(1)Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 12

Full text

(1)Digital Comprehensive Summaries of Uppsala Dissertations from the Faculty of Science and Technology 12. Solving Problems in Surface Engineering and Tribology by Means of Analytical Electron Microscopy ERNESTO CORONEL. ACTA UNIVERSITATIS UPSALIENSIS UPPSALA 2005. ISSN 1651-6214 ISBN 91-554-6148-4 urn:nbn:se:uu:diva-4785.

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10) !""# "$ # % % % & ' (

(11)

(12) )

(13)

(14) *

(15) ' +

(16) *' !""#'

(17) &

(18) % *

(19)

(20)

(21)

(22) (

(23) % ,

(24) *

(25) ' ,

(26)

(27) '

(28)

(29)

(30)

(31) !' -1 ' ' ./0 1 2##324 3-23 . ) 5

(32) )

(33)

(34)

(35)

(36)

(37) %

(38)

(39)

(40)

(41) %

(42)

(43)

(44) '

(45) ''

(46)

(47)

(48)

(49)

(50)

(51) %

(52) %

(53) %

(54) ) % %

(55) %%

(56)

(57) %

(58) %%

(59) %

(60) ' .

(61) ) 5

(62)

(63) )

(64)

(65)

(66) % %' +

(67)

(68)

(69)

(70)

(71)

(72)

(73)

(74) )

(75) ' (

(76)

(77)

(78) %

(79) 2% (+

(80) ) )

(81)

(82) %

(83)

(84) %

(85) ' 6 )

(86)

(87)

(88)

(89)

(90)

(91) )

(92) )

(93)

(94)

(95) %

(96) %

(97) (+ ' (.

(98) %

(99)

(100) % % (/!

(101) ) )

(102)

(103)

(104) '

(105) )

(106)

(107) 7( ,8+$+

(108) )

(109) .

(110)

(111) %

(112) % ,( 93 )

(113) -

(114)

(115) % ' , %

(116)

(117)

(118) ) %

(119) % )

(120)

(121) ' (

(122) )

(123)

(124)

(125)

(126)

(127) %

(128)

(129)

(130) '

(131)

(132)

(133)

(134) + 2

(135) % % ' , :"

(136) + 2 % ) )

(137) % ' % ) %

(138)

(139) %

(140) )

(141) %

(142) )

(143) ' ;+<+

(144) % 5

(145)

(146)

(147) % 5

(148) +

(149) % % ' +

(150) )

(151)

(152)

(153)

(154)

(155)

(156) ) %

(157) % ) ) .

(158)

(159)

(160)

(161)

(162)

(163) 5 ' ( *

(164) % *

(165)

(166)

(167)

(168) +

(169) !

(170) ! " ! # $ %&'! ! ()%*+* ! = *

(171) +

(172) !""# .0 4# 24! 3 ./0 1 2##324 3-23

(173) $

(174)

(175) $$$ 23>-# 7 $<<

(176) '5'< ?

(177) @

(178) $

(179)

(180) $$$ 23>-#8.

(181) Till min familj.

(182)

(183) List of Papers. I. The effect of carbon content on the microstructure of hydrogenfree titanium carbide films grown on high speed steel by physical vapour deposition E. Coronel, U. Wiklund and E. Olsson, In manuscript. II. Microstructure of d.c. magnetron sputtered TiB2 coatings M. Berger, E. Coronel and E. Olsson, Surface & Coatings Technology, 180, (2004), 240-244. III. An analytical TEM study of (Ta,Al)C:C coatings in as-deposited and oxidised states E. Coronel, D. Nilsson, S. Csillag and U. Wiklund, In manuscript. IV. TEM studies of tribofilm formation - compatibility between two metal doped DLC coatings and lubricant E. Coronel, N. Stavlid and U. Wiklund, In manuscript. V. On the wear mechanisms of CVD diamond when sliding in nitrogen and argon atmosphere E. Coronel and J. Andersson and U. Wiklund, In manuscript. VI. Surface analysis of laser cladded Stellite exposed to self-mated high load dry sliding D. Persson, E. Coronel, S. Jacobson and S. Hogmark, Submitted to Wear. VII. Wear induced material modification of cemented carbide rock drill buttons U. Beste, E. Coronel and S. Jacobson, Accepted for publication in International journal of refractory metals and hard materials. The author of this thesis has performed all high resolution analytical electron microscopy of all papers..

(184)

(185) Contents. Part I I.1 I.1.1 I.1.2 I.1.3 I.1.4 I.1.5 I.2 I.2.1. I.3. I.3.1 I.3.2. I.3.3 I.3.4 I.3.5 Part II II.1 II.2 II.2.1. Background Introduction Tribology Surface Modification The Active Surface Tribological Testing General Tests used in this thesis Tribofilm Analysis Surface Engineering Deposition of Coatings Magnetron Sputtering Electron beam evaporation CVD Laser Powder Cladding Analytical Microscopy Electron Scattering Elastic Scattering Inelastic Scattering Microscopical Techniques Transmission ElectronMicroscopy Scanning Electron Microscopy Focused Ion Beam Conventional TEM sample preparation Energy Dispersive X-ray Spectroscopy Electron Energy Loss Spectroscopy Thickness measurements Relative Quantification Energy Filtered TEM Three Windows Technique. 9 11 11 12 13 13 13 14 15 17 17 17 18 19 20 21 21 22 25 26 27 28 28 30 31 32 34 35 36 37. Contributions Introduction Structure and composition of as-deposited and heat treated thin coatings (Papers I, II, III) Effect of carbon content on the microstructure of TiC coatings (Paper I) Conclusions. 39 41 41 41 45.

(186) II.2.2. Effect of substrate bias on the microstructure of TiB2 coatings (Paper II) Conclusions II.2.3 The role of Al on structure and stability of (Ta,Al)C:C coatings (Paper III) As-deposited coating Heat treated coatings Conclusions II.3 Revealing mechanisms of friction and wear (Papers IV, V, VI) II.3.1 Mechanisms of tribofilm formation for lubricated Me-DLC coatings (Paper IV) Surface morphology of tribofilms Structure of tribofilms Composition of tribofilms Conclusions II.3.2 Wear mechanisms of diamond surfaces in different atmospheres (Paper V) Sliding tests Analysis of wear debris Internal film characterizations Composition of diamond surface Conclusions II.3.3 Friction mechanisms of Stellite 21 (Paper VI) Sample preparation Microstructure Composition Conclusions II.4 Solution of a practical problem-Wear of cemented carbide in rock drilling (PaperVII) II.4.1 Rock drilling- A tough application of cemented carbide II.4.2 Sample Preparation II.4.3 Wear mechanisms of drill buttons Drilling quartzitic granite Drilling magnetite II.4.4 Conclusions II.5 Summary of Contributions II.6 Future challenges of HR-TEM Sammanfattning på svenska Bakgrund Bidrag och resultat Acknowledgements References. 45 47 47 48 50 51 53 53 54 54 55 58 59 59 60 61 61 62 62 62 63 63 65 67 67 68 68 68 70 70 71 72 73 73 74 77 79.

