Takadoum J. Materials and Surface Engineering in Tribology

Подождите немного. Документ загружается.

Materials for Tribology 175

to 100 for a white object and varies continuously between these extremes. L* is

represented as a vertical axis, whereas the horizontal plane comprises the two axes

a* and b*. The coordinate a* indicates position between red and green (positive

values of a* indicate red while negative values of a* indicate green), whereas the

coordinate b* indicates position between yellow and blue (positive values of b*

indicate yellow while negative values of b* indicate blue) (see Figure 3.49).

Figure 3.49. Representation of a point P at coordinates L, a, and b in the color space L*a*b*

We limit ourselves here to coatings produced using PVD (generally applied

using reactive cathodic sputtering) as this technique is the most widely used in

industry within the decorative sector.

The substrates that can be used with this deposition technique are those that can

withstand temperatures of the order 200°C without degassing.

White

Yellow

Blue

Black

Red

Green

176 Materials and Surface Engineering in Tribology

Among the coating materials that are used, titanium dioxide (TiO

) has the

advantage of generating so-called interferential films where the color varies as a

function of the film thickness. For example, it is possible to obtain white films for

coating thicknesses less than a few nanometers, yellow films when the thickness

exceeds 40 nm, red films for thicknesses around 50 nm, and blue and green films for

thicknesses between 50 and 100 nm [MANA 04].

The fact that the films are based on optical interference, however, has limited the

widespread use of TiO

coatings on an industrial scale. In fact, only flat surfaces can

be coated satisfactorily using this technique and, for complex surface geometries,

different parts of the substrate are not coated with the same thickness, leading to

differences in colors. Moreover, titanium dioxide films are brittle and their hardness

is limited; the combined interferential nature of the coating and its low abrasion

resistance therefore limits the use of this material in decorative applications.

In contrast to TiO

, titanium nitride (TiN) is a hard, non-interferential coating

which presents good resistance to abrasion and appears golden yellow. Its

remarkable mechanical properties and its attractive color mean that it is one of the

materials most widely used for decorative purposes.

However, the radiance and golden yellow color of TiN are not rigorously

identical to those of gold and, in spite of systematic work carried out on the

production of TiN films under varying experimental conditions, it has not been

possible to achieve a genuine golden yellow color [MANA 04].

There is no possible combination of the three parameters a*, b* and L* for the

different layers generated that corresponds totally to the reference parameters of the

gold standards (according to the normalized NIHS colors or Normes de l'Industrie

Horlogère Suisse). Indeed, even when two parameters coincide with the

characteristics of one of the gold standards, the third is very far from the expected

value. It is for this reason that in industrial applications such as watch-making,

jewellery and accessories, the 1 μm thick film of TiN is coated with a additional

layer of “gold flash” one-tenth of a micron thick.

This solution, widely used in industry, does however present a major drawback.

In particular, the thin layer of gold (a soft material) erodes relatively quickly and

reveals the slightly darker titanium nitride layer, which causes the surface to look

tarnished [CONSTANTI 06]. To solve this problem, most research in the field has

been aimed at developing hard, genuinely gold-like films. Among some of the

solutions envisaged are mixed TiN/ZrN films, TiZrN films or multilayered TiN/ZrN

coatings.

Materials for Tribology 177

These multilayered TiN/ZrN coatings appear to open up interesting

opportunities. Indeed, alternate nanometer-scale films of TiN/ZrN deposited using

reactive cathodic sputtering have managed to reproduce the gold standards 1N14,

2N18 and 3N18 by combining TiN (golden-yellow but too reddish) and ZrN

(golden-yellow, but too greenish). Optimizing the experimental process has yielded

the same color values as the 1N and 2N gold standards for the a* and b* values.

However, the third value L* requires further improvement [CONSTANTI 06,

MANA 04].

Between the hard, non-interferential golden-yellow titanium nitride and the

softer, interferential titanium dioxide, the class of titanium oxinitrides Ti

presents intermediate properties that yield hard, non-interferential coatings

presenting desirable colors for decorative applications. Indeed, depending on the

oxygen/nitrogen ratio, titanium oxinitrides can take on a wide range of colors

[MARTIN 02, VAZ 04].

films have recently been generated using the reactive sputtering

technique with pulsed gases [CHAP 05, MARTIN 02]. Many ternary compounds

ranging from TiN to TiO

were thus obtained and characterized, showing that the

shift from nitride to oxide is paralleled by a gradual decrease in hardness, the

Young’s modulus and conductivity.

