Accepted Manuscript
Title: Effect of layer thickness on thermal properties of multilayer thin films produced by PVD Authors: B. Tlili, C. Nouveau, M.J. Walock, M. Nasri, T. Gharib PII: DOI: Reference: To appear in: S0042-207X(11)00353-8 10.1016/j.vacuum.2011.09.008 VAC 5485 Vacuum

Received Date: 20 February 2011 Revised Date: 5 September 2011 Accepted Date: 18 September 2011

Please cite this article as: Tlili B, Nouveau C, Walock MJ, Nasri M, Gharib T. Effect of layer thickness on thermal properties of multilayer thin films produced by PVD, Vacuum (2011), doi: 10.1016/ j.vacuum.2011.09.008 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Effect of layer thickness on thermal properties of multilayer thin films produced by PVD
B. Tlilia,b, C. Nouveaub, M J Walockc,a, M. Nasria, dT.Gharib.
UR. Mécanique Appliquée, Ingénierie et Industrialisation (M.A.2I), ENIT, BP 37, Le Belvédère, 1002 Tunis, Tunisie b Laboratoire Bourguignon des Matériaux et Procédés (LaBoMaP), Centre Arts et Métiers ParisTech de Cluny, Rue Porte de Paris, F-71250 Cluny, France c a

Department of Physics, University of Alabama at Birmingham, USA d Photothermal Laboratory, I PEI 8000 Nabeul, Tunisia.

Abstract

Deposition). The multilayers were obtained from the combined deposition of different layers Cr, CrN and CrAlN thick films on on AISI4140 steel and silicon substrates at 200°C, and evaluated with respect to fundamental properties such as structure and thermal properties. Cr, CrN and CrAlN single-layers were also prepared for comparison purposes. The structural and morphological properties of PVD layers were characterized by X-ray diffraction (XRD) and scanning electron microscopy (SEM) coupled with EDS+WDS microanalyses, stresses were determined by the Newton’s rings methods using the Stoney’s equation and surface hardening and hardness profiles were evaluated by micro hardness measurements. The XRD data and HRTEM showed that both the Cr/CrN, CrN/CrAlN and Cr/CrN/CrAlN multilayer coatings

AC C

exhibited B1NaCl structure with a prominent reflection along (200) plane, and CrAlN sublayer micro-structures composed of nanocrystallites uniformly embedded in an amorphous matrix. The innovation of this work was to use the thickness of three different coating types to determine the thermal properties. Furthermore, an empirical equation was developed for the thermal properties variations with temperature of AISI4140 steel coated with different multilayer coatings. The thermal conductivity of CrAlN single layered was lower than the

EP

TE D

M AN U

Cr/CrN/CrAlN, CrN/CrAlN and Cr/CrN thin layers were deposited by PVD (Physical Vapor

SC

* Corresponding author: Tlilii_Brahim@yahoo.fr. Phone: +33 (06) 85 30 58 95

RI PT

ACCEPTED MANUSCRIPT multilayer and the bulk material AISI4140. Moreover, the influences of structure and composition of the multilayer coatings on the thermal properties are discussed. The thermal conductivity of nanoscale thin film is remarkably lower than that of bulk materials because of its various size effects.

1 Introduction

Thin films based on carbides and nitrides of transition metals have been very successful in improving the performance of cutting tools. Although TiN is the most widely used in industrial applications, CrN has been gaining importance in recent years due to its superior corrosion and wear resistance, friction behavior and low internal stress [1-3]. However, there are an increasing number of applications where properties of single layer coatings are not sufficient. Recent research is focused on the development of nanolayered multilayer coatings in order to meet the requirements in terms of hardness, thermal stability, coefficient of friction, corrosion resistance and internal stress [4-11]. Nanolayered multilayer coatings are composed of two or three different layers at the nanometer scale. The bilayers of the multilayers can be metals, carbides, nitrides or oxides of different materials or a combination of one layer made of nitride, carbide or oxide of one metal and the second layer made of another metal. A variety of multilayer systems such as TiN/CrN [4, 5], TiAlN/CrN [6, 7], TiN/NbN [8-11], etc., have been studied extensively. However, there are very few reports on

