Thesis (doctoral)--Helsinki University of Technology, 2001. / Includes bibliographical references. Also available on the World Wide Web.

Recent developments in the field of nano-structured materials for technological as well as scientific prospective are quite interesting. In this context carbon plays a dominant role. Few examples such as carbon nanotubes (CNTs), fullerene, nanostructured diamond, as well as, amorphous carbon film, particularly, diamond-like carbon (DLC) coating are the areas of today’s research. This thesis deals with ECR plasma deposition of carbon in two different forms, i.e., Diamond-like carbon (DLC) and carbon nanotubes (CNTs) In the case of DLC coatings the chemical vapor deposition (CVD) and sputtering CVD configuration has been used. The carbon nanotubes have been grown using CVD configuration. DLC films were deposited by ECR-rf CVD mode, as well as, ECR sputtering mode. In case of CVD films, about 0 — 100 Watts rf bias was employed in steps of 20 Watts, corresponding to effective negative self bias voltage of 15 — 440 V. CH4 and C2H2 have been used as source gas for CVD films. Microwave power was optimized at 300 Watts. In case of sputtering, a cylindrical graphite target (diameter 9 cm and length 6 cm) kept at the exit of the Ar plasma was biased with -200 V. Films were deposited on floating substrate (temperature ~100 oC). Films were deposited on Si, quartz, and steel substrates and characterized by FT-IR, Raman, UV-Visible, Photoluminescence spectroscopy (PL), spectroscopic ellipsometry. Nanoindentation was used to evaluate the film’s elastic property. Pin-on-disk measurement was used to study the tribological property of the films. Electrical properties of the films deposited on Si [p-(100), 10 Ω cm] were studied using picoammeter / source measuring instrument by two probe method. FT-IR analysis showed sp3C-H absorption peak at 2930 cm-1 for the CVD films, while sputtered films did not show any C-H absorption. Raman spectroscopy was used to evaluate bonding aspects as well as hydrogen content of the films. Comparison of sp3C : sp2C among the films was done based on I(D) / I(G) of the Raman peaks, while hydrogen content was estimated based on background slope of the Raman spectra. It was observed that increase in rf bias induces more sp2C while hydrogen content decreases. An optimum substrate bias of 40 Watts was predicted from the Raman spectra. For sputtered films Raman spectra indicated the formation of nanocrystal diamond in a-C matrix. UV-Visible-NIR optical transmission spectroscopy was used to determine the band gap (Tauc), E0, of the films. It showed that increase in rf bias increases the absorption coefficient α. The films deposited from CH4 with a substrate bias of 0 and 20 Watts (i.e., high hydrogen content in the film) followed (hνα)1/2 = const. (hν –E0), while other films hνα = const. (hν –E0) ( h is Plank constant ν is frequency of light). E0 varied from 1.1 — 2.5 eV. It was assumed that for π--π* transition follows root relation while π--σ * transition follows linear relation. Spectroscopic ellipsometry was used to determine optical constants, film thickness, and interface thickness. Deposition rate found out to be ~100 nm / mints for C2H2, ~10 nm / mints for CH4, and ~2.5 nm /mints for sputtered films. Formation of interface layer of thickness about 5 —30 nm due to high energy ion bombardment takes place for the films deposited at 40 Watts rf bias or higher. Band gap and related phenomena was revisited from the data that was obtained from this instrument which reasonably matches with the earlier results. PL experiments were carried out at room temperature using lamp excitation source as well as laser excitation source (457.9 nm wavelength). In case of lamp excitation source any wavelength from 200 —900 nm region can be selected. PL spectra showed that there are two sources of PL signal, one from nanocrystal diamond and other from sp2C phase. To obtain PL signal from diamond UV excitation wavelength was required. This diamond phase is highly efficient emitter as compared to sp2C phase. Based on the closeness of diamond’s optical centre labeling of the peaks was done. For CVD films N3 ( 457 nm), H4 (495 nm), H3 (520 nm), [N-V]0 (~590 nm) optical centers of diamond was observed. For sputtered films [N-V]0 (2.08 eV), H3 (2.38 eV), H4 (2.50 eV), N3 (2.81eV), N3 (2.96 eV), 3.3 eV ( undocumented peak), 5RL ( 4.14 eV) optical centers of diamond as well as band-edge emission (5.01 eV ) was observed. Nanoindentation technique was used to estimate the elastic property and related phenomena of the films. It shows that the films are having hardness of 5—17 GPa and reduced modulus of 20 —120 