Regarding the colors obtained, the films are interferential within the composition

range (between 30 and 100% oxygen); however, when oxygen content is between 0

and 30%, they are non-interferential and present much higher values for hardness

than TiO

Increasing the oxygen content has yielded a varied palette of colors. It has been

possible to continuously vary the color from golden yellow for TiN to metallic

bronze, and then blue, purple, and dark and light green for compounds higher in

oxygen content.

Red is a rare and highly sought-after color in decorative applications. It is

therefore significant that recent research has reported that the Fe-O-N system has

made it possible to obtain compounds presenting the same reddish color as hematite

(Į-Fe

). Iron-oxinitride films were then generated using magnetron sputtering by

simultaneously injecting oxygen and nitrogen into the experimental reactor

[PET 06].

Carbonitrides are yet another class of coatings which must be mentioned as they

offer interesting solutions for decorative applications. Indeed, these materials are

characterized by high mechanical properties and highly sought-after colors. The

principal examples are titanium carbonitride (bronze, gray), titanium/aluminum

178 Materials and Surface Engineering in Tribology

carbonitride (black, gray and aubergine), chromium carbonitride (gray) and zirconia

carbonitride (brass yellow) [CAT 05].

3.5. Characterization of coatings: hardness, adherence and internal stresses

3.5.1. Hardness

We define three degrees of hardness principally differentiated by the depth of

indentation (d) as summarized in Table 3.6. The load required to reach these depths

will vary according to the material being tested. With nano-hardness, the required

load ranges from a few micronewtons to several hundreds of millinewtons. With

micro-hardness, the required load ranges from a few tens of millinewtons to a few

newtons. Finally, with macro-hardness the required load ranges from a few newtons

to several tens of newtons.

Table 3.6. The three types of hardness

As a consequence, the choice of hardness test will be tailored to the type of

material under study as well as to the application envisaged. Nano-hardness will

therefore be preferred to characterize thin layers (less than a micron) while micro-

hardness will be used to analyze the effects of surface treatments such as shot

peening or thermochemical treatments (carburizing, nitridation, etc.). Macro-

hardness will typically be used to measure the hardness of steels following

treatments such as quenching or tempering.

In section 1.2.4.1, we described the basic principles of the hardness test and the

different methods used to measure the hardness of materials. In the case of coated

materials, carrying out hardness tests presents two major difficulties. Specifically,

measurements carried out under very light loads can be affected by a size effect due

to the close proximity of the surface (see section 3.5.1.1), while measurements

performed under very heavy loads can be affected by the proximity of the substrate

(see section 3.5.1.2).

Nano-hardness Micro-hardness Macro-hardness

Indentation depth

(μm)

0.001–1 1–50 50–1000

Materials for Tribology 179

3.5.1.1. The indentation size effect

Hardness is a mechanical property of the material and its value (determined

experimentally) should therefore not depend on the measurement test conditions.

However, we note that measured hardness values are systematically higher when the

load (and penetration depth) is lower. Many reasons have been suggested to account

for this indentation size effect (ISE): work hardening; the presence of a superficial

oxide layer; the influence of impurities segregating on the surface after annealing of

the material; the presence of a free surface acting as an obstacle to dislocation

movements; or even elastic return. Indeed, although elastic return is often negligible

in macro- and micro-hardness, it becomes significant when the impressions obtained

present indentation depths ranging from a few tens to a few hundreds of nanometers

[BUL 01a]. Another theory based on the concept of geometrically necessary

dislocations (GND) to accommodate the deformation gradient in the plastified zone

has also been suggested [NIX 98].

Two expressions are generally used to characterize the influence of the ISE on

the hardness:

– Model 1 [NIX 98]



[3.38]

where H is the measured hardness, H

is the measured hardness for an indentation

depth large enough so that ISE is not present, d is the indentation depth and d* is a

characteristic length that depends on the shape of the indenter determined from the

fitting of the experimental points by the model.

– Model 2 [BUL 01a]



[3.39]

where H is the measured hardness, H

is the measured hardness for an indentation

depth large enough so that ISE is not present, d is the indentation depth and k is a

constant determined from the fitting of the experimental points by the model as

shown in Figure 3.50. It clearly shows the validity of the model and illustrates the

significant influence of elastic return on the ISE. Indeed, nano-indentation tests

carried out using a blunt-tip indenter clearly yielded greater elastic return than those

performed with a sharp-tip indenter.