(187) Part I. Background.

(188)

(189) I.1. Introduction. The development of new materials and surface layers is one of the most significant motors for the technical progress of our society. New materials may be developed by curiosity or even by chance, but much more often as a respond to pronounced demands. The demands are of several categories; economical (longer life, lower price, easier to machine, etc.) technical (higher performance, lighter, stronger, less prone to wear and corrosion, etc.) and – today of increasing importance – environmental and health related (nontoxic, renewable, etc.). Successful development of new and improved materials is a complicated activity, requiring the collected efforts and skills from people of several fields of expertise. The development is always iterative, including several loops of material synthesis, testing, and modifications based on the evaluation. The research behind the present thesis encompasses work with coating synthesis and tribological testing and evaluation based on a solid knowledge about the inner structure and composition of the materials. By employing advanced surface analysis and high-resolution microscopy, the materials development and tribological evaluation processes are made much more efficient. Moreover, this evaluation can be performed both on the as-synthesised material and after tribological testing or practical use. Knowledge on how the material interacts on the atomic scale with the counter surfaces, provides even more information and gives hints on how to make the materials even better suited to performing their intended tasks.. I.1.1. Tribology. Tribology covers the science and technology of surfaces in contact and relative motion. This broad definition makes it extremely common and hence an important field of study. It includes friction, wear and lubrication and associated surface layers on the contacting bodies1. Tribology is also of great economic significance in all industrial sectors. In late years this has been especially evident in the development and use of tools, machine components and vehicles to meet today’s increasing economical and environmental restrictions. 11.

(190) Surfaces in tribological contact are subject to very harsh environments characterized by extreme local pressures, temperatures and deformation. Under these conditions, surface films are formed through reactions between the contacting materials and the surrounding atmosphere or lubricant, but also through phase changes resulting in a surface layer with different properties compared to the original surface. These surface films, often referred to as tribofilms, may be very thin i.e. from a couple of nanometers to a few micrometers. Despite their insignificant thickness they will govern the behaviour of critical tribological components. Obviously, the study of tribofilms or related material properties requires high resolving power which is provided by techniques such as Auger electron spectroscopy (AES), transmission electron microscopy (TEM), scanning electron microscopy (SEM) or atom force microscopy (AFM), among others.. I.1.2. Surface Modification. Low friction films have been used by humans for a very long time and yet there are areas of applications which just recently started to benefit from the advantages of surface modification or coating technology. The ancient Egyptians, for instance, used wooden sledges to drag the large stones used to build the pyramids. The sledge was dragged on a wooden track and recovered wall paintings reveal that a lubricant, probably water, was used to alter the surface properties and reduce the friction. This early documented use of friction-reduction technology shows that man has addressed this problem for a long time, and that the use of lubricants and oils today bear a close resemblance with the use then. However, in vacuum technology, space applications or food industry, where there is a need for low friction components, one cannot make use of liquid lubricants. In the case of vacuum technology and space applications it is difficult to keep the lubricants in position due to the low pressure2-4, and in the food industry a lubricant might contaminate the food5. To solve this problem, solid low-friction materials have been developed and used successfully. Today, thin layers of solid lubricants are used to reduce friction and wear in many applications with sliding contacts such as on video tapes, computer hard disks and reader heads6. The great economic significance of tribological development can be easily understood when considering the shortened work time in metal cutting. In the beginning of the 20:th century it took more than one hour of turning to remove the surface layer of a 0.5 m long steel rod with a diameter of 0.1 m using a carbon steel tool material7. When high speed steel (HSS) was introduced the working time was reduced to 30 minutes and with the first cemented carbide (CC) materials it could be done within only 6 minutes. In order to further increase the life time and quality of cutting tools as well as to decrease the working time, Sandvik introduced the Gamma Coating (GC) 12.

(191) process in 1969. This process made it possible to deposit thin (some µm thick) titanium carbide (TiC) coatings on CC tools. These new cutting tools reduced the working time so that the work that took more than one hour in the beginning of the 20:th century now could be performed in less than 1 minute. The success story of materials science and surface engineering in tooling application is still unfolding and today it is followed by a similar development in the tribological field of machine components.. I.1.3. The Active Surface. Immediately after a tribological surface has been set to work, reactions may build up a surface layer with significantly different properties compared to the original surface. Very often these layers determine the tribological behaviour of the component. A pin on disc experiment with a Teflon pin and an aluminium disc could serve as an illustrative example. In this experiment it was somewhat surprisingly observed that the Teflon pin produced significant wear of the disc, although Teflon is a very soft polymer material. It was found that very hard alumina particles from the aluminium surface were formed and embedded in the Teflon. Teflon thus was more capable of retaining alumina than aluminium itself and hence hard alumina particles abraded the soft aluminium disc. This shows the importance of studying the active surfaces of a tribological system in order to understand the friction and wear mechanisms encountered. Here, the counter surface sliding against aluminium was no longer the original soft Teflon surface but an alumina reinforced abrasive Teflon pin.. I.1.4. Tribological Testing. General The most reliable tribotest to simulate a real component in its authentic application is of course, a field test. However, this is often neither practically nor economically feasible. Instead, simplified laboratory tests mimicking the real application have to be used. The compatibility between the test and the real application has to be ensured, however. This is usually done by verifying that the wear mechanisms are the same for the test and the authentic application. A vast number of tribotests have been developed during the years to assess different wear, friction, seizure and lubrication mechanisms8-10. Here follows a description of the tests performed in this work.. 13.

(192) Tests used in this thesis A pin-on-disc or ball-on-disc experimental set-up was used to measure friction during sliding as a function of sliding distance, see Fig. 1, either with a rotating disc or a disc moving back and forth. It is a versatile set-up because different materials can be chosen for the ball and disc or plane in addition to the possibility of controlling temperature, atmosphere and lubrication.. Figure 1. Schematic image of the ball-on-disc rotating set-up and ball-on-plane fretting experimental set-up.. In this test friction is calculated by measuring the normal load and lateral forces as the disc or plane moves. After finishing the friction measurement it is possible to asses the wear from the weight reduction or by making a depth profile along the wear track. Further, wear debris accumulated along the rim of the wear track can be analysed with microscopic and spectroscopic techniques to establish its nature. This analysis may be of interest for coatings on both components and tools to examine any tribofilm formed on the surface, and hence identifying friction and wear mechanisms. The tribological load scanner was used to study friction, wear, seizure and surface damage as a function of load for a single stroke or after reciprocating strokes11, 12, see Fig. 2.. Figure 2. Schematic image illustrating the motion of the rod during a single stroke tribological load scanner experiment.. Parameters such as speed, load range and number of strokes can be independently varied in a load scanner. Experiments can also be performed with addition of lubricants. Samples are typically 100 mm long rods with a diameter of 10 mm.. 14.

