magnetic Force Microscopy (MFM) and Scanning Near field Optical
Microscopy (SNOM) can be mentioned. MFM is used for the investigation of magnetic surfaces and SNOM allows to perform optical spectroscopy on a molecular level, thus yielding chemical information on the nm scale.
Optical Methods and Scattering
Roughness determination by optical methods is the complementary approach to probe techniques. Here the information about surface morphology is not acquired in real space but in reciprocal space. Generally, if one considers the reflection of an electromagnetic wave (optical or X-ray) on a surface, two components are present: They are displayed in Fig. 4.20. For reflection on a smooth surface the specular component is dominant (Fig. 4.20a), for rough surfaces the diffuse component dominates (Fig. 4.20b).
Fig. 4.20.: Reflection types at surfaces:
(a) specular reflection
(b) diffuse reflection
(c) combination of specular and diffuse reflection; real measurement signal
The change in the intensity of the specular component during film growth can yield important qualitative information about roughness evolution. Quantitative information about surface morphology can be obtained from the diffuse component by determining the fourier coefficients of the surface fourier spectrum. These can be calculated from the spatially resolved intensity of the diffuse component. The shape of a (for simplicities sake) one dimensional surface profile h(x) is given by its fourier series as
As radiation visible light as well as X-rays are used. Especially X-rays allow the determination of the roughness of inner interfaces within a large variety of materials due to their high penetration depth. Also particle beams as e. g. electrons or ions can be used for scattering experiments and yield results similar to electromagnetic radiation due to the wave/particle dualism, but on a different length scale because of their significantly shorter deBroglie wavelength.
Scattering based techniques are non destructive and allow the determination of surface parameters in real time and in situ during a deposition process. Disadvantages are the loss of the phase information and the low reflection coefficients of X-rays which gain reasonable values only in grazing incidence. The low intensity of the diffuse component in the case of Xrays often requires intensive syncrotron light sources which is linked to high apparatve and administrative effort.
The structure of thin films is mainly responsible for their mechanical properties.
Therefore at first the connection between film growth, structure and morphology and the mechanical properties resulting thereof shall be discussed.
Structure Zone Models
For the growth of a film and for the development of its structure three factors are important: the roughness of technical surfaces, the activation energies of surface and volume diffusion of the film forming atoms and finally the binding energy of ad-atoms to the substrate. Substrate roughness leads to shadowing which, in turn, triggers a porous structure. Shadowing can be overcome by surface diffusion at elevated temperatures.
The energies mentioned above are proportional to the melting temperature of many pure metals, Tm [K]. Therefore one can guess that one of the three effects, shadowing, surface diffusion and volume diffusion, is dominant in one given region of T/Tm, i.e. the ratio of the substrate temperature T [K] and the melting temperature Tm [K]. Within this region the dominant effect has the primary influence on the microstructure. This is the basis of the so called structure zone models.
Model of Movchan und Demchishin
Movchan and Demchishin, in 1969, investigated the structure and the properties of films evaporated in HV (10-4 to 10-3Pa) as a function of T/Tm. The materials deposited were Ti, Ni, W, ZrO2 und Al2O3 and the thickness reached values up to 2 mm. Their results yielded the three zone model.
The M.-D. model was expanded in 1974 by Thornton by experiments using a hollow
cathode discharge at Argon pressures between 0,1 and 4 Pa. Thornton added one additional variable, the Ar pressure, to describe the influence of a gas atmosphere (without ion bombardement) on the film structure. Additionally, a transition zone (zone T) was introduced between zone 1 and zone 2 (see Fig. 4.27b). This transition zone is not very pronounced for metals and single phase materials, but can well be observed for refractory compounds and multiphase materials which are produced by evaporation in HV or, in the presence of inert or reactive gases, by sputtering or ion plating. The other zones show equal properties in both models.
T Substrate temperature [K]; Tm Melting temperature [K]; pA Argon pressure
Zone 1 comprises the microstructure forming at low values of T/Tm. Ad atom diffusion is not sufficient to overcome shadowing. Therefore needle shaped crystallites emerge from relatively few nuclei. The crystallites broaden with increasing height by lateral capture of atoms so that inverted cones are formed. Their tips consist of spherical caps. The film is porous and the single crystallites have a distance in the region of some 10 nm. They exhibit a high dislocation density and show compressive stress inside due to the high defect density.
Globally, on the other hand, zone 1 films are subjected to tensile stress because increasing film thickness leads to the coalescence of neighboring crystallites.
Zone T is characterized by the fact that ad atoms can compensate the effects of shadowing because of their mobility due to surface diffusion. Additionally, especially at low working gas pressure, a current of energetic particles is present which increases the density of nuclei by the formation of surface defects. A fiber shaped and, compared to zone 1, much denser structure is formed.
Zone 2 is defined by the region of T/Tm in which surface diffusion is the dominant factor of growth. A columnar microstructure is formed. The column diameter increases with the substrate temperature T while porosity decreases.
Zone 3, finally, comprises the region of T/Tm in which growth is dominated by volume diffusion. A recrystallized dense microstructure of three dimensional, equiaxed grains is formed. This temperature region is important for epitactic growth of semiconductors by evaporation, sputtering and CVD.
Influence of the inert gas on the microstructure
According to Thornton’s model the transition temperatures T1 and T2 are decreasing with falling inert gas pressure pA. This is mainly due to the fact that, as mentioned before, a permanent stream of energetic particles is present. The reason for this is the low number of collisions of the film forming particles with working gas atoms. This, on the one hand, leads to the formation of surface defects which increase the nucleation density. On the other hand their impulse is transferred to loosely bound adsorbents (e. g. single ad atoms) and increases the transient mobility of the adsorbates. Finally, the additional energy input also heats the substrate. All these effects lead to a temperature reduction for the transition from zone 1 to zone T.
