Tài liệu Cảm biến trong sản xuất P6 ppt

are currently used to examine and assess the microgeometry of thestructures to be measured which cannot be resolved by light-optical methods, as asupplement to optical measuring methods.

3 Sensors for Workpieces

98

3.1.7

Further Reading

1 Adam, W., Busch, M., Nickolay, B.,

Senso-ren für die Produktionstechnik; Berlin:

Springer, 1997.

2 Deutsche Gesellschaft für

Zerstö-rungsfreie Prüfung, Handbuch OF 1:

Verfahren für die Optische Formerfassung;

Ei-genverlag, 1995.

3 Dutschke, W.,Fertigungsmeßtechnik;

Stutt-gart: Teubner, 1993.

4 Ernst, A.,Digitale Längen- und

Winkelmess-technik; Landsberg/Lech: Verlag Moderne

A Weckenmann, Universität Erlangen-Nürnberg, Erlangen, Germany

Precision measurement of structures in the micrometer and sub-micrometerranges is becoming more and more important Because of the never-ending min-iaturization it is central to the precision of production and metrology of microelec-tronics and micromechanics, but also to the measurement of the size distribution

of microparticles, for example, in environmental protection A number of ing methods are available to perform these tasks They range from conventionaloptical microscopy and its extension into the ultraviolet range, through electronmicroscopy, to the high-resolution near-field microscopy methods such as atomicforce microscopy

measur-Optical microscopy includes conventional bright- and dark-field microscopy,confocal scanning microscopy, in the visible and ultraviolet spectral ranges, andinterference microscopy As a non-microscopic additional feature, far-field diffrac-tion images of the objects are evaluated Fundamental research into the interac-tion of the radiation used with the objects and theoretical modeling are impor-tant, additional aids in using these methods Non-optical, high-resolution micro-scopy methods (scanning electron microscopy (SEM), atomic force microscopy(AFM), etc.) are currently used to examine and assess the microgeometry of thestructures to be measured which cannot be resolved by light-optical methods, as asupplement to optical measuring methods After further extensive research intothe interaction of the scanning probes with the object structures and specific ex-tension of microscope systems, eg, adding precision length measurement sys-tems, high-resolution microscopy methods can also be used for calibration So farthis has not been possible because the principle of optical methods places a limit

on the resolution that can be achieved

Copyright © 2001 Wiley-VCH Verlag GmbH ISBNs: 3-527-29558-5 (Hardcover); 3-527-60002-7 (Electronic)

Tactile Measuring Method

Tactile measuring methods for surface measurement are still the most importantmethods, especially in the area of metal-cutting and non-cutting machining opera-tions in industry and research It is the only operation that is anchored in na-tional and international standards Particularly the parameters and measurementconditions are fixed, so that the comparability of the measurement results can besecured The surface roughness and topography greatly affect the mechanical andphysical properties of parts Properties such as fit, seal, friction, wear, fatigue, ad-hesion of coatings, electrical and thermal contact, and even optical propertiessuch as gloss, transparency, etc., can be adjusted by manufacturing design Thesurface laboratory is concerned with the assessment of roughness, waviness, tex-ture, groove depth, and other special surface shapes The contact stylus method isgenerally set-up off-line in the measuring room or in the workshop Only in spe-cial cases are oil-proof calipers integrated into the processing equipment The pro-file method is based on the linear sampling of the workpiece surface with a dia-mond needle whose tip has the shape of a cone or a pyramid (Figure 3.2-1) Theradius of the tip is 2 and 10lm and its angle usually 908

The static measuring force applied is less than 1 mN Thereby, equidistant file supporting points are measured directly to calculate various roughness andwaviness characteristics The commencement of this method dates back to about

pro-1930 Nowadays, measurement systems with digital signal processing and profileevaluation are available The instruments can be adjusted to fit the workpiece flex-ibly by modularly compiling the stylus instrument, feed mechanism, and evalua-tion system Contact stylus instruments generally register a two-dimensional verti-cal profile cut in the workpiece surface Latterly, its application has expanded by

Fig 3.2-1 Probe

tip (courtesy: PTB)

introducing a successive cross traverse for the three-dimensional measurement ofsurface topography.