(193) I.1.5. Tribofilm Analysis. The small dimensions encountered in tribology make microscopy indispensable when assessing results of tribological contacts. There are many different microscopy techniques available but one of the most versatile instruments is the scanning electron microscope (SEM) equipped with an energy dispersive X-ray spectroscopy (EDS) detector. This combined instrument has the ability to image a surface with nanometre resolution and simultaneously perform elemental analysis at the sub-micrometer level. Its imaging resolution is surpassed only by Scanning Probe Microscopy (SPM) techniques1. A focused ion beam (FIB) instrument combined with an SEMfunction provides the capability of producing a metallographical crosssection or transmission electron microscopy (TEM) sample at a precise location. The TEM can provide chemical as well as crystallographic information with atomic resolution and its spatial resolution is unsurpassed by any other technique. However, the actual volume that is studied is extremely small, and care has to be taken when generalising the results to the whole sample. In fact a rough calculation gives that the combined volume of all samples studied in all TEM´s since they became commercially available, in the 1960’ies, is only13 0.6 mm3. Spectroscopy techniques often used to study tribofilms are AES, Electron Spectroscopy for Chemical Analysis (ESCA), Glow Discharge Optical Emission Spectroscopy (GDOES) and Secondary Ion Mass Spectroscopy (SIMS). These techniques provide good statistics, i.e. relatively larger areas are analysed with a high depth- and spectral resolution.. 15.

(194)

(195) I.2. Surface Engineering. There are many means of producing surface layers suitable for tribological applications. In some cases the original surface is modified by heating14 or ion bombardment15 and in other cases a new surface is produced through deposition of new material. The coatings studied in this work were deposited using different deposition techniques, such as Physical Vapour Deposition (PVD), Hot Filament Chemical Vapour Deposition HF-CVD and Laser Powder Cladding (LPC).. I.2.1. Deposition of Coatings. Many different processes are used to deposit coatings on components or tools. Among these are the Chemical Vapour Deposition (CVD) and PVD processes. In CVD, different gaseous precursors are brought to react on a heated surface, i.e. the substrate, which might be a tool or a mechanical component. The surface is often heated to temperatures exceeding 1000°C16. The high temperature restricts the number of materials which may be considered as substrates. For instance, any tool steel would soften at these elevated temperatures. From this reason the hard, stiff and heat resistant CC materials are the most commonly used tribological substrates for CVD deposition. During the years there has of course been a desire to put coatings also on steels and other more temperature sensitive materials resulting in an evolution of deposition processes performed at lower process temperatures17. Several different deposition techniques are available in the PVD family, such as sputtering, electron beam evaporation, cathodic arc and laser ablation. Of these magnetron sputtering and electron beam evaporation have been used in this work and are described below.. Magnetron Sputtering The magnetron sputtering device consists of a water cooled cathode mounted on an array of magnets or electromagnets. The poles of the magnet are mounted so that the centre of the cathode surface is one magnetic pole and the outer rim is the other. The magnetic field lines extend with near parabolic shape from the central pole to the outer one, as depicted in Fig. 3. By 17.

(196) orienting the magnetic field and the electric field in this particular way there will be a closed EuB drift path on the magnetron surface.. Figure 3. Schematic image of a rectangular magnetron with the magnetic field lines and the ExB drift path indicated.. The coating chamber is first evacuated and then filled with argon, where after a plasma is ignited. The free electrons will tend to be partially constrained to, and travel along the drift path, and as a result, ionisation will be most intense in the vicinity of these tracks, producing a locally much denser plasma. Ionised atoms are accelerated towards the cathode by the negative potential. There they will erode the target material by sputtering and hence provide released target atoms which will then condense on the surroundings, atom by atom. The process of sputtering can be compared to billiard ball collisions where the atoms in the target material are knocked out from their initial position by incoming energetic ions as shown in Fig. 4.. a). b). Figure 4. When an argon ion impinges onto the target material a) it will scatter out target-material atoms b). The expelled atoms are then deposited onto any exposed surface.. Electron beam evaporation Electron beam evaporation performed in an evacuated chamber, can be used to deposit a wide range of materials (considering it being a heat evaporation technique) including compounds, alloys, metals and ceramics18. The materi18.

(197) als to be deposited are put into a water cooled crucible, often made of copper. The water cooling allows the melt to remain separate from the crucible by a barrier of solid material and hence prevents any reaction between the melt and the crucible. A typical electron beam evaporation unit bends the beam into a 270q arc by a coil positioned as depicted in Fig. 5, i.e. opposite to the electron filament. This construction allows protection of the filament as well as the possibility to scan the beam to vary the extent of the melt. The material in the crucible is heated, melted and brought to evaporation by the high energy electron beam. As with sputtering, atoms condensate everywhere on the surrounding surfaces.. Figure 5. Schematic image of a typical 270° electron beam evaporation unit. CVD Synthetic diamond can be produced by several different processes such as combustion synthesis, microwave plasma assisted diamond deposition, plasma torch and HF-CVD deposition. The diamond films studied in this work were produced by means of HF-CVD. In this process, tungsten filaments are heated to temperatures ranging 2200í2800°C, see Fig. 6. A working pressure of 30í40 mBar was used and the gas was a mixture of H2 and CH4. Deposition of diamond is achieved through reactions occurring on the filament surface, in the gas and on the growing diamond surface19.. 19.

(198) Figure 6. The HF-CVD sample holder shown together with the W filament wires in front.. Laser Powder Cladding The laser powder cladding (LPC) process uses gas-atomized powder that is melted onto the substrate by a laser beam, see Fig. 7. The rapid solidification of the clad layer, due to efficient cooling from the substrate material, results in a metallically bonded fine structured coating.. Figure 7. Principles of the laser cladding process. The laser beam is focused to melt the added metal powder onto the substrate. Inert gas protects the clad layer from detrimental oxidation.. 20.

(199) I.3.. Analytical Microscopy. As the requirements on the quality of tribological components continue to rise, the need for understanding the factors that govern their properties also raise. Many of today’s coating have structures so small that characterization on a nanometre level is required in order to reveal the microstructure or nanostructure of the coating. Often, it is necessary to combine several characterization techniques in order to fully uncover the overalll as well as the local chemistry and structure of such coatings. In this work, electron microscopy techniques have been extensively used and the reason for using electrons instead of light is that the acquired resolution, G, is directly proportional to the wavelength, O, of the light or electromagnetical wave as shown in formula (1)13.. G. 0.61O P sin E. (1). This means that when using visible light the highest resolution we may achieve is around 0.3 µm13 for green light. Electrons accelerated with 300 kV, however, have a significantly shorter wavelength and hence makes it possible to achieve a much higher resolution. The theoretically calculated resolution is though not attainable in an electron microscope. This is due to lens aberrations such as spherical aberration, astigmatism and chromatic aberration.. I.3.1. Electron Scattering. Electron scattering or diffraction is essential for any kind of electron microscopy. Scattering denotes a deflection by collision while diffraction denotes a deviation of a wave direction by an obstacle. As the electron can attain both wave and particle properties, both scattering and diffraction are used to describe electron and matter interaction. A negatively charged electron passing through a material interacts strongly with the atoms and gives rise to different signals, see Fig. 8. A primary electron, i.e. incident electron, may interact with the atomic nucleus, bounded electrons or free electrons but it may also be transmitted through a material unaffected. If the material does not scatter 21.