Influence of ion bombardment on the microstructure
Ion bombardment generates point defects at the substrate and therefore increases the density of nuclei. The energy transfer to the ad atoms increases their transient mobility.
Therefore, at given T/Tm a denser crystallite structure is generated when compared to the situation without ion bombardment.
This means that ion bombardment influences the film structure in a way that the zone limits, especially the border between zone 1 and zone T are shifted to lower values of T/Tm. This effect, which cannot be explained by additional heating alone, is indeed observed.
All coatings are in a state of more or less pronounced internal stress which is composed of two components, thermal stress σT and intrinsic stress, σi:
σ = σT + σi
The thermal stress σT is caused by different Coefficients of Thermal Expansion (CTS) of the coating and the substrate and is given by:
ES = Elastic Modulus of the film
αs = mean CTE of the film
αu = mean CTE of the substrate
TB = Substrate temperature during coating
TM = Substrate temperature during measurement
Therefore the thermal stress σT can, in principle, be calculated as opposed to the intrinsic stress σi
The intrinsic stress σi depends on the deposition parameters and is caused by structural disorder within the film, i. e. by incorporated foreign atoms and by coating atoms which are located out of a potential minimum. These stresses can be compressive or tensile depending on the deposition parameters, so that the coating tries to contract (tensile stress) or to expand (compressive stress) parallel to the substrate surface.
If low melting point materials are deposited at sufficiently high substrate temperature (T/Tm>0,5) the intrinsic stresses remain low due to the high number of diffusion events while thermal stresses are dominating. Thermal stresses can be reduced by temperature treatment after deposition. The flow of material which is triggered by temperature treatment can lead to the formation of hillocks or voids with sizes in the μm and sub μm region depending on whether the film is set under compression or tension by the temperature change. Hillock formation is observed for Al, Pb or Au coatings which were deposited by PVD methods at elevated substrate temperature.
High melting point materials are generally deposited at rather low T/Tm values (<0,25) so that the intrinsic stresses are dominant relative to the thermal ones. For sufficiently thin coatings (<500nm) the intrinsic stresses can be considered constant throughout the whole film thickness. For evaporated coatings they are mostly tensile, while sputtered coatings mostly show compressive stress. Stresses can approach the yield strength of the bulk material and can reach values of 103Nmm-2 for refractory metals. In some cases intrinsic stresses even exceed the bulk yield strength which indicates certain solidification processes in thin films.
The bonds within the interface between coating and substrate have to withstand the shear forces which are produced by the intrinsic and thermal stresses. Since the contribution of the intrinsic stress to the shear force grows with film thickness the film can spall off the substrate after a certain critical thickness is exceeded. Under adverse conditions this may even happen at thickness values as low as 100 nm.
Intrinsic stresses may be influenced by the process parameters like substrate temperature, deposition rate, impingement angle and energy distribution of the film forming atoms, gas incorporation and residual gas composition in regard to their type, compressive or tensile (s. This was demonstrated for evaporated and sputtered coatings in extensive work. Therefore the possibility to produce mostly stress free and, resulting from this, rather thick films by PVD methods is given.
For all applications of coated materials sufficient adhesion regarding the respective application is paramount to guarantee a reasonable lifetime of the coated part. Adhesion is a macroscopic property which depends on the inteatomic forces within the interface between substrate and film, on the internal stresses and on the specific load the coating is subjected to. The latter can be mechanical (pulling or shearing), thermal (high or low temperatures or thermal cycling), chemical (corrosion, either chemical or electrochemical) and can also result from other effects.
According to Mattox a “good adhesion” is generally reached if:
1. A strong atom-atom bonding exists within the interface zone,
2. Low internal stresses exist within the film,
3. No easy mode of deformation or fracture exists, and
4. No long term degradation is present in the composite substrate/coating.
The adhesion depends primarily on the choice of the partners present in the
composite, on the interface type, on the microstructure (and therefore on the deposition parameters of the coating) and on the pre treatment of the substrate. In the following these dependences shall be discussed with a special focus on PVD coating methods.
Interface between Substrate and Coating
Nucleation and Film Growth
It have been shown there that the aggregation of ad-atoms leads to the formation of nuclei and to the growth of islands which – dependent on the deposition parameters – coalesce to a more or less continuous film.
The nucleation density and the growth of nuclei determine the effective contact area at the interface or, vice versa, the surface which is exposed to voids. Given a low nucleation density the adhesion is low due to the low contact area and the easy fracture propagation through voids and pores. The nucleation density can be enhanced by ion bombardment, surface defects, impurities, the surrounding gas and therefore in general by the choice of a suitable deposition method.
Mattox distinguishes the five interface types described in the following:
The substrate surface is rough and exhibits pores in which the coating material is locked. This yields sufficient, purely mechanical adhesion for many applications. One has to take into account that substrate roughness leads to shadowing and therefore to void formation and a porous structure. On the other hand a crack through a coating which was deposited on a sand blasted surface cannot propagate easily because it has to change its direction frequently or it often has to pass more solid material. Examples for mechanical interlocking are thick coatings which were produced by thermal spraying or which were electrochemically deposited on sintered ceramics or etched polymer.
Monolayer / Monolayer Transition
Here an abrupt transition from the substrate to the film within a few atomic layers is observed. Such a transition is formed if no diffusion or chemical reaction is present between the partners, i. e. if there is only low particle energy available, if there is low solubility or if impurities are present on the substrate. Evaporated coatings which were produced under these conditions exhibit this characteristic feature.