AC C

the multilayer coatings based on CrN and CrAlN [12]. CrN films have been used for cutting tools, molding dies and machine parts. But the oxidation resistance of CrN is limited up to 800°C [13]. For hard protective coatings, thermal stability is a main property as they are exposed to high temperatures during the cutting process. The addition of Al to CrN system permits to work at higher temperatures where the oxidation occurs [14]. CrAlN coatings have been reported to be stable up to a temperature of 900°C depending upon the Al content in the coatings [15]. CrAlN coatings also exhibit higher hardness and a lower friction coefficient

EP

TE D

M AN U

SC

RI PT

Keywords: Multilayers Cr/CrN/CrAlN, thermal conductivity, PVD, model.

ACCEPTED MANUSCRIPT compared to CrN coatings [16, 17]. The structural and thermal properties have been studied by X-ray diffraction (XRD), Photo-thermal deflection (PTD) technique [18].

2 Experimental details

the composition presented in Table 1. The substrates were hardened and tempered to a hardness of 4.5 HV0.05. They were ground, polished and cleaned with trichloroethylene, acetone and alcohol in an ultrasonic cleaner. 2.1 Coatings Deposition and characterizations

The multilayers were deposited on mechanically polished steel (AISI4140), with a surface roughness Ra = 0.2 µm and silicon 10x10mm² (for internal stress and thickness measurements) SiO2 10x10mm² substrates by using a dual RF magnetron sputtering system (NORDIKO type 3500-13, 56 MHz) equipped with two targets of high purity (Cr of 99.995 % and Al of 99.999 %) as shown in Fig. 1, the working pressure is 0.4 µbar. Before deposition,

5 min. The deposition conditions such as target power, bias voltage, and deposition time are given in Table 2. The deposition temperature was around 200°C. Very high purity nitrogen

pressure was 10-7 mbar. In both cases, the targets were 101.6 mm of diameter and 3 mm thick.

AC C

The substrate/target distance was 80 mm. The Cr/CrN or Cr/CrN/CrAlN multilayers can be defined as graded coatings. Actually, the CrN or CrAlN layers were achieved after decreasing the argon concentration inside the deposition chamber from 100% to 80%, and increasing the nitrogen content to 20%. The morphology, thickness and the composition of the coatings were determined by Scanning Electron Microscopic observations (SEM) and EDS+WDS microanalyses (Jeol JSM-5900

EP

(N2 of 99.9999%) was introduced into the vacuum chamber as the reactive gas. The residual

TE D

the substrates and the targets were ultrasonically cleaned and etched in pure argon plasma for

M AN U

SC

RI PT

The substrate material used in the present study was commercial AISI 4140 steel with

ACCEPTED MANUSCRIPT
LV). XRD analyses (SIEMENS D500, Co-K(radiation)) permitted the determination of the crystalline orientations of the layers. The residual stresses (σ) were determined by interferometer (Newton’s rings method [21]) by the measurement of the curvature radius (R0, R) of the Si (100) specimens before and after

 σ =  s   s  −  6  (1−νs )   t f   R R0     

1  E   t2   1 1 

 E  respectively [28]. The  s  term is the biaxial modulus of the Si (100) substrate (E = 1.805 (1−ν )   s
1011 N/mm2). Both ts and tf are the thicknesses of the Si substrate and of the layer, respectively, while R0 and R are the curvature radius before and after deposition, respectively. 2.2 Photo-thermal deflection (PTD)

The thermal properties are determined by photo-thermal deflection (PTD) [23, 24]

determine the local thermal properties of a stacking formed with succession of thin layers described above (section 2.1). We proceeded by determining the thermal conductivity and

of the bi-layers of Cr/CrN and multi-layer Cr/CrN/CrAlN are determined using a

AC C

mathematical model and an appropriate experimental setup. 2.2.1 Experimental setup The PTD setup developed in our laboratory as shown in Fig. 2. It has been applied here

to study the photo-thermal signal; the sample is heated by a light from a 100W power halogen lamp with power modulated thanks to a mechanical chopper at a variable frequency. A He-Ne laser probe beam of diameter d =100µm, skimming the sample surface at a distance z, is deflected. This deflection can be detected by a four quadrant photo-detector and converted