(200) or in other way respond to electron irradiation, it is invisible in electron microscopes. Especially in SEM analysis, when studying bulk materials, the incident electrons often loose all their energy and get absorbed by the specimen.. Figure 8. A schematic image showing the different emerging signals as a specimen is irradiated with an electron beam. Transmitted electrons are only observed if the sample is thin enough.. Electron scattering is often divided into elastic scattering and inelastic scattering. Both types of scattering give structural and phase information.. Elastic Scattering Elastically scattered electrons are responsible for most of the contrast observed in TEM images and they build up the intensity distribution in diffraction patterns. In the SEM they are responsible for the Z contrast observed from back-scattered electrons. To better understand the mechanisms behind the contrast observed both in the TEM and the SEM it is necessary to look closer into the interaction processes between electrons and matter. Elastic scattering may be divided into scattering from single atoms or scattering from several atoms together. Interaction with a single atom may occur in two different ways, both of which involve interaction with Coulomb forces. Firstly, it may interact with the electron cloud resulting in a small angular deviation or secondly, it may penetrate the electron cloud, interact with the nucleus and deviate through a larger angle, see Fig.9. The deviation can exceed 90° resulting in backscattering of electrons.. 22.

(201) Figure 9. Schematic figure of the interaction between high energy electrons and an atom. Coulomb interaction in the electron cloud, i.e. with a large impact parameter b, results in low angle scattering while Coulomb interaction with the nucleus, i.e. with a small impact parameter b, results in high angle scattering or even backscattering.. Scattering at larger angles can be explained by Rutherford scattering, in which the screening by the atomic electrons is neglected. Electron scattering V from atoms per unit solid angle : can be described by the differential cross section. dV d:. f. 2. (2). where f is the complex scattering factor, i.e. a function of the scattering vector q depicted in Fig. 10 or the scattering angle T. The differential cross section may also be described as a function of the elastic form factor F(q):. dV d:. 4 2 F q

(202) 2 4 a0 q. 4J 2 2 Z f x q

(203) (3) 2 4 a0 q. where a0 is the first Bohr radius, q is the magnitude of the scattering vector, J=(1-v2/c2)-1/2 is a relativistic factor, Z is the atomic number and ƒx(q), the Fourier transform of the electron density inside the atom, is the atomic scattering factor or form factor for an incident x-ray. Since Z occurs in equation 3, it is obvious that an incident electron is scattered by the total electrostatic field in an atom. Applying classical mechanics and neglecting the screening influence of the atomic electrons gives the following expression for the differential cross section:. 23.

(204) dV d:. 4J 2 Z 2 a02 q 4. (4). This equation gives a sensible approximation for large scattering angles, i.e. a small impact parameter. There are several different approximations available to include electron screening such as the Wentzel expression in which the nuclear potential is attenuated exponentially. Cross sections with higher accuracy are calculated using Hartree-Fock and Hartree-Slater methods20.. Figure 10. Diagram of elastic scattering; k0 and k1 are the wave vectors of the incident and scattered electron wave while q is the scattering vector. The magnitude of the scattering vector is q=2k0sin(ș/2). Bragg’s Law The following treatment of electron diffraction is not physically valid but provides a good mathematical model. Bragg’s law applies to reflecting waves at glance angles and not to transmitted waves as is the case in TEM analysis. However, the atomic planes seams to behave as mirrors for the incident electron beam. A crystal can be thought of as being an array of atoms or slits. When an electron wave impinges on the crystal it undergoes diffraction. To understand diffraction we need to use the wave nature of electrons. The waves passes the crystal and when these waves exit the sample, interference effects will occur resulting in constructive and destructive interference. This yields a periodic oscillation of diffraction intensity. For certain directions, which may be calculated from Bragg's law:. nO. 2d sin T B

(205). (5). . O 24. §d· 2¨ ¸ sin T B

(206) ©n¹. (6).

(207) constructive interference is obtained, where O is the wavelength of the electron wave, d is the distance between the atomic planes, TB the Bragg angle and n is an integer. From this equation we can observe that the shorter the distance between the lattice planes the larger the scattering angle since O is constant. The integer number n can be thought of as introducing additional atomic planes, for instance n=2 can be visualized as introducing an additional atomic plane in the centre between the original ones. In electron diffraction, reflections of specific sets of atomic planes are observed. These are shown in diffraction patterns, which can be used to identify the atomic structure of observed crystals since only specific sets of planes meet the diffraction conditions, i.e. are present in the diffraction pattern. Structure factor analysis selection rules, i.e. rules for which reflections might be observed for a specific atomic structure. From the diffraction pattern it is also possible to deduce the planar spacing dhkl for specific (hkl) planes, where hkl are the Miller indices of that specific set of atomic planes.. Inelastic Scattering Inelastic scattering which, involves energy transfer provides the possibility to extract chemical information, such as the constituent elements in the sample. An electron that passes through the sample may undergo phonon scattering or plasmon scattering or excite an inner or outer shell electron in an atom. These inelastic processes and upcoming signals reveal information about the sample. Even without being irradiated by electrons, the atoms in the sample vibrates about their lattice sites because of their thermal energy. An electron passing through a sample may generate more vibration in the form of phonons. Phonons are quanta of atomic vibration and they are of the order of kT in energy where k is Boltzmann's constant and T is the absolute temperature. This means that the energy loss for an electron upon generating a phonon is below 0.1 eV20. Phonon scattering is, in contrast to other scattering processes, temperature dependent and increases with factor of 2-4 between 10 K and room temperature. It also increases with the atomic number, Z, of the scattering atom. Plasmon scattering is due to excitation of electron waves in the sample. The valence electrons in the sample may be seen as oscillators which reacts with each other through electrostatic forces. The incoming electron produces an electric field, which through electrostatic forces starts oscillations of the valance electrons. This oscillation occurs at a characteristic angular frequency, Zp. The energy loss suffered by an primary electron differs between different materials but is in the range of approximately 2 to 40 eV An incoming electron may interact with an atomic electron, transfer energy to it and then continue through the sample. If the transferred energy exceeds a certain limit the atomic electron is ejected and hence escapes the 25.