This transition, which forms in the case of a chemical reaction between the substrate and a coating, demands the formation of a chemical bond between the involved partners and is promoted by elevated temperatures. Within a zone which can comprise many atomic layers an intermetallic compound, an oxide or some other compound may form.
Intermetallic compounds are often brittle, prone to crack formation and may trigger fractures within the composite. Oxides may form thermal barriers or good transitions from metal to a ceramic. Examples for the latter case are oxidic interlayers which are formed by CVD or by thermal spraying.
If substrate and coating material are mutually soluble and if suitable temperature is present a continuous transition of the chemical composition, of the lattice parameters and of the stresses from the substrate to the coating is formed by interdiffusion.
For different diffusion rates porosities may be formed at the interface due to the Kirkendall effect if the transition zone is sufficiently thick.
The diffusion transition forms everywhere, where two mutually soluble materials with atomically clean surfaces ace in close contact to each other at elevated temperatures. This is the case for many coating processes as e. g. ion plating, CVD, vacuum plasma spraying or melt dipping. It can also be achieved for evaporation and sputtering if the substrate is treated properly before deposition (Ion Etching) and if it is heated during deposition.
Pseudo Diffusion Transition
This transition is formed between materials without mutual solubility in the presence of strong energy input to the substrate or to the forming film. This can be accomplished by ion bombardment or ion implantation. Energetic ions or neutrals are penetrating the lattice of the substrate to a certain depth depending on their energy and stick there without further diffusion.
Also ion bombardment prior to deposition can increase the solubility and, together with that, the diffusivity in the interface zone due to the formation of a high defect concentration and of internal stresses. Another example is the contact of two molten metals which mix and then are rapidly quenched as it is the case e. g. for plasma build up welding.
Practically the obtained interface zones are often a combination of the different transition types. Regarding adhesion those transitions are best which equally distribute the internal stresses resulting from different CTE, from lattice mismatch and from different solubilities among a sufficiently large volume without generating positions of reduced strength or increased brittleness. These conditions are best met by diffusion joints which can be achieved often by heating of the substrate during and/or after deposition.
Microscopic Bond Types
As mentioned above the adhesion between substrate and film depends on the binding energy between substrate material and coating material and on the structure of the interface zone. Chemical, electrostatic and van der Waals bonds as well as a combination of these bond types may be formed. With a binding energy of 0,5 – 10 eV the chemical bond is the strongest among these. Electrons are transferred between film and substrate atoms and are newly distributed. This requires an intimate contact between the partner atoms. In the case of covalent, ionic or metallic bonding the bond strength depends on the degree of the electron transfer. In the first two cases compounds form which generally are brittle, in the last case mostly ductile alloys are generated.
Van der Waals bonding is based on a polarization interaction which does not require an intimate contact of the partner atoms but is significantly weaker (0,1 – 0,3eV) than the chemical bond and also decays rapidly with distance.
It can be shown that a typical chemical bond (4 eV) should withstand a mechanical stress of about 104 Nmm-2 and a typical Van der Waals bond (0,2 eV) should bear a stress of 500 Nmm-2. Adhesion measurements often yield lower values than expected by these estimates. This has three reasons: first, intrinsic stresses have to be taken into account which are added (according to tonsorial calculus) to the external stress and which may approach the theoretical adhesion value. Second, the strength of the interface layer is determined by the defects and the microstructure and therefore lower than expected for an ideal atomic bond. Finally, a low adhesion value may also be due to the structure of the interface, e. g. because of a low contact area which results from a low nucleation density. In this case the contact area is significantly lower than the geometric interface surface.
Since ion bombardment increases the nucleation density and therefore the contact area to the substrate and zone boundaries within the Thornton model are shifted towards lower T/Tm values the adhesion of coatings which are produced by bias sputtering is higher than the one for evaporated coatings. Even higher adhesion values are found in the case of an interface zone which is generated by diffusion and/or chemical reaction as e. g. in coatings produced by ion plating or CVD.
Substrate Pre Treatment
The purity of the substrate surface has a decisive influence on the microstructure and on film adhesion. The criteria which define a clean, or better, a suitable substrate surface are dependent on the given application of the coating. If gas adsorption or desorption, surface diffusion or similar interface processes shall be investigated a surface is demanded which only is covered by a tiny fraction of a monolayer. Therefore an atomically clean surface is required as it can only be produced by UHV techniques.
If, on the other hand, a steel construction shall be protected against corrosion by a Zn layer, sand blasting is a perfectly adequate pre treatment. Rust and scale are removed and a mechanical transition zone is prepared by roughening.
In the following special cleaning methods shall be discussed which are applied in the case of PVD processes. In this case, cleaning consists of two steps: one happening outside the vacuum chamber and one within the chamber as part of the deposition process.
The vacuum chamber is often positioned in a way that the process chamber is part of a clean room while all other components are located in a service compartment.
In principle a coarse cleaning step of the substrates has to be performed out of the deposition chamber. Often they are heavily contaminated or covered by protective substances (e. g. corrosion protection oil) and have to be cleaned by manual scrubbing, sand or glass blasting, polishing, etching in diluted acids or bases or other methods. Many different procedures for a subsequent cleaning are given in literature for different substrate types. Often a combination of multiple solvents and ultrasonic cleaning is chosen. The excellent cleaning efficiency of ultrasonic treatments is based on a “micro scrubbing” effect caused by the implosion of cavitation bubbles within a liquid.