GPa depending on the deposition parameters. All the films show highly elastic response at lower load, i.e., at low indentation depth where elastic recovery is 85—95 %. At higher load substrate effect comes into the picture. Further morphology in and around the region was evaluated using scanning probe microscopy (SPM). It was shown that substrate effect comes into picture that is based on film’s thickness as well as its elastic property. Films were further characterized by pin-on-disk experiments. C2H2 based films were used because of high deposition rate. Since 40 Watts, 60 Watts, and 100 Watts films adhere well with steel only on these films tribological test was possible. A hardened bearing-steel was used as substrate and a 2 mm diameter cylindrical pin made of tool steel was use as pin. Studies were carried out with three different loads of 20, 40, and 60 N. Friction coefficient varied from 0.02 — 0.04 and wear rate was found to be 10-6 — 10-9 mm3 / N m. A sputtered film of 0.1 μ m on the top of the CVD film, in many respects, enhances the tribological properties. It was shown that certain amount of wear is required for low friction of DLC. Electrical characterization of the films deposited from CH4 showed that they are highly insulating with resistivity of 1013 —1011 Ω-cm, and current conduction mechanism has been found to be predominantly space charge limited conduction (SCLC). Similar to the observations of Tauc’s relation, the film deposited with 0 and 20 Watts bias behave differently and followed the relation , where as, all other films exhibited the relation ( α, n are constants). It signifies that for 0 and 20 Watts rf biased films traps are uniformly distributed across the band gap while for others it decreases from the conduction band. For 0 and 20 Watts rf biased films no Ohmic current was observed at a detection level of 10-11 A. 40 Watts and higher rf biased films showed that three distinct regions in the I-V curves; initially Ohmic region, next to it SPLC region, and finally breakdown region. Increase in rf bias causes increase in Ohmic current. Film deposited from C2H2 showed diode-like behavior with higher conduction current limited by resistive control, and the resistivity of the films was ~ 109 — 105 Ω-cm. Difference in resistivity between the films deposited from CH4 and C2H2 was explained by considering the impurities in the source gas resulting in nitrogen doping concentration. Increase in Ohmic current for the CH4 films was explained by assuming the widening of the σ--σ * gap. Similar diode-like behavior was observed with the sputtered film. The last part of the work deals with the growth mechanism of aligned CNTs and their field emission (FE) properties. Nanotubes were grown at 700 0C on Ni coated (thickness 40 nm, 70 nm, and 150 nm) Si substrate using a mixture of CH4 and H2 gas. Microwave power of 500 Watts was optimized for nanotube growth. Nickel nanoparticle formation mechanism from a continuous Ni film was explained by considering the stress that is generated due to the difference in thermal expansion coefficients of Si and Ni at 700 oC. Though the thicker film such as 150 nm does not form nanoparticle due to stress, hydrogen induced fragmentation of the film due the brittleness of the film even causes formation of finer nanoparticles. A substrate bias in the range 0— 250 V was used to align the nanotubes. Perfectly aligned CNTs were obtained at -250 V substrate bias. The density of the tubes varied from 108 —109 / cm2 while its length was 0.5 — 2 μ m. Due to hydrogen induced fragmentation of the films, 150 nm Ni thick film showed smallest diameter 2 — 5 nm CNTs. 40 nm films showed nanotube diameter of 10 — 30 nm and 150 — 300 nm while 70 nm showed 10 — 30 nm diameter nanotubes. Diameter of the nanotubes was estimated using transmission electron microscopy (TEM). Field emission analysis of these CNTs was done using Fowler-Nordheim (F-N) plot and the investigation revealed that the field emission properties strongly depend on density and aspect ratios. The non-linearity in the F-N plot or current saturation phenomena was explained in terms of change in work function due to heating effect during FE which was pronounced in case of longer nanotube. Suitable efficient cold-cathode emitters for a particular usage (assuming that the variables are applied field and emission current) could be designed from the obtained results. An ammonia gas sensor using thick nonaligned CNTs was realized. For this purpose a thick film of CNTs (~ 0.5 μm) was deposited. This sensor can detect 100 ppm level of ammonia. About 1.5 — 4.5 % change of resistance depending on ammonia concentration (100 —1000 ppm) was observed.