The amplitude resolution can be as good as 10 nm at any measurement point,and the best possible local resolution in the horizontal axis is 0.25lm The mea-suring range for contour measurements extends to 120 mm along the plane ofthe face and 6 mm in amplitude The contact stylus instrument is traceable to theunit meter through reference standards

Alignment of the cantilever is problematic Additionally, the measuring ment is sensitive to vibrations and oscillations A further problem in some cases

instru-is a curved form of the surface of the workpiece

For the adaptation of different workpiece geometries, a variety of different tile profile meters exist, whose properties clearly determine the quality of the sur-face measurement They can generally be traced back to the basic reference sur-face, skidded and double skidded system

tac-3.2.1.1 Reference Surface Tactile Probing System

In the skidless system (Figure 3.2-2), the stylus is located at the end of a probingarm that is guided over the surface of the object to be measured held in a linearguide in the vertical direction The styli are rigidly connected with a referenceplane that is usually located in the feed mechanism The excursions of the styluscaused by the surface roughness are transmitted to a measuring transducer andconverted to measuring signals, depending on the type of transducer, in analog ordigital format The measuring pick-up and the object to be measured are me-chanically decoupled, and only the stylus itself slides over the surface of the objectbeing measured For that reason, skidless systems are extremely sensitive to vibra-tions

3.2.1.2 Skidded System

The skidded system (Figure 3.2-3) uses the surface to be measured as a guide andhas much smaller dimensions than the skidless system The stylus contacts thesurface to be measured with a skid and acquires the surface profile relative to thepath of the skid with the probe tip Depending on the measurement task, the

Fig 3.2-2 Skidless system

landing skid is mounted before, behind, or lateral to the probe tip The able distance between the landing skid and the probe tip can lead to falsificationsduring the transfer of the profile, depending on the surface attributes of the work-piece to be measured However, it is less precise because of the mechanical filter-ing that occurs while sliding over the surface The skids act as an amplitude-inde-pendent, non-linear, high-pass filter and eliminates, depending on the probe andworkpiece geometry, the macro-geometric form and waviness of the workpieceprofile This system is used for fast measurements in production.

unavoid-3.2.1.3 Double Skidded System

The double skidded system (Figure 3.2-4) uses the surface under test as a ence, it is self-aligning, insensitive to vibrations, and requires large measuringsurfaces because of its size

refer-The double skidded system can lead to considerable profile falsification owing

to its landing skid, especially with profile tips that jut out

3.2.2

Optical Measuring Methods

Optical 3D measuring methods permit fast, wide-area sampling point acquisition

In several measurements from different views, it is possible to measure all ing zones and zones of the workpiece relevant to determining the form and sur-face characteristics of the workpiece with the required resolution After transfor-mation of the measured data into a common coordinate system, the sample is re-presented by a 3D set of sampling points From the measured data it is possible

wear-to determine the form, surface, or wear characteristics The advantages of this

Fig 3.2-3 Skidded system

Fig 3.2-4 Double

skidded system

method are that the measuring process can be automated to a great extent and istherefore independent of the influences of the operator, it has a high measuringrate, and the surface of the measured object is acquired as a whole Especiallysuitable for measured value acquisition for microgeometry are devices that oper-ate on the principle of white-light interferometry or scattered light methods.

3.2.2.1 White Light Interferometry

Special white light interferometers permit wide-area form acquisition on opticallyrough surfaces A measuring system called coherence radar is based on the princi-ple of the Michelson interferometer, where the mirror in the measuring beam isreplaced by the object to be measured The light-emitting diode (LED) to be used

as the light source causes white light interference that displays a typical tion as a function of the phase shift between the measuring and reference beam,which is at a maximum when no phase shift exists, ie, the object being measured

modula-is in the reference plane (Figure 3.2-5)