Glow Discharge Cleaning
After sufficient HV is achieved in an evaporation plant an inert gas (e. g. Ar) with a pressure of approx. 10 Pa is introduced into the chamber at reduced pumping speed. A glow discharge is ignited between an insulated ring or rod shaped Al cathode and the substrate as anode. The substrate holder is usually kept at 2 kV voltage and a current of approx. 1 mAcm-2 is flowing.
If the substrate is an insulator it is charged negatively due to the higher mobility of the electrons compared to the heavy ions. An ion current from the plasma to the substrate is generated and the substrate is cleaned by sputtering. Electrically conductive substrates are hit by electrons drawn from the plasma which triggers the desorption of impurities and heats the substrate. In addition sputterered electrode material is deposited on the substrate and forms nucleation sites.
To clean an insulating substrate prior to deposition by sputtering, an RF voltage has to be supplied to prevent positive charging of the substrate. In the case of coating by ion plating continuous sputter cleaning is an integral part of the deposition process.
DC sputter cleaning (“Sputter Etching”) is the adequate method to clean conductive substrates by sputtering.
Improvement of Adhesion
As mentioned before, bias sputtering and ion plating results in better adhering films than simple evaporation.
Friction and Wear
The coefficient of friction and the wear rate (∆d/∆s = reduction in film thickness ∆d per friction distance ∆s) are mostly measured by a pin on disk device under defined conditions (load on pin and material, constant turning velocity of the disc) in a well controlled ambient (dry nitrogen, water vapor, UHV).
For selected applications also test rigs are used which check the wear resistance of tribological coatings by special test procedures.
Ductility is an important criterion for the amount of plastic deformation a thin film can sustain and serves as a guideline how far apart can be deformed after coating without film failure. With the increasing amount of decorative coatings and hard material films produced by PVD and CVD methods the measurement of ductility gains increasing importance.
The ductility of massive metals is evaluated by the well known tensile test. A measure of ductility is the relative breaking elongation given in percent. Ductile metals have a high, brittle materials a low breaking elongation. It is with
Lzz = length of the sample at the point of fracture
Lo = length of unloaded sample
ЄB = breaking elongation in %
The breaking elongation of thin films ЄB can be determined by a bending test. The film which is deposited onto a ductile metal strip is bent around cylindrical bodies with decreasing diameter and the bending radius R is determined for which the film shows the first cracks.
Under the constraint of d << R
where d is the film thickness. A variation of this technique is the three point bending test where the ductile metal strip is supported in the middle and the ends of the strip are loaded by two edges with increasing force until the first cracks show up. This test can be performed in situ e. g. in a SEM.
Hardness is defined as the mechanical resistance which a body opposes to a penetrating second, harder test body. The measured values depend on the elastic and plastic properties of the material, on the shape and properties of the test body and on the specific set-up. One distinguishes e. g. hardness according to Vickers, Rockwell and Brinell where these three types of hardness measurement show different shapes of the test body.
Micro Hardness Measurement
For the measurement of the micro hardness of thin films one mostly uses the micro hardness measurement according to Vickers or Knoop. The terminus “micro hardness” does not, as one might think, imply small hardness values but hardness measurement procedures producing very small test impressions. A generally accepted (but not standardized) definition of micro hardness means test loads of approx. 0,02 – 2 N, where practically the range from 0,05 N to 5 N is preferred. Utra micro hardness measurements use test loads from 10-2 down to approx. 10-5 N.
It is easily understandable that a hardness measurement for very thin films which are used in optical and electronically applications is not possible with a commercially available micro hardness tester which uses an optical microscope. Because of the low film thickness
(<1 μm) one has to work with very low test loads which leads to impressions with diameters far below the diffraction limit. Bangert and Wagendristl developed a Ultra micro hardness measurement system which can be implemented into a SEM and allows measurements with test loads down to 10-4 N. The state of the art, finally, is the so-called “nano indenter” where also the dependence of the tip displacement on the load can be measured and where the geometry of the impressions is determined by an AFM.
The global economic and technological relevance of micro electronics is out of question. The development, production and investigation of micro electronic devices is also a significant part of thin film technology. Understanding conductivity, finally is the basis for developments which happened during the last few decades, starting from a simple diode to a VLSI circuit or a Gibabit RAM chip.
The electrical conductivity of thin films deviates from the bulk conductivity due to the size effect (film thicknesses < 1 μm) and because of other effects as e. g. grain size or incomplete continuity of the film. The conductivity can be determined by the sheet resistance RN.
Definition and Measurement
The sheet resistance is the resistance which one measures at two electrodes which are located at opposite sides of a film with a square area. RN (which is also often termed RΩ) is independent on the size of the quadratic area because the electric resistance of a cuboid with thickness d, width b, length l and specific resistivity ρ is given by:
The sheet resistance is measured by the four point method. Four point shaped electrodes are put in contact with the film surface. The electrodes may be arranged linear with equal distance from each other or on the edges of a square. For the linear arrangement a current I is generated by the outer electrodes (taken from a constant current source in the mA to μA range) and the voltage drop between the inner electrodes is measured.
Conductivity of Discontinuous Thin Films
If no continuous metallic coating is present then the conductivity of a discontinuous metallic film on an insulating substrate is determined by the mean thickness D.
Fig. 4.43.: Schematic of a discontinuous film
The experimental findings are:
They can be explained by the following mechanisms.
1. Thermal emission of electrons into the vacuum or into the conduction band of the insulating substrate. This mechanism is dominating for d > 10 nm und T > 300 K.