EP

thermal diffusivity of a single layer of Cr deposited on the substrate then, thermal properties

TE D

which is a non-destructive technique. In this work we used this technique in order to

M AN U

SC

where Es and νs are Young’s modulus (181 GPa) and Poisson’s ratio (0.28) of silicon (100),

RI PT

deposition and calculated with the Stoney’s equation [21-22]:

eq (1)

ACCEPTED MANUSCRIPT into an electrical signal which is measured by a lock-in amplifier (EG&G5210). Through the intermediary of the interfaces of the mechanical chopper and the Look-in amplifier, a microcomputer will set the desired modulation frequency and read the values of the amplitude and phase of the photo-thermal signal and then draw their variations according to the square

2.2.2 Theoretical Model

The PTD technique consists in heating an absorbing sample using a modulated light pump beam. The optical absorption of the sample will generate a thermal wave that will propagate into the sample and in the surrounding fluid medium, inducing a temperature gradient, and as a consequence, a refractive index gradient in the fluid. A Laser probe beam skimming the sample surface and crossing the region with inhomogeneous refractive index gradient is deflected. Its deflection ψ may be related to the thermal properties of the sample and to the temperature T0 at the sample surface. 2.2.3 Deflection calculation

undergoes a periodic deflection as shown in Fig. 3 due to periodic variations of the refractive index of the fluid in the vicinity of the sample. A study of the signal deflection yields a

EP

relationship between the amplitude and phase [25, 26] of the probe beam deflection ψ as:
− 2 l dn µ ψ= T0 e f n µf dTf x

TE D

The laser probe beam which passes parallel to the sample surface at a distance x0

M AN U

SC

AC C

ϕ=−

x

µf

+θ +

5π 4

Where l is the width of the pump beam in the direction of the probe laser beam, n , µ f , and Tf

are the refractive index, the thermal diffusion length and the temperature of the fluid respectively. ϕ and θ are the amplitude and phase respectively at the temperature T0 of the

RI PT

root modulation frequency.

eq (2)

eq (3)

ACCEPTED MANUSCRIPT sample surface which are function of the thermal properties of the different media, x is the distance between the probe beam axis and the sample surface. 2.2.4 Calculation of the periodic elevation temperature T0 at the sample surface We consider the case of a stacking sample formed by n layers on a Si substrate as

the thickness of the i-th layer respectively. The temperature inside each layer is written as:

M AN U if n

Tn (x,t) =( Xn eσn ( x) + Yn e−σn ( x) −Zn eαn ( x) ) e jωt

Tn−1(x,t) = (Xn−1 eσn−1 (x+ln ) + Yn−1e−σn−1 (x+ln ) − Zn−1eαn (x+ln ) ) e jωt −−−−−−−−−−−−−−−−−−−−−−−−−−−−

T1 (x) = (X1eσ1 (x+ln ....+l2 ) + Y1e−σ1 ( x+ln ....+l2 ) −Z1eα1 (x+ln ....+l2 ) )e jωt Tb (x,t) = W eσb (x+ln ....+ls ) e jωt

TE D

Ts (x) = (Xs eσ s (x+ln ....+l1 ) + Yse−σ s (x+ln ....+l1 ) −Zseαs (x+ln ....+l1 ) )e jωt

By writing the heat flow expressions in each medium one obtains:

EP

Φf (x,t) = Kf σ f T0e

−σ f ( x) jωt

e

Φn (x,t) =− Knσn (Xn eσn ( x) −Yn e−σn ( x) −

AC C

αn α ( x) jωt Z e )e σn n if n n

Φn−1(x,t) = − Kn−1σn−1(Xn−1 eσn−1 (x+ln ) −Yn−1e−σn−1 (x+ln ) − −−−−−−−−−−−−−−−−−−−−−−−−−−−−