(208) attractive force of the nucleus. When an inner shell electron is ejected the atom is left in an excited state with a higher energy than its ground state. It may return to a lower energy state if an outer shell electron fills the hole. In this process the atom may release the excess energy by emitting either radiation in the form of an X-ray photon or an Auger electron, i.e. a low energy electron. The fluorescence yield, Z, is the possibility ratio of X-ray emission to inner shell ionization and can be calculated using the following equation13:. Z. Z4 aZ4. (7). where a is approximately 106 for K shell. X-ray analysis is thus well suited for heavier elements but not so well for light elements. Also other factors, like absorption in the sample, further reduces the sensitivity to detect light elements. Secondary electrons (SE) are emitted from the sample as a result of electron irradiation. There are different types of secondary electrons and these might be divided into different groups. The fast SE are those that are ejected from an inner-shell and these electrons may have up to 50% of the incoming electron energy. So called slow SE are ejected from the valence or conduction band and have energies typically below 50 eV. The third kind of SE are the Auger electrons, which also have low energies. These latter electrons are ejected by a characteristic X-ray giving its energy to an outer shell electron. Detecting these signals all have different pros and cons. The slow SE have low energy and are hence strongly absorbed. This results in that the slow SE that may escape from the sample are those ejected from near the surface, i.e. only the surface is probed upon using these electrons for analysis. The same is true for Auger electrons that have an energy of only some hundreds of eV. The difference between the two is that Auger electrons have characteristic quantized energies that can be traced to a specific element. This is not the case with either fast or slow SE since the process behind the creation of these electrons can be quite complex.. I.3.2. Microscopical Techniques. There are two types of electron microscopes namely SEM and TEM. The surface technique SEM gives good topographic information together with, chemical information, although the spatial resolution is limited and not always enough for the required characterization. Advantages of the SEM are though that it can analyze relatively large areas, it is easy to use and the sample preparation is quite straight forward meaning that little or nothing 26.

(209) has to be done to the sample before it is put into the microscope. On the other hand the contrast in the image can be difficult to interpret correctly since many effects may alter the secondary electron signal from the sample. In the case of chemical analysis using the EDS detector it is important to realize that the x-ray signal is not emerging from the surface alone but also from an interaction volume that may reach 1 µm into the sample. In a homogenous sample it might not be a problem but whenever dealing with inhomogeneous samples, such as coatings there is uncertainty about what is included in the probed volume.. Transmission ElectronMicroscopy A TEM is used for studying materials at high magnifications. It is a very powerful tool in materials science and its uniqueness lies in the possibility to extract for instance crystallographic, elemental and chemical information with extremely high spatial resolution. The TEM has an electron source, focusing lenses, i.e. condenser lenses, an objective lens, several magnifying lenses and electron intensity detectors, such as fluorescent screens and cameras. As mentioned earlier, electrons are diffracted by crystals in the specimen and hence a diffraction pattern is formed at the back focal plane of the objective lens. At the back focal plane of a lens parallel rays coming into the lens will be imaged to a point. Since Bragg diffraction diffracts electron beams into specific angles, the back focal plane thus exhibit bright spots separated by an angle of TB for the specific diffracting planes. This presents several possibilities when imaging the specimen. If a diaphragm, in this case the objective aperture is introduced in the back focal plane of the objective lens this enables imaging either with electrons scattered by specific planes or by “unscattered” electrons. The former imaging mode is called dark field (DF) while the latter is called bright field (BF). There is also a selected area diffraction aperture (SAD), positioned at an image plane, with which it is possible to choose areas in the sample that contributes to the diffraction pattern. More information is available in books by Williams & Carter and Joy13, 21. The sample preparation prior to TEM analysis is highly dependent on the nature of the sample. A sample needs to be thinned to at least 100 nm over as large area as possible without the sample becoming too brittle to handle. The classical approach is to grind a dimple in a thinned disc with a grinding wheel. The final thinning is done in vacuum using a beam of argon ions with a small angle of incidence. During this stage some argon implantation will take place as well as amorphization of the surface layer13. A low angle of incidence of Ar ions reduces its thickness and hence the effect of these artefacts. It is though important to keep these effects in mind as they can distort the results of the analysis. Other artefacts can occur during polishing, for instance the use of SiC polishing laps and diamond laps will damage the sample within a depth of approximately 2 times the polishing particle size. 27.

(210) The main contribution to contrast in the TEM is elastic scattering so when studying inelastic signals from the TEM to get chemical information, the elastic scattering might redistribute the inelastic scattering from the sample. In EDS analysis strong diffracting conditions should be avoided since the diffraction contrast might modulate the EDS signal. Furthermore, thickness effects will also change the signal. The same applies to the TEM techniques of electron energy loss spectroscopy (EELS) and energy filtered TEM.. Scanning Electron Microscopy The SEM is considered to be one of the most important instruments when studying tribological wear mechanisms and is the most widely used electron microscope. It is used for characterizing all kind of materials, including biological samples. In an SEM an electron beam is scanned over the sample and the emissions generated by the beam are detected. The intensity of the signal controls the contrast in the image. An SE image shows surface topography due to the placement of the detector at the side of the sample. This leads to shadow and thus topographic contrast. Intensity of BSE images is dependent on the atomic number because backscattering increases with atomic number. BSE images also exhibit shadowing effects due to BSE trajectories being straight lines.. Focused Ion Beam In the FIB station, a beam of ions is scanned over the studied surface and the out coming electrons or ions are collected by a detector. The amount of out coming electrons or ions is directly proportional to the intensity of the acquired image. I.e. an FIB instrument works essentially like an SEM with the main difference that the scanning beam is composed of ions instead of electrons. The impinging ions sputter the material and even at low currents mass removal can be observed. The possibility to sputter or etch a material is used to dig into materials to expose their internal structure. It has been proven by many scientists that the FIB instrument is a very versatile and valuable tool when studying materials exposed to tribological contact or materials in general. The main reason is that the area of study and subsequent etching can be selected with nanometre precision. Also, the crystal-orientation modulated signal is beneficial when studying polycrystalline samples. The heavy ion penetrates deep into the material under investigation and the penetration depth depends on the crystallographic orientation. For certain crystal orientations, ions may tunnel along atom rows and hence end up very deep into the material. Ions interact with matter and SEs are produced. However, the SE signal strongly depends on the depth where they are produced since they have a short escape depth. This means that the contrast in the im28.

(211) age will depend on the penetration depth of the ions which in turn depends on the crystallographic orientation. The instrument used in this work has also an additional SEM column, providing a possibility to acquire SEM images of the area being etched. The two columns are 52° apart but focus on the same point, i.e. the working point, see Fig. 11. This provide a possibility to study the etched wall and measure features along the ion beam direction at the same time as material is being etched. It is hence possible to acquire SEM images during etching without having to tilt or rotate the sample, see Fig.12. In many FIB applications it is essential to protect the surface of the material that will be investigated from the intense ion beam, especially in tribology were the main interest is devoted to the very surface region. This is accomplished by depositing a thin metal film by ion or electron beam induced CVD inside the FIB instrument.. Figure 11. Schematic image showing the orientation of the electron and ion columns, the sample and the working point.. a). b). Figure 12. Image taken from the same are with a) the ion column and b) with the electron column.. 29.