2. Quantum mechanical tunneling: possible via the substrate and via vacuum.
The demands on thin film resistors are various. One of the most important features is the large area of resistance values which has to be covered. It extends from approx. 100MΩ/ to 100MΩ/. A variation of the sheet resistance which comprises this many orders of magnitude cannot be realized by changing the geometrical parameters of the coatings alone.
Different materials with selected specific resistivities have to be used.
For films with relatively low resistance values metals or metal alloys are selected. For higher resistance values so-called cermets (ceramic metal combinations) are used. For extremely high resistance values the classical insulators as e. g. quartz are the natural choice.
Other important demands are a low TCR, a good stability of the electrical properties and cost effective production. Therefore not only the choice of material, but also the selected deposition processes and process parameters are important, because they also influence the coating properties.
NiCr-Thin Film Resistors
For low resistivity values NiCr films are preferred because of their excellent properties in regard to low noise, stability and high durability. The sheet resistance and the TCR depend on the composition of the alloy. The electrical properties are critically influenced by the residual gas pressure and the composition of the background gas in the case of PVD processes.
Cermet Thin Film Resistors
The field of application of thin film resistors made of NiCr extends to about 1000MΩ/.
This limit is given by the fact that for very high resistance values the films are so thin that extreme demands on substrate roughness are imposed. Also the reproducibility of film properties decreases and the TCR increases.
For higher resistance values cermets are chosen. These are mixtures of metals and metal oxides or other chemical compounds. These materials allow for a sheet resistance accompanied by a low TCR. Various material combinations have been investigated and applied.
R&D worldwide is permanently investigating semiconducting materials and semiconductor technologies to comply with emerging demands posed on electronic and optoelectronic devices. This is especially valid for thin film technology without which modern semiconductor industry would be unthinkable.
Si is the most common and best investigated semiconducting material (90% of all devices). The importance of compound semiconductors (GaAs, AlAs, InP) is steadily increasing in respect to electronic and optoelectrinic applications.
Deposition processes for semiconductors range from high purity CVD via Molecular Beam Epitaxy (MBE) to Metallo Organic CVD (MOCVD).
Magnetic Thin Films
Of course also magnetism is a direct consequence of the electronic structure of a material. In connection with thin film technology it only shall be mentioned here that by using PVD methods as e. g. magnetron sputtering the storage density of magnetic storage media could be tremendously increased from 1 MByte/dm2 (3 1/4″ diskette) to some 10 GByte/dm2 (modern hard disc) due to the possibility to deposit very fine grained magnetic coatings.
One of the first technologically relevant applications of thin films was optical coatings.
Because of its huge importance the invention of ant reflex coatings in the first half of the 20th century was indicated to be one of the few fundamental advances in applied optics. The field of interference optics has also considerably widened by the development of new coating technologies, by improvements in instrumentation and by the development of new scientific and technological fields.
For optically active films energy conservation is valid:
T = Transmission
R = Reflection
A = Absorption
S = Scattering
Generally, scattering may be neglected so that it is possible to calculate transmission from reflection and vice versa if a film is not absorbing. Therefore, the following considerations can be limited to calculating one of these quantities. An statement about the reflected and the transmitted part and about the polarization state of a plane wave with known direction of oscillation which hits an interface can be made by the equations which were first derived by A. J. Fresnel.
Anti Reflection Coatings
If an isotropic, absorption free thin coating with a thickness d1 in the region of the wavelength of light and a refractive index n1 is deposited onto a substrate with refractive index n2 and the whole system is is located within a medium with refractive index n0 then multiple reflections will be present at the interfaces n2/n1 and n1/n0.
From the above considerations one can see that it is impossible to obtain full transmission for an extended range of wavelengths for different angles of incidence for a single layer system. Nonetheless it is possible – with the aid of modern calculation-, coating and measurement tools – to manufacture optical systems with high transmission values. This is possible only by employing multilayered systems.
Zeroes in reflection can be achieved for single layers only on substrates with a refractive index higher than the one of the coating. From the available coating materials which satisfy today’s demands in respect to durability towards influences from the ambient and towards mechanical loads MgF2 has the smallest refractive index of n = 1,38. After (4.38) zero reflectivity can be achieved only for a substrate with ns = 1,9 if the phase condition (4.84) is met.
The glasses usually employed in optics have refractive indices ranging from 1,5 – 1,7 which is far below 1,9. Zero reflection can therefore not be achieved by a MgF2 single layer.
With double or, even better, with multilayers one can achieve zero reflection also on substrates with low refractive indices. For their calculation the same methods as for the derivation or reflectivity and transmittivity of single layers can be employed.
The dependence of the reflection of λ/4-double layers on glass (n=1,52)
In real life mostly coating materials with the calculated refractive indices to match the amplitude condition and therefore allow for a zero in reflectivity are unavailable. To obtain an optimum approximation of the real reflectivity to the desired one often coatings with film thicknesses other than λ/4 are used. This allows to tune the width of the region with low reflectivity as well as the intensity of suppression of reflectivity within this region.
In comparison to double layers the suppression of reflectivity within a given region is by far stronger for multilayer coatings. Because of this very small deviations from the amplitude or phase condition are easily visible by a change in the reflected color.
Even higher degrees of transmissivity can be achieved with multilayer systems.
Without coating modern optical systems with up to 100 interfaces between glass and air would be unthinkable. Termini like “fully coated” or “multi coated” can be found on nearly every optical component.
For highly reflecting coatings metal mirrors are the easiest choice. In addition they can be covered by absorption free interference layers which enhance reflectivity and act as protective layer at the same time. The reflectivity of metallic layers is not only dependent on the material but also on the deposition method and on the roughness parameters of the substrate since a smooth surface is prerequisite for high reflectivity.