180 Materials and Surface Engineering in Tribology

0 200 400 600 800 1000

Plastic depth (nm)

Micro-hardness (GPa)

Blunt nano-hardness (GPa)

Micro-hardness (GPa)

Sharp nano-hardness (GPa)

Figure 3.50. Hardness variation with plastic depth for PVD TiN,

comparing experimental data (symbols) with model fits (solid lines).The

indentation size effect behavior is clearly visible and depends on

indenter geometry and data analysis method [BUL 01a]

3.5.1.2.

Hardness tests for coated materials

The experimental determination of the mechanical and tribological properties of

coatings is often a delicate procedure due to the proximity of the substrate which can

significantly affect measurement results.

In the case of hardness tests, if h and d represent the thickness of the coating and

the depth of penetration of the indenter into the surface (see Figure 3.51), then as the

ratio d/h increases beyond a critical value (d/h)

the measurements will be affected

by the proximity to the substrate. The hardness measured will therefore be

overestimated or underestimated depending on whether the hardness of the substrate

is higher or lower than that of the coating (see Figure 3.52).

Materials for Tribology 181

Substrate

Film

Indenter

Figure 3.51. Indentation of a coated material

Figure 3.52. Hardness variation with the d/h ratio (see Figure 3.51). H

is the hardness of the

substrate: a) the coating is harder than the substrate; b) the coating is softer than the

substrate. The indentation size effect has been neglected

182 Materials and Surface Engineering in Tribology

Many models have been designed to analyze hardness tests performed on coated

materials, as described in the following.

3.5.1.2.1. Buckle’s model [BUC 65]

Using experimental data, Buckle was the first to suggest a model allowing the

influence of the substrate to be taken into account. He divided the volume under

stress into twelve layers of equal thickness d parallel to the surface and attributed to

each layer a relative weight P

and hardness H

(see Figure 3.53). The contribution of

a given layer i to the total hardness measured will be given by the product H

. The

layer thickness is related to the indentation depth by the relation h = nd, where n is

an integer corresponding to the order (i) of the last layer d which still forms part of

the coating, with the zone i= n+1 then belonging to the substrate.

18.5

16.5

7.5

5.5

3.5

0.5

100

Figure 3.53. Buckle’s model of indentation. The volume under stress is divided into twelve

layers of equal thickness d parallel to the surface [BUC 65]

If the hardness of the substrate is given by H

and if H

is the hardness of the

film, Buckle’s model allows the hardness of the coated material (noted H

and

hereafter referred to as composite) to be expressed:

Materials for Tribology 183

()

fis i

iin

PH P



¦¦

[3.40]

This equation gives the hardness of the composite for an indentation depth d.

If we write:

and



[3.41]

then equation [3.40] may be written:

()cd f s

aH bH  with a + b = 1 [3.42]

If n is known, equations [3.41] can be used to calculate the values of the

coefficients a and b for different values of d or d/h. Figure 3.54 shows the variation

of a as a function of the d/h ratio. It is interesting to note that if the d/h ratio is lower

than 0.1, the measurement is only weakly influenced by the proximity of the

substrate. This observation forms the basis of the so-called one-tenth rule of thumb

that states that hardness measurements of a coated material can be deemed reliable

and unaffected by the substrate provided the depth of penetration of the indenter

does not exceed one-tenth of the coating thickness. This rule, established

experimentally based on Vickers hardness tests, is generally valid for this type of

hardness but only in the case of hard coatings deposited onto soft substrates. When

the coating is softer than the substrate, a significant pill-up forms around the

indentation due to the plastification of the soft film (see Figure 3.55) and this

considerably affects the predictions of the model. Indeed, this pill-up partially

withstands the applied load as reported in [CHA 88] and confirmed in [IOS 96].

184 Materials and Surface Engineering in Tribology

0.1

d/h

Figure 3.54. Variation of a (see equation [3.41]) with d/h ratio (see Figure 3.51)

Generally speaking, publications report values for the (d/h)

ratio ranging

between 0.07 and 1 [BEE 05, CHE 95, IOS 05, KOR 98]. Table 3.7 presents some

results obtained with different coatings. These results confirm that the one-tenth rule

is generally true in those cases when the coating is harder than the substrate.

However, significant variations are noted in the reverse situation when the substrate

is harder than the coating.

Figure 3.55. Indentation of a nickel film (hardness = 250 HV) deposited onto

a harder steel substrate of hardness 700 HV; a pill-up forms around the indentation