(212) TEM samples are rapidly produced using the FIB instrument by etching trenches on either side of the area of interest, see Fig. 12. The ion current is reduced during the thinning process to accomplish a smoother cut and less radiation damage of the final sample.. Conventional TEM sample preparation TEM specimens can be prepared in a variety of ways but in this work they have been prepared by classical TEM sample preparation techniques and by FIB sample preparation. The goal with the specimen preparation is to achieve an electron transparent area at the site of interest. In classical preparation, the sample has to be cut into small pieces first, see Fig.13. The pieces are glued together and the resin let to cure. After this, the "sandwich" and resin is introduced into a brass tube with an outer diameter of 3 mm and the resin let to cure. Now the brass tube is cut into thin 0.7 mm discs which are polished from both sides in several steps first with silicon carbide paper grade 800 and finally with 1 µm diamond lap. At this stage the disc should be about 100 µm thick and it is time to focus on the interesting area. Using a dimpel grinding machine, see Fig. 14a, the disc is ground in the centre to a thickness of 20 µm or less. The dimpled area is also finally polished with 1 µm diamond to achieve a smooth surface.. Figure 13. The sample is cut into smaller pieces. b) Glue is applied and c) the pieces of sample are glued together under heat and pressure. d) The “club sandwich” is encased with epoxy into a brass tube and then cut into discs.. The last step is to further thin the area until it is suitable for TEM analysis, i.e. around 100 nm or thinner. This is done by ion etching, see figure 14b. The sample is bombarded with argon ions at low incident angle until the disc has a small hole in the middle. At this stage the material in the vicinity of the hole should be thin enough to allow electrons to be transmitted through the sample.. 30.

(213) a). b). Figure 14. a) : Schematic figure of the dimple grinder setup. A dimple is produced by the simultaneous rotation of the sample and the grinding wheel. b) A simplified view showing how the sample is rotated while Ar ions are accelerated towards the sample at low angles, typically between 1ºand 10º.. The easiest way to produce samples is suitable if the studied material is in powder form. Then the powder is simply mixed with ethanol to a slurry and then a thin carbon grid is dipped into the slurry and dried. The carbon grid will support the sample.. I.3.3. Energy Dispersive X-ray Spectroscopy. EDS analysis is often performed inside the electron microscope in order to retrieve the chemical composition of the analyzed material. A typical spectrum is depicted in Fig. 15.. Figure 15. A typical spectrum from an EDS analysis inside the TEM.. 31.

(214) Primary electrons may interact with atoms by transferring enough energy to eject a bounded inner-shell electron, thus leaving the atom in an excited state. The excited atom returns to its ground state by replacing the missing electron with an electron from an outer shell. In this transition, an Auger electron or a characteristic X-ray is emitted. The probability of X-ray emission versus Auger electron emission increases with atomic number and hence making EDS a suitable technique for heavier elements. The energy of the emitted X-ray is equal to the difference in energy between the two electron shells involved in the relaxation process and hence unique to the atom of origin. If now the intensity of the out coming X-rays is plotted as a function of their respective energy a spectrum is acquired. The energy-position of the peaks in the spectrum reveals the elements present in the analyzed material, i.e a qualitative elemental analysis is performed. However, quantitative analysis is often wanted and in order to perform good quantitative analysis effects such as absorption, fluorescence, bremstrahlung, escape peaks, sum peaks and k factors have to be considered. For sufficiently thin TEM samples it is possible to use the thin film approximation, in which absorption effects and fluorescence effects can be neglected. The mass concentration ratio of two elements, A and B, can be written as22. CA CB. k AB. IA IB. (8). where CA is the weight concentration of element A, IA the measured X-ray intensity and kAB the Cliff-Lorimer factor, accounting for differences in ionization cross sections, fluorescence yield and detector efficiency between the two elements. The kAB values can be measured using standard samples but they are also often calculated when performing standardless quantification.. I.3.4. Electron Energy Loss Spectroscopy. Electron Energy-Loss Spectroscopy (EELS) can be performed in a TEM where it will benefit from the high spatial resolution of the microscope. Due to its high energy resolution, it is used to study chemical states of elements, measure sample thickness and to perform qualitative as well as quantitative elemental analysis. The nature of the intensity distribution in the spectrum makes it however best suited for light element analysis. As described earlier, there are many processes in which an electron might loose energy inside the specimen and hence the energy distribution of the transmitted electrons will extend over a large energy interval. For instance, when an incoming electron loses energy by exciting one single atom in the 32.

(215) specimen by knocking out a core electron it looses a characteristic amount of energy. The electron energy loss spectrum, i.e. the intensity of transmitted electrons as a function of energy loss, actually depicts the probability of an electron loosing a specific amount of energy as a function of energy loss. The differential cross section can be written as follows20:. dV i d:. 4J 2 Z 1 2 4 a 0 k 0 T 2 T E2

(216) 2. ° ª º ½° T 04 » ®1 « 2 2 ¾ °¯ «¬ T T E2 T 02

(217) »¼ °¿. (9). where T is the scattering angle, TE=E/(Jm0v2) is the characteristic scattering angle for inelastic scattering with E being the energy loss,. șE. E/(Ȗ m 0 v 2 ). (10). where E is the mean energy loss, and T0 = (k0r0)-1 where r0 is the screening radius defined for a Wentzel potential and equal to a0Z-1/3 using the ThomasFermi model. Calculation of cross sections is required for quantitative EELS analysis. Often an angle limiting aperture, the objective aperture, is used to select the central beam. In that case the differential cross section should be recalculated to take also the limiting angle aperture into account. The EEL spectrometer bends the electron beam and disperses the electron as a function of their energy. A typical EEL spectrum as the one shown in figure 16, is composed by the intense zero loss peak, the low loss region, and the characteristic energy loss region.. Figure 16. A typical EELS spectrum showing the different energy-loss regions.. The phonon scattering is not resolved since the resolution of the EEL spectrometer is seldom better than 0.5 eV and phonon scattering is of the order of 0.1 eV. Next to the zero loss peak the plasmon peak is observed, both surface and volume plasmons may exist. Depending on the thickness of the sample we see one or several equidistant plasmon peaks. Since the probability of an electron exciting a plasmon can be expressed in terms of a mean 33.