For mirror layers preferably Al is used since it shows good reflectivity from UV to IR. Upon storage on air a thin oxide layer is formed which reduces reflectivity but enhances chemical stability.
Silver mirrors exhibit a higher reflectivity than Al mirrors, but have the disadvantage that they discolorate after short times due to sulfide formation and lose their reflectivity. Gold and copper have a high IR reflectivity while their optical properties are very wavelength dependent in the visible region.
Dielectric Mirrors – Multilayer Systems
An additional possibility for reflectivity enhancement is given by multilayered systems.
Eqn. (4.84) yields that reflection enhancement is given when a material is deposited on a substrate with a refractive index higher than that of the substrate. The reflectivity increase is higher the higher the refractive index of the coating is when compared to that of the substrate. An additional result of (4.84) is that for a given refractive index reflectivity reaches a maximum if the optical thickness is a multiple of λ/2.
An amplification of this effect can be reached if λ/4 layers with alternating low and high indices of refraction are deposited. Maximum reflection is achieved if the number of films is odd and in the beginning and the end of the stack a film with the higher refractive index is located.
The application of highly reflecting coating systems are laser mirrors on the one hand and so-called cold light mirrors on the other hand. Laser mirrors with 19 or more films of high quality show losses below 0.01%. Cold light mirrors show high reflectivity in the visible region and high transmittivity in the IR region. So a high light yield at low warming is achieved (e. g. for projectors). Also Infrared optics and X-ray mirrors use multilayers as reflecting elements since the reflectivity of single layers is too low in this spectral region.
In the simplest case a Fabry Perot filter consists of two partially reflecting metal layers (mostly Ag) between which a spacer layer made from an absorption free material is located. The optical thickness of the spacer layer determines the location of the transmission band.
is transmitted, where nd is the optical film thickness. k is the so-called order of the Fabry perot filter.
The finite absorption of the metal layers leads to disadvantages as e. g. side maxima or difficulties regarding production in respect to the position of the transmission region.
These disadvantages can be avoided by using dielectric multilayer systems. The mode of operation is the same, only that the metallic reflectors are replaced by highly reflecting multilayers. A reflection enhancing film stack consisting of an odd number of layers and a layer with a high refractive index at the bottom and the top of the stack is put on each side
of the spacer layer (see Fig. 4.55.)
By properly choosing the parameters it is possible to tune the width, shape and position of the transmission area as well as the transmittivity in a wide range.
Transmissivities of more than 90% and half widths of 0,05 nm can be achieved (see Fig.4.56). If one uses this kind of filters one has to take into account that, for unprotected systems, the peak is slightly shifted due to humidity and/or temperature.
Edge filters are optical coating systems which have an excellent transmissivity in one spectral region and an excellent reflectivity in an adjacent spectral region. The transition region between these two areas is very narrow (see Fig. 4.57.). Examples are temperature protection filters and blocking or transmission filters for certain spectral ranges (IR, VIS, UV).
By a proper choice of the coating materials, of their respective thicknesses and of their sequence it is possible to realize nearly every desired filter design in respect to bandwidth, edge position and transmission.
Characterization of Optical Coatings
The reproducible production of thin films for optics is much easier to achieve today due to the available measurement methods, the advanced deposition plants and the higher degree of automatization.
Nonetheless the control of the final products is necessary even if it is only by random inspection. For optical coatings sticking to the pre defined optical values is often not enough because also selected mechanical properties as e. g. hardness, adhesion or chemical stability have to be given for a coating subjected to application.
The measurement methods for film thickness, mechanical properties and chemical composition were treated in detail in the proper chapters. The measurement of optical constatnts is mostly done with an ellipsometer.
There are many possible set-ups for an ellipsometer, but only the RAE (Rotating Analyser Ellipsometer) set up is of practical relevance (see Fig.4.58.).
Fig. 4.58.: RAE-ellipsometer: a light source; b polarisator; c substrate; d analysator; e photo detector
Typical values for the angle between a and e are 130 to 140°. If one designates the angle between polarisator and the plane of impingement with P and the corresponding angle of the analysator with A then the signal intensity of the photo detector for the ellipsometric values ψ and ∆ is given by;
The big advantages of optical methods can fully been exploited if one chooses a direct measurement on the substrate with the real film system rather than an indirect measurement on a test sample.
Conventional methods of chemical analysis as atomic emission, atomic absorption spectral analysis, X-ray fluorescence and mass spectrometry play an important role for the production of the coating materials used in several processes. In most cases samples of some milligram are necessary. Given this amount the detection limit may well be in the ppm region.
On the other hand the increased demands to micro analysis of thin films in respect to spatial and areal resolution as well as information depth have led to the development of many new methods. They are mostly based on the interaction of photons, electrons, ions or other particles with the coating or surface to be analyzed. Depending on the exciatation and on the type of analyzed material particles are either scattered elastically or in elastically or secondary particles are emitted. These particles are analyzed by a suitable detector system.
Electron Probe Microanalysis (EPM)
Electron Probe Microanalysis (EPM), also called X-ray micro analysis, is the eldest among these methods. The excitation is performed by electrons and X-ray quanta are detected. The method is suitable for analyzing films of at least 1 μm thickness. Two basic set-ups are existing: the “micro probe” with wavelength dispersive crystal spectrometer (WDX) which often is combined with a TEM. In the SEM mostly energy dispersive Si(Li) semiconductor spectrometers are used. The method os then called Energy Dispersive X-ray analysis, EDX.