Using a linear table, the object to be measured is pushed through the referenceplane and the position of the linear table is stored as a vertical coordinate for eachsampling point as soon as the maximum modulation of the interference signal ac-quired with a charge-coupled device (CCD) camera is detected One advantage ofthis measuring method is that, unlike, for example, the triangulation method, illu-mination and observation are in the same direction, which makes measurementspossible on structures with a large aspect ratio that are often encountered in mi-crosystem technology Both the topography of the measured object and, derivedfrom it, the roughness of its surface can be measured

Fig 3.2-5 Principle of a white light interferometer

Depending on the optical arrangement, measuring fields from about 50´50

mm to about 200´200 lm can be implemented The lateral resolution depends

on the measuring field and the number of columns or rows of the CCD camera(typically 512´512 pixels) and the pixel geometry; the resolution in the longitudi-nal direction is limited by the roughness of the workpiece surface and the travers-ing speed of the linear table (typically 1–2lm at 4 lm/s) The measuring time is

in the minutes range, depending on the maximum structure depth to be sured

mea-3.2.2.2 Scattered Light Method

The scattered light method is used to measure the roughness of workpiece faces Light reflected from the workpiece has a spatial distribution that depends

sur-on the surface roughness Smooth surfaces reflect incident light fully according tothe law of reflection of geometric optics (angle of reflection with respect to thesurface normal equal to the angle of incidence) On rough surfaces, portions ofthe scattered light are also reflected in other directions Figure 3.2-6 shows the ar-rangement principle of a scattered light sensor The collimated light of an LED isdeflected on to the workpiece surface via a beam divider The diameter of themeasuring spot is about 1 mm The scattered light is mapped with a lens on to alinear image sensor (photodiode or CCD line) so that the intensity of the lightscattered in different directions can be measured at different locations on the de-

tector To assess the surface, the scatter value SNis usually used This is tional to the second statistical moment of the measured intensity distribution and

propor-therefore describes its width Larger SN values indicate a greater proportion ofscattered light, usually describing a rougher surface One problem with the accep-

tance of this measuring method is that the measured SNvalue does not correlate

Fig 3.2-6 Block diagram

of a scattered light sensor

with the roughness quantities Raand Rz which have been introduced into tactileroughness metrology and which are standardized.

Angular speckle correlation (ASK) Angular speckle correlation offers two

advan-tages On the one hand, the requirements for the laser system are small, sinceonly one wavelength is necessary for the implementation of the measurement Onthe other hand, owing to the difference in the angles of illumination, the mea-surement area can continuously be re-adjusted according to the measurementtask The disadvantages of an adjustable difference angle result in high require-ments for the mechanical precision Figure 3.2-7 displays a typical experimentalset-up for ASK measurements with an adjustable difference angle One of the illu-mination beams is faded out when taking the first picture, and for the second pic-ture the other is faded out It is necessary to move one of the pictures respective

to the deviating illumination angle of the applied ASK, so that the offset oppositethe second picture can be counter-balanced The distance moved can roughly be

Fig 3.2-7 Setup of an angular speckle correlation

calculated from the geometry of the setup and is always the same for a fixed

set-up The exact value can be calculated in an evaluation program

Spectral speckle correlation (SSK) The setup of SSK is simplified to the extent that no

second illumination beam path is necessary The adjustment of the measurementsystem during practical application is far easier and one can achieve an increase

in stability With the possibility of taking two pictures of the surface at the sametime, the measurement time can be reduced The disadvantage of this measurementsystem is the higher requirements for the laser system At least two different wave-lengths must be generated, so that an adaptation of the different roughness areasbecomes possible A larger number of wavelengths is, however, more advantageous.The evaluation of the two pictures taken takes place via a two-dimensionalcross-correlation coefficient Experimental prerequisites for the correct evaluationconsist in the observance of Shannon’s theorem This means that the spatial sam-pling frequency, in this case the reciprocal pixel size of the CCD camera, has to

be at least twice as large as the spatial signal frequency In other words,

dspeckleˆp4
2xkf

wherek is the wavelength of the light used, f the focal length of the lens and x0

the diameter of the illuminated area on the surface

3.2.2.4 Grazing Incidence X-Ray Reflectometry

The total reflection of X-rays from solid samples with flat and smooth surfaceswas first reported by Compton in 1923, which can be assumed to mark the birth