(218) free path, obviously an electron might excite two plasmons and hence have lost twice the energy of one plasmon and so on. Beyond the plasmon peaks the intensity decays exponentially until the first characteristic ionizing edge. Depending on the probability of a core electron being either excited to an empty state or just excited to a free level, the energy loss near edge structure (ELNES) will look quite different as can be seen in the case of graphite and diamond, see figure 17. From this it is possible to extract chemical information with the high spatial resolution provided by the TEM.. a). b). Figure 17. EELS spectra showing the differences in energy-loss near edge structure (ELNES) between graphite (a) and diamond (b). Actually diamond should not exhibit the observed ʌ* peak at 284 eV energy-loss, However in case it is present due to amorphous carbon originating from the sample preparation.. Thickness measurements From the EELS spectrum it is possible to retrieve the thickness of the observed sample. This is achievable since the inelastic scattering will increase with increasing thickness. The intensity ratio between the zero loss intensity and the total intensity of the spectrum is governed by the total mean free path, O, for all inelastic scattering. t. § It · ¸¸ © I0 ¹. O ln¨¨. (11). If O is known for the studied sample the local thickness can be retrieved. O varies between approximately 50 nm to 200 nm for different materials. If an angle limiting aperture is introduced then O(E) should be used.. 34.

(219) Relative Quantification Relative quantification can be performed if the background is subtracted from the edge (see figure 18). The following equation can be used for such quantification of two materials A and B:. NA NB. I A ( E , ')V A ( E , ') I B ( E , ')V B ( E , '). (11). where N is the number of atoms per unit area contributing to the edge, I the edge intensity above background (see figure 18) and V the cross section. Note that the angle limiting aperture or collection aperture with semi angle E and the integration windows ' affect both the intensity I and the cross section V. Prior to a quantitative analysis, the background contribution has to be subtracted. The background under the edge is often calculated using the power law model23:. I. AE r. (12). where I is the intensity at energy E and A and r are fitting parameters. The fitting parameters are calculated using the intensity in the window or windows, depending on the employed technique, i.e. two windows or three windows technique20.. Figure 18. An EELS spectrum of the carbon K-edge of TiC with the background calculated beyond the edge. IK , IB and ǻ are also shown.. 35.

(220) If the cross sections are not known, changes in the ratio will reflect changes in composition, though they are not quantitative but qualitative.. I.3.5. Energy Filtered TEM. Energy Filtered TEM (EFTEM) offers the possibility to image large sample areas using electrons with well defined energy hence making it possible to acquire elemental distribution maps within minutes. An energy filter is basically a spectrometer with additional lenses making it possible to reconstruct the image of a sample area from the EEL spectrum. At the energy plane, the plane where an energy loss spectrum is formed, a slit is inserted thus selecting an energy range for the transmitted electrons contributing to the energy filtered image. This is possible due to a number of post slit lenses correcting aberrations in the image due to the fact that the electrons have already been dispersed as a function of energy. The attainable resolution of energy filtered TEM images, roughly 1 nm, is mostly limited by the delocalization error, the chromatic aberration and the diffraction limit24. Delocalization error arises because a fast electron can pass at a certain distance from an atom and still ionize it. Chromatic aberration in the objective lens, i.e. electrons possessing different amounts of kinetic energy are focused with different strength affects the resolution. Diffraction limit is due to the size of the objective aperture and the fact that a lens will not image a point object to a point image but to a disc. All these errors can be summed in quadrature, e.g.. x2 y2 z2. (13). The effect on the resolution of each of these parameters is depicted in Fig. 19 together with their sum in quadrature.. 36.

(221) Figure 19. Attainable resolution at the carbon K energy (284 eV) for 300 kV electrons and using a energy selecting slit of 20 eV width.. Chromatic aberration can be reduced by using a small energy selecting slit and a small objective aperture. The objective aperture should be chosen considering the resolution that is required. However, larger objective aperture gives more intensity in the image, hence better signal to noise ratio.. Three Windows Technique To produce an elemental map, the background intensity in the image has to be subtracted. This is performed by acquiring images with energy interval positioned at lower energies than the edge of interest, see Fig. 20. The two pre edge images are then used to calculate the background intensity at the post-edge image, i.e. the image with energy interval under the edge.. 37.

(222) Figure 20. An EELS spectrum of the carbon edge in TiC showing possible position of the pre edge windows and the post edge window.. The background is calculated for each pixel in the post-edge image using the pre-edge images and then subtracted from the initial value of the post-edge image. In this way an image with the background subtracted, i.e. an elemental map, is produced. If two elemental maps are acquired then the same methods used in EELS quantitative analysis can be used here. Using equation 11 and hence dividing two maps and multiplying them with the corresponding cross sections gives an elemental ratio image with quantitative values23. Artefacts such as thickness effects due to varying thickness and diffraction effects, often observed in crystalline TEM samples, affects the intensity distribution also of an elemental map. A way to cancel these effects is by producing a jump ratio image. This image is calculated by simply dividing the post edge image with a pre edge image. Jump ratio images have low noise levels but are not quantitative even though they often follow the intensity observed in elemental maps.. 38.

(223) Part II. Contributions.

(224)

(225) II.1. Introduction. The primary aim of this work was to forward the knowledge of tribological surfaces by investigating the internal microstructure of thin coatings and surface layers used in tribological applications, and study their modifications due to tribological contact. Knowledge of the tribological response of the surface material to mechanical, thermal and chemical loads down to the atomic level will aid in the design of new and better coatings and materials. It will also be useful when selecting the best tribological materials for given applications. In all papers included in the thesis this has been achieved through the use of state of the art analytical electron microscopy.. II.2 Structure and composition of as-deposited and heat treated thin coatings (Papers I, II, III) A number of different PVD coatings for tribological applications were studied. Two of the coatings i.e. TiC and TiB2 are commonly used for tool applications while TaC:C is a low friction coatings intended for use on mechanical components. The fine scale microstructure was analyzed with respect to sensitivity to changes in substrate bias, composition and oxidation stability. Techniques used were analytical electron microscopy together with spectroscopy techniques such as EDS, EELS and EFTEM. This combination is well suited for uncovering the atomic and chemical structure of coatings.. II.2.1 Effect of carbon content on the microstructure of TiC coatings (Paper I) TiCx coatings were deposited by sputtering carbon and electron beam evaporating titanium. During deposition the substrates were cyclically exposed to both sources by means of a carrousel mechanism. Different carbon to titanium ratios were achieved by varying the power of the magnetron25. 41.

(226) From SEM images it was concluded that the carbon content has a significant effect on the coating morphology. High carbon content leads to the formation of large loosely attached columns of material while low carbon content results in a fibrous structure, see Fig. 21.. a). b). Figure 21. SEM micrographs of fractured cross-sections of thin TiC coatings with 59 at% of carbon (a) and 43 at% of carbon (b). The columnar morphology of the coating in (a) emanates from protruding carbides in the substrate surface.. The microstructure of TiC coatings was studied with TEM in order to uncover atomic scale structure.. a). b). Figure 22. Bright field TEM micrograph of the 43 at% of carbon TiC coating (a) and a corresponding selected area diffraction pattern (b) showing the 200 texture of the thin coating.. A low carbon content led to the formation of columnar grains with a strong texture with the <001> direction being parallel to the substrate normal, see Fig. 22. An increase of carbon content, however led to the formation of ran42.