These methods fail for light elements (z < 8) and for films which are thinner than the information depth given by the penetration depth of the electron beam (approx. 1 μm).
Auger Electron Spectroscopy (AES)
The sample is probed by electrons with energies between 1 and 10 kV. By this bombardment electrons are removed from inner shells which are refilled by electrons transiting from higher energy levels (see Fig. 4.61.). The energy released during this process can either be transferred to a X-ray quantum which is emitted with the energy hvF (X-ray fluorescence) or can be transferred to a third electron within the atom (Auger effect) which also is released with a characteristic energy.
The information depth is given by the escape depth of the Auger electrons and amounts to approx. 1 – 10 nm. For analysis the number of Auger electrons emitted per unit time is measured as a function of their kinetic energy (determined e. g. by a cylinder mirror energy analyzer) by a secondary electron multiplier. The maxima of the so gained AES spectrum are tabulated and are characteristic for the type and concentration of the atoms within the sample. To suppress the background resulting from scattered electrons the spectrum can be differentiated by electronic means.
The atomic number of detectable atoms is Z > 3. The detection limit is about approx. 0.1%. Depth profiles can be measured by removing material by sputtering, but this method is obviously destructive.
Fig. 4.61.: Energy level diagram for Auger-electron (AES)- and photoelectron (ESCA)- spectroscopy. Q: work function of electrons
Photo Electron Spectroscopy (UPS, XPS)
For excitation either X-rays (then the method is called XPS) or UV light (then the method is called UPS) can be used. For XPS e. g. Aluminum-K-radiation with an energy of
hv1 = 1486 eV (see Fig. 4.61.) is used. The emitted photoelectron therefore carries a kinetic energy of Ekin = hv1 – Ek- F which is hv1 reduced by the binding energy characteristic for the element, Ek, and the work finction F. Similar to AES analysis the number of electrons emitted per unit time is measured in dependence on Ekin.
The atomic number of detectable elements is also Z > 3, the information depth is 0,5-10 nm and the detection limit is 0,1%. An advantage of ESCA is that the analytical information contains evidence about the binding state of the atoms which can be obtained from the line shift (chemical shift) which is dependent on the type of bond (e. g. metallic or oxidic).
Secondary Ion Mass Spectroscopy (SIMS)
The sample is bombarded with positively charged ions, mostly Cs (for the detection of electronegative elements) or O2 (for the detection of electropositive elements) of an energy of approx. 10 keV (see Fig. 4.62.). Also Ga, Au or Bi ions may be used. They are emitted from low intensity sources and therefore allow for vertically and laterally highly resolved
SIMS. Between the impinging ions and the target atoms collision processes occur which lead to the formation of collision cascades within the target. Some of them reach the surface and lead to the emission of single atoms or of molecular fragments (sputtering effect).
From the emitted particles some are positively or negatively charged and can be detected e. g. within a quadrupole mass spectrometer. Newer devices are also using Time of Flight mass spectrometry (TOF-SIMS) which can detect even very low ion currents. The secondary ion mass spectrum is characteristic for the distribution of elements on the sample surface and also (due to the molecular fragments) for their binding state. The mean escape depth is some nm. The SIMS spectrum has no background. Therefore the detection limit is very low (fractions of particle number down to 10-4%). All elements (even H) and their isotopes can be detected.
While AES and ESCA are non destructive (if the sample is not subjected to additional sputtering) the SIMS method is destructive, but this has the intrinsic advantage that depth profiles of the atomic concentration can be measured.
Secondary Neutral Mass Spectroscopy (SNMS)
In the case of SIMS the processes of ionization and emission of the particles to be analyzed are coupled (“matrix effect”). Therefore the intensities of the single lines of the mass spectrum are not representative for the coposition of the investigated surface. To obtain a quantitative analysis a calibration using samples of known composition is necessary. To increase the accuracy of measurement, especially in the case of sophisticated multilayer structures and diffusion or implantation profiles, it is preferable to separate ionization and emission events.
This is the case for the SNMS method where the neutral particles generated due to ion bombardment, which are the main species emitted for sputtering, are post ionized by suitable means and analyzed in a quadrupole mass spectrometer. The atomic concentration of a given mass is then a function of the sputter yield and the ionization probability for a given material in a given set-up. The sputtered particles can be ionized by a RF low pressure plasma (Ar) or by laser radiation.
Ion Scattering Spectroscopy (ISS)
This method allows the analysis of the uppermost atomic layer of a sample whose surface is probed by ions with known energy E0 (some 100 to 103 eV) and mass m0. The primary ions elastically scattered under a given angle (e. g. Ǿ= 90°) are analyzed in respect to their energy E in a detector. The obtained ISS spectrum exhibits maxima at values at E/E0 which correspond to mass ratios m/m0 according to the theory of elastic scattering.
Rutherford Backscatter Spectroscopy (RBS)
RBS analysis is a high energy version of ISS (see Fig. 4.63). It is executed by
Hydrogen, Helium or ions of other light elements within the energy region between 0,1 to 5 MeV. Energy spectra of the backscattered ions are obtained by variation of the energy E0 of the primary ions. From these spectra the depth profiles of the atomic concentrations of all light elements within depths of 0,1 – 10 μm can be determined non-destructively. A big advantage of the RBS method which exhibits a detection limit of approx. 10-3% is that the method is absolute while the other methods mentioned above require calibration by standard samples.