Réduire l’énergie générée par le frottement et dissipée dans les systèmes mécaniques réels est un des challenges actuels en tribologie. Ce point représente une importance toute particulière dans le domaine des transports terrestres. En réponse à cette nécessité, les constructeurs automobiles se concentrent sur la réduction de la consommation d'énergie en sélectionnant des lubrifiants et des matériaux appropriés d'une part et les systèmes mécaniques performants d'autre part. Grâce à leurs propriétés physiques et tribologiques en termes de réduction de la friction et de l'usure, les couches minces de DLC (Diamond-like Carbon) sont considérées comme l'une des solutions. Le comportement tribologique de couches minces de ta-C (carbone amorphe très dur dépourvu d’hydrogène) et de a-C:H (carbone amorphe hydrogéné) est ici exploré. D’autre part, des “lubrifiants verts“ et des additifs participant aussi à la réduction du frottement et de l'usure sont testés. Ces essais sont effectués dans différentes conditions en utilisant une nouvelle méthodologie expérimentale. Le tribomètre dynamique oscillant développé au LTDS possède la capacité de quantifier avec une très grande précision et sans recourir à une quelconque mesure de force, des niveaux de frottement faibles (dans la gamme 10 – 5 à 10 – 2), et permet en plus d’identifier différentes contributions du frottement. Dans un premier temps, une loi de frottement linéaire a été utilisée afin d’évaluer deux contributions de frottement. La première, μ0 est le coefficient transitoire de frottement quand la vitesse s’annule au changement de direction (du type frottement de Coulomb). La seconde, μ1, est une contribution dépendante de la vitesse de glissement. Ensuite, une étude numérique a été réalisée en appliquant une loi de frottement quadratique afin de mieux comprendre l'aspect dynamique des réponses libres. Cela nous a permis d'étudier numériquement la décroissance d’amplitude des oscillations déterminée grâce à la méthode de moindres carrés. Nous nous intéressons aussi à l‘étude des formes de l'enveloppe de ces oscillations en relation avec le modèle de frottement. Les résultats numériques et expérimentaux pour les lois de frottement linéaire et quadratique sont ensuite comparés. En complément, nous avons mesuré la force tangentielle correspondant aux tribosystèmes testés. Finalement, nous avons étudié numériquement un système dynamique masse-ressort à un degré de liberté, modélisé par la loi de frottement LuGre. Cette loi est décrite par le phénomène d’hystérésis et l’effet de décalage de temps qui ont été détectés expérimentalement. Les résultats expérimentaux obtenus avec le tribomètre oscillant montrent qu'ils sont qualitativement comparables à ceux obtenus en configuration classique cylindre-plan, travaillant à vitesse de glissement constante. Dans tous les cas, les résultats montrent la supériorité du tribosystème ta-C/ta-C dans la réduction du frottement μ0. De plus, les résultats numériques et expérimentaux sont cohérents. Cette étude montre qu’une loi de frottement quadratique est capable de décrire correctement toutes les formes d’enveloppes obtenues expérimentalement : droite, convexe, concave et une combinaison des formes convexe et concave. / Controlling friction is a one of the most significant challenges in the field of tribology. Its major purpose is directed towards the reduction of energy in real mechanical systems, especially in the area of transportation. In response to this necessity, the automobile industries are emphasizing on minimizing the consumption of energy by selecting the appropriate lubricants and materials on one hand and mechanical system with high performance on the other hand. DLC (Diamond-like carbon) coated surfaces are considered one of the solutions thanks to their physical and tribological properties in reducing friction and wear. In this study, we have been interested in investigating the friction behavior of both amorphous hydrogenated (a-C:H) and hydrogen-free tetrahedral amorphous carbon (ta-C) DLC coatings. Furthermore, some “green lubricants” and additives are tested which play a role in reducing friction and wear. These latter tests are performed at different operating conditions using a new experimental methodology, known as the dynamic oscillating tribometer. This original tribometer, developed at the Laboratory of Tribology and System Dynamics (LTDS), is able to measure the oscillatory motion corresponding to various tribosystems having low friction. This technique has the ability to quantify with remarkable precision and without any force transducer, low friction values (in the range of 10 – 5 to 10 – 2 ) and also to evaluate different friction contributions from the displacement and velocity-time responses of a mass-springdamper oscillating system. First, a linear friction law has been used for the systems tested in order to calculate two contributions of friction. The first one, μ0 is the transient friction coefficient at zero speed and the second one (Coulomb-type fiction), μ is a velocitydependent friction coefficient. Then, a numerical study is carried out in order to better understand the dynamic aspect of the oscillatory vibratory free responses. A quadratic friction law is used to model on the mechanical system of the apparatus. This allows us to study numerically the behavior of the decreasing amplitudes of the damping responses which are determined using the Averaging method. Furthermore, we are interested in studying the various forms of these oscillations’ envelop in relation with the polynomial expansion of the friction model. Also, numerical and experimental results are compared using both the linear and quadratic friction models. In addition, a particular attention is given to the measured tangential forces corresponding to our tribosystems. Finally, we study numerically a damped mass-spring single degree-of-freedom mechanical system, induced by LuGre friction law described by both hysteresis and lag effect which have been detected experimentally. Results show that friction tests performed with the oscillating tribometer can be qualitatively compared to those obtained with a conventional tribometer. Moreover, ta-C/ta-C surfaces reveal the lowest friction μ0 when tested by both oscillating and reciprocating tribometers. Also, both numerical and experimental results are found to be of good agreement. This study shows that a quadratic friction law is able to describe correctly the envelopes observed in our experimental tests: straight, convex, concave and a combination of both convex and concave.