of the experimental technique of X-ray specular reflectivity Since the angle of dence is very shallow and almost parallel to the surface, measurement using X-raytotal reflection is also called the grazing incidence experiment If the surface isnot ideally smooth but somewhat rough, the X-rays can be diffusely scattered inany direction The experimental technique is known as X-ray diffuse scattering (X-ray non-specular reflection) Its development began immediately after the pioneer-ing work in 1963 of Yoneda, who reported intensity modulation in X-ray diffusescattering, known as Yoneda wings or anomalous reflection

inci-Nowadays, X-ray reflectometry based on total reflection has become a powerfultool for the analysis of surfaces and thin-film interfaces, and will continue withfurther progress This is mainly due to the significant development of experimen-tal techniques and instrumentation, especially the advent of synchrotron radiationand the progress achieved in detector technology The advances in theoreticalmodeling and techniques for analyzing experimental data are also important

Total reflection and the penetration capability of energy-rich X-rays are used forcoating thickness measurement The refractive index for X-rays is always < 1 If

the angle of incidence is made smaller, the X-radiation penetrates only up to avery small angle, the critical angle If the angle of incidence is reduced stillfurther, external total reflection on the interface occurs The beam is reflected as

by a mirror In coating-substrate systems, part of the radiation is reflected andpart of it penetrates the film There are now two angles of total reflection at theair-coating and coating-substrate interfaces The two partial beams interfere andform interference Surface roughness and the optical densities of coating and sub-strate material affect the acuity of the resulting interference image The most in-tense and sharpest interference images are obtained if the refractive index of thesubstrate material is less than the refractive index of the coating material

The main limitations of the X-ray reflectivity technique are the limited range ofthe wave-vector transfer and the loss of the phase of the reflected amplitude.Nevertheless, an accuracy of approximately 0.2 nm has been reported in determin-ing the thickness and roughness of a double-layer sample

3.2.3

Probe Measuring Methods

Over the last decade, fundamental research into surface physics has given rise to

a new class of analyzer, the scanning probe microscope These devices allow themapping of a surface in a lateral range of 150´150 lm down to atomic resolutionaccording to similar measuring principles with slight technical variations Fig-ure 3.2-8 shows the principle of the structure of a scanning probe microscope.Other members of this class are the magnetic force microscope, the opticalnear-field microscope and microscopes that work by a thermal or capacitive inter-face or with ion flows

However, scanning probe microscopes are not only useful for characterizingsurfaces with high spatial resolution The sharp tips of the scanning tunneling,

Fig 3.2-8 Schematic of scanning probe microscopy (SPM)

scanning force, and lateral force microscopes can also be used as local sensorsand as nano-tools for carrying out experiments or for making surface modifica-tions on the atomic scale In this way, time-stable atomic-scale structures can begenerated, modified, and removed under environmental conditions Chemical re-actions can be induced locally with the AFM tip and crystal growth can be moni-tored in situ and in real time Forces and interactions can be investigated on the(sub)atomic scale and the phenomenon of energy dissipation due to friction can

be studied quantitatively on a microscopic scale

3.2.3.1 Scanning Electron Microscopy (SEM)

In many areas of research it is important to obtain chemical, morphological mation in the sub-micrometer range Because of the limited resolution of opticalmicroscopes (theoretically 0.15lm), bundled electrons accelerated by electricalhigh voltage (up to 3 MV) in a high vacuum are used instead of light becausethey are strongly deflected by scattering at atmospheric pressure Rotationallysymmetric electrical and magnetic fields perform the same functions as lenses in

infor-an optical microscope, concentrating the electron beam coming from the hot ode on to the object The object to be measured is penetrated by the electrons todifferent degrees in the transmission electron microscope depending on the thick-ness and density of the electrons in such a way that the corresponding intensitydistribution in the electron image represents the structure The electron image isacquired on a photographic plate or fluorescent screen, yielding an approximately

cath-200 000-fold magnification In SEM (Figure 3.2-9), an electron beam (diameterabout 10 nm) is moved over the object in a scanning pattern, ie, row by row Theelectrons, both those scattered back and the secondary electrons that escape fromthe surface of the sample, are amplified by the scintillator and photomultiplierand provide the signal for brightness control of a synchronously controlled cath-ode-ray tube (large depth of field)