As displayed in Fig. 4.63., high energy ions may be used to realize two other important analysis methods of solid state physics: the method of Ion Induced X-ray radiation
(IIX) and the method of nuclear reactions triggerd by ions (Nuclear Reaction Analysis, NRA)
APPLICATIONS OF THIN FILMS
The kick-off issues will be devoted to descriptions of the multitude of coating applications and the discussions of processes and materials used to deposit specific thin-film optical coatings.
Optical coatings have a ubiquitous presence and serve many roles in our daily lives. Recognition and appreciation of the presence of coated surfaces is a part of the reader’s experience. As an example, ophthalmic Anti-Reflection coatings are commonplace, and all instruments and optical aids associated with vision employ AR coatings to reduce reflection and increase transmitted contrast. Figure 1 shows the performance for a typical four-layer Titania and Silica AR design.
Ophthalmic AR coatings on relatively soft polymer lenses must possess relatively severe durability robustness. They are tested to salt water immersion, transition between boiling and ice water, and must resist moderate abrasion that simulates cleaning with minimum degradation. Eyeglass lenses are often over coated with organ silicones to enhance abrasion and scratch resistance and to provide hydrophobic properties.
Display screens, camera lenses and automobile instrument panels also are AR coated to reduce ambient reflections. Outside the car, the headlight reflecting surfaces are coated with protected aluminum, and the windscreen with a coating that reflects heat to reduce energy consumption for cooling and heating while transmitting visible light. The coating material is
Indium Tin Oxide (ITO), one of several transparent conducting oxides (TCO).
Coatings with the same function are used in huge volume (many square kilometers) for architectural and residential windows. Low emissivity and high transparency in these thermal control coatings are provided by a multi-layer design based on a very thin layer of silver immersed in dielectrics. The multi-layer is deposited by sputtering and reflects energy longer that wavelength 1000 nm while transmitting visible light.
The medical treatment and research fields are major users of optical coatings in their diagnostic, analysis and treatment instruments. Imaging, spectroscopic, and reactant assay devices are some examples. In the case of non-disposable endoscopic imaging instruments, the lens coatings must have low reflection under immersed environments, and must survive sterilization without degradation of the coatings. Laser surgery is commonplace for particular treatment. Coatings that tolerate high energy and long-term repetitive exposure to lasers of various energies are required. Coatings are the weak link in optics that convey lasers; therefore special processes have been developed since the birth of laser sources. Eye surgery, skin resurfacing, tumor destruction and even tattoo erasure are typical examples of the use of laser light. Those applications involve different wavelengths, fluencies and exposure times. All components require high reliability; the stability of coated surfaces is at the top of the list. In non-optical areas, wear resistant coatings that are compatible with tissue and bone interfaces have been developed. We shall discuss coatings and materials related to these applications in greater depth at a later date.
In the entertainment industry, CD/DVD recording and playing devices use sophisticated optical systems that intensively use AR and polarizing coatings. The optical systems in image projectors use beam dividing and polarizing coatings, and the lenses require wide-band, wide-angle AR coatings on their different optical glasses. New coating challenges introduced by the production of 3-D imagery are currently being addressed.
Surfaces that are exposed to external potentially damaging environments require a different set of coating materials and modified deposition processes compared with surfaces that are internally protected. Manufacturing and product sorting industries use coated optics in machine vision systems. Often the manufacturing environment introduces contamination, and the coated surfaces of lenses must survive frequent cleaning. Here again the challenge is to provide coatings that tolerate moderate abrasive and solvent cleaning without deterioration such as haziness, scratches, or stress cracking.
The aeronautic environment, commercial and military, is especially harsh for coated windows and imaging optics. Optical coatings are subjected to rain impact, sand erosion, temperature swings, and aggressive cleaning maintenance.
Deployment in a marine environment additionally requires resistance to exposure to salt spray exposure and cleaning. Coatings specialized for glass surfaces and for softer materials used at infrared wavelengths require special considerations. We shall discuss the specific materials and their deposition processes in future notes.
Optics for military applications probably exemplify the most challenging coating requirements. Surfaces of windows and imaging optics must maintain their properties during and after exposure to harsh environments during field and sea deployment. They must be durable to cleaning in non-ideal conditions. Often the same optics must pass a laser beam as well as transmit Visible / Near-IR and far-IR imagery. Coating designs are complicated and consist of dozens of layers.
Absorption and stress-related aspects must be finely controlled, as must deposition thickness accuracy.
Thin films are very thin layers of materials of desired properties deposited on surfaces called substrates. There are many techniques of depositing thin films which we mentioned above. The techniques to be used depends on the materials to be deposited. We want to use solution growth technique because copper II oxide (CU2O) which materials we want to work with is insoluble in water and also being a compound will not be successfully and chiefly deposition by other techniques.
Thin films are applicable in so many areas like sensor, shielding of materials, electronics, high speed detectors etc Solid films are used to protect corrosives metals like iron from rusting. Tin oxide (SnO2) and other desired electrical properties can be used to fabricate components like resistors and transistors Thin films technology has extra advantage, because of the high absorption coefficient and P-type semiconductor of Tin oxide ( with a band gap of range (1.4-3.2) eV are suitable for the use of solar cells.
Before you can use a thin film, you need to know the properties of the material of the thin film, this is the purpose of this research .The most important properties of thin films include optical and electrical properties, like absorbance, reflectance, transmittance, band gap, electrical and conductivity. These properties can be found mainly by crystallography and spectroscopy with appropriate experiments.
The aim of the present work is to obtain the Tin oxide (SnO2) thin film within a shorter time by varying the concentration of complexing agents such as Ethylenediaminetetraacetic Acid and Triethanolamine, as well as by varying deposition temperature of Tin oxide (SnO2) thin films which we will be studying.