(227) domly ordered nanocrystalline TiC grains dispersed in an amorphous carbon matrix. According to earlier studies this element composition is suggested to correspond to TiC+C26. The high amount of carbon in the 59 at% of carbon TiC coating led to the formation of randomly ordered nanocrystalline TiC grains with agglomeration of larger grains at the edge of the growth columns, see Fig. 23.. a). b). Figure 23. Dark field TEM micrograph of a TiC film with 59 at% of carbon (a) and corresponding SAED pattern (b). Bright areas in (a) correspond to Bragg diffraction conditions. The diffraction pattern in (b) reveals a structure of randomly oriented TiC nanocrystals.. Elemental analysis was performed by EFTEM. Both types of coatings were analyzed at thin parts, see Fig. 24. It was found that the Ti/C ratio increased as a function of distance from the substrate/coating interface until a steady state was achieved, see Fig. 25. It was concluded that this was a result of the temperature of the titanium evaporation source. The amount of available titanium increases with the temperature of the evaporation source which had not saturated when carbon sputtering begun. The initially deposited titanium layer, clearly observed in Fig. 25, has the purpose to enhance the adhesion of the coating to the substrate. Its Ti/C ratio is not comparable with the Ti/C ratio for the coating since there is no carbon in the interlayer.. 43.

(228) a). b). Figure 24. TEM micrograph of thin areas of the 43 at% (a) and 59 at% (b) carbon coatings (b). Note that much of the coatings has been milled away during specimen preparation.. a). b) Figure 25. Mapping of the Ti/C ratio and corresponding intensity profiles along the marked areas of the coatings with 43 at% (a) and 59 at% (b) carbon content.. 44.

(229) Conclusions x x x. Low carbon content resulted in coatings with columnar grains textured with the <001> direction parallel to the substrate normal. High carbon content resulted in randomly ordered TiC crystals in an amorphous carbon matrix. These observations indicate that the morphology of carbon based coatings can be tailored by varying the carbon content.. II.2.2 Effect of substrate bias on the microstructure of TiB2 coatings (Paper II) Thin coatings of TiB2 were deposited by sputtering from a TiB2 target. Though TiB2 offers excellent properties such as high hardness good chemical stability and good thermal and electrical conductivity its use for tribological applications has been limited. This is largely because TiB2 is prone to develop excessively high residual stresses during deposition27. However, it was found that by applying a positive bias to the substrate during deposition instead of the normal negative, it was possible to deposit TiB2 with low compressive residual stress maintaining its high hardness28. A positive substrate bias results in high mobility of adatoms on the growing film surface during deposition through thermal activation by predominantly electron bombardment. Negative bias cause bombardment solely by ions which more likely introduces defects and high compressive stress. Three coatings deposited on cemented carbide at three different biases; Vs = -110 V, floating potential Vf § -6 V and Vs = +50 V were studied. TEM cross section specimens were prepared by classical TEM specimen preparation. All three coatings exhibited a columnar structure with the grain axes being parallel to the substrate normal, see Fig. 26.. 45.

(230) a). b). c). Figure 26. TEM micrographs of the three TiB2 coatings deposited with different bias voltage. The images are montages of several individual micrographs. a) -110 V, b) Vf and c) +50V.. Near the substrate, however, a layer with more random orientation of the grains was observed. The texture developed gradually during growth, see Fig. 27. This same behaviour was also observed for the TiC coatings studied in Paper I. The thickness of this layer was approximately 50 nm and similar for all coatings.. Figure 27. SAED patterns acquired at different levels in the TiB2 coating deposited with Vf. Each SAED enclosed several grains.. 46.

(231) From Fig. 26 it was observed that the columnar grain size decreased as the bias increased from negative to positive voltage. For the Vf and Vs= +50 V the grain diameter was approximately 5 nm and observed to be constant from 200 nm from the substrate interface and up to the surface. A model for the development of film texture through evolutionary selection during growth has been proposed by van der Drift29. The texture results from differences in growth rate for different crystallographic direction. In this work the (0001) planes parallel to the substrate surface should be favored. This results in a texture development where finally the [0001] direction grows on the expense of other crystallographic orientations. At the beginning of the growth, however, [0001] grains will grow in the lateral direction also, but just until they meet each other thus hindering further lateral grain growth. This happens after 200 nm of film growth for the Vf and Vs = +50 V TiB2 films. In the case of the Vs = -110 V, the surface should have experienced higher intensity of ion bombardment than the other two films. This could have rearranged the atoms on the surface and led to the reduction of nucleation sites in the nucleation zone. Eventually, this would result in larger grain size in the film made with negative bias.. Conclusions x. x x x x. By increasing the bias from negative to positive voltage it is possible to reduce the grain size and at the same time reduce the internal stresses while maintaining a high hardness. The latter means that the thickness of TiB2 coatings can be dramatically increased without the risk of spontaneous coating detachment. All coatings were found to be dense with no voids or pores. All coatings exhibited a texture with the <0001> direction parallel to the substrate normal. The same advantage should be achievable in other coating systems that suffer from excessively high residual stresses such as TiN or (AlCr)N.. II.2.3 The role of Al on structure and stability of (Ta,Al)C:C coatings (Paper III) Carbon coatings have attracted much interest due to their tribological properties such as low friction and good chemical stability. However, metal additions have been observed to reduce friction with maintained wear rate30. The. 47.

(232) addition of metals may also serve to alter the wear rate by changing the oxidation properties of the coating. Aluminium added to a TaC:C coating was shown to improve the oxidation resistance with almost unaltered friction properties for a certain concentration of Al31. However, the details of the role of aluminium was unknown. Tests of oxidation resistance were performed at different temperatures and Al contents. This work concerned a coating with about 11 at% Al, 15 at% of Ta,70 at% of carbon, 1.4 at% of O and 1.5 at% of Ar. The coatings were deposited by magnetron sputtering from a single carbon target partially covered with foils of tantalum and aluminium. Cross section TEM samples were produced by classical TEM specimen preparation techniques on one asdeposited coating and one heat treated at 500º C for about two hours.. As-deposited coating The coating exhibited the same microstructure throughout the whole thickness with an unclear layered structure, see Fig.28a.. a). b). Figure 28. TEM micrographs of the as deposited (Ta,Al)C:C coating. a) A layered structure of the coating seen in low magnification. Observe that much of the coating has been milled away during sample preparation. b) TEM lattice image showing small crystallites.. Electron diffraction suggested that the coating was composed of randomly ordered carbide grains. They could be revealed at high magnification, see Fig. 28b. It was also observed that the addition of Al changed the morphology of the TaC coating by removing the columnar structure usually developing in TaC coatings32. The layered structure was studied in detail using STEM EDS. The intensity in HAADF STEM images follows Rutherford scattering, i.e. as a function of Z2. Further, the HAADF signal is incoherent in nature meaning that 48.

No results found