The resolution limit is determined by the diffraction phenomena at the ture of the imaging system and the wavelength of the particles With a 100 kVelectron microscope, a resolution of 0.2 nm (k=3.7 pm, A=0.4–0.8, error of lens

aper-aperture Cs= 0.3–1 mm) is achieved according to the equation

dtheor:ˆ A
4 kCs

p

3.2.3.2 Scanning Tunneling Microscopy (STM)

In scanning tunneling microscopy, a conductive, atomically sharp needle isguided row by row over a conductive surface If the probe is lowered near the sur-face to be measured, interaction due to the quantum physical tunneling effect oc-curs with the surface in the form of tunneling current (Figure 3.2-10) To obtain ameasurable signal, the distance from the tip to the sample must only be about

10 Å With highly sensitive amplifiers, it is possible to detect currents up to 2 fA(10–15A) Influences from vibrations with amplitudes up to 1 lm (floor or sonicvibrations) and thermal drift of the components in the range of approximately0.1lm/cm are problems encountered in implementing this measuring method

It is possible to choose between two different modes If, during the measuringoperation, the height of the tip is controlled in such a way that the tunneling cur-

rent remains constant (IT(x, y) = constant), the probe displacement in the vertical

direction provides a measure of the profile height of the surface at the measuringpoint (Figure 3.2-11)

In the second mode, the height of the needle tip is kept constant

(Z (x, y) = constant) and the variation in tunneling current is acquired during the

Fig 3.2-9 Design of a scanning electron microscope

Cathode Tungsten or Lanthanum Hexaboride

measuring process (Figure 3.2-12) With this mode, however, there is a dangerthat the tip might come into contact with the surface because of irregularities ofthe sample and that incorrect measured current values could be obtained because

of electrical contact

The advantage of this measuring method is its very high resolution of about0.01 nm The disadvantage is its low measuring range (laterally maximum

100lm, in the z-direction maximum 10 lm)

Fig 3.2-10 Schematic

represen-tation of tip and sample

interactions (tunneling effect)

Fig 3.2-11 Constant-current

mode

Fig 3.2-12 Constant-height

mode

3.2.3.3 Scanning Near-field Optical Microscopy (SNOM)

SNOM is high-resolution optical microscopy implemented by scanning a smallspot of ‘light’ over the specimen and detecting the reflected (or transmitted) lightfor image formation (Figure 3.2-13) This is the only similarity with confocal mi-croscopy, where the focal point is scanned Operation of conventional light micro-scopes suffers from the diffraction limit, which limits the optical resolution of themicroscope to only approximately a half wavelength of the light being used Theresolution of the SNOM image is defined by the size and the properties of theaperture, not by the wavelength used This means that SNOM provides an im-provement in spatial resolution of at least one order of magnitude over conven-tional optical microscopes However, the attainable resolution of approximately

50 nm is smaller than that for STM or AFM SNOM utilizes tiny apertures of ameters in the range typically 50–100 nm, ie, smaller than half the wavelength ofvisible light Typically such apertures are prepared in the metal coating at theapex of an optically transparent, sharp tip Light cannot pass through such anaperture, but an evanescent field, the optical near-field, extends from it The opti-cal near-field decays exponentially with the distance, and is thus only detectable inthe immediate vicinity of the tip

di-The optical resolution limit for SNOM is governed by the light intensity passingthrough the aperture, usually by heating and pulling a fabricated fiber tip A prac-tical limit is usually encountered at aperture diameters between 80 and 200 nmbut in ideal cases diameters down to < 20 nm have been achieved

If the aperture is brought close to the sample surface, the presence of the ple causes a disturbance of the optical near-field, which leads to the emission oflight from the location opposite the aperture Scanning the aperture at a distance

sam-of typically 10 nm from the sample with an accuracy sam-of*5Å, in order to prevent

Fig 3.2-13 Design of SNOM in combination with AFM