αn−1 Zn−1eα (x+l ) ) ejωt σn−1 if −ln − ln−1≤ x ≤ −ln if − ln....−l1 ≤ x ≤ −ln....− l2 if − ln....−ls ≤ x ≤− ln....− l1 if − ln....− ls ≤ x ≤ −ln....−l1

α1 α (x+l ....+l ) jωt Ze )e σ1 1 α Φs (x) = − Ksσs (Xs eσ (x+l ....+l ) −Yse−σ (x+l ....+l ) − s Zseα (x+l ....+l ) )e jωt σs
Φ1 (x) = − K1σ1(X1eσ1 (x+ln ....+l2 ) −Ye−σ1 (x+ln ....+l2 ) − 1 s n 1 s n 1 1 n 2 s n 1

Φb (x,t) = − KbσbW eσb (x+ln ....+ls ) e jωt

SC if if
0≤ x ≤ lf − ln ≤ x ≤ 0

Tf (x,t) =T0e

−σ f ( x) jωt

e

if − ln − ln−1≤ x ≤ −ln if − ln....− l1 ≤ x ≤ −ln....− l2

if − ln....− ls ≤ x ≤− ln....− l1

if − ln....− ls ≤ x ≤ −ln....− l1

RI PT
0 ≤ x ≤ lf − ln ≤ x ≤ 0

shown in Fig. 4, where Ki, Di and li are the thermal conductivity, the thermal Diffusivity and

ACCEPTED MANUSCRIPT
The temperature and heat flow continuity at the different interfaces

x = 0, x = −ln , x = −l n + ln -1 , …, x = - ln − l n -1 …. − ls

permit to obtain:

The temperature and heat flow continuity at the interface x = -ln permit to obtain the following equation :
− 1  X n −1   e −σ nln 1 1     − rn −1  Yn −1  =  c n e −σ nln 1 −1  0 0 E / E  E   0 n n −1  n −1    e σ n ln −c n e σ nln 0 − e −α n l n − c n rn e −α n ln 1 Xn      Yn      En 

M AN U
0

 e −σ nln −1  1 1    − rn −1  , Dn =  c n e −σ nln Where G n =  1 − 1  0 0 E / E  n n −1    0

e σ n ln −c n e σ n ln

In the same way we can write in the interface: x i = -ln - l n -1 ... - l i

 X1   X n   m11      Then  Y1  = M 2 . M 3 ....M n  Yn  =  m21 E   E  m  1  n   31

TE D

m12 m 22 m32

m13   X n    m23   Yn  m33   E n   

AC C

The writing of the heat flow and temperature continuity at the interfaces x=0 and x= -l3 –l2-l1 give respectively:

E E 1 1 X n = (1 − g )T0 + (1 + r3 ) n , Yn = (1 + g )T0 +(1− r3 ) n . 2 2 2 2 (4)

−σ l σl −α l And (1− b) e 1 1 X1−(1+ b) e 1 1 Y1−(r1−b)e 1 1 E1 = 0

Then X 1 =

m11 m ((1 − g )T0 + (1 + rn ) E n ) + 12 ((1 + g )T0 + (1 − rn ) E n ) + m13 E n and 2 2

EP

 X 1 =m11 X n + m12Yn + m13 E n In this case we write   Y1 = m21 X n + m22Yn + m23 E n

SC

Gn

 X n −1   Xn   X n −1  Xn  Xn            1  Yn−1  = Dn  Yn  or  Yn −1  = G n .Dn  Yn  = M n  Yn  E  E  E  E  E   n −1   n  n −1   n  n

− e −α n ln   − c n rn e −α n ln  and M n =G n−1 .D n .  1 

RI PT
 X i −1   Xi       Yi −1  = M i  Yi  E  E   i −1   i

ACCEPTED MANUSCRIPT
Y1 = m21 m ((1 − g )T0 + (1 + rn ) E n ) + 22 ((1 + g )T0 + (1 − rn ) E n ) + m 23 E n . 2 2 T0 E + (m11 (1 + rn ) + m12 (1 − rn )+ 2 m13 ) n and 2 2

Or too X 1 =(m11 (1 − g ) + m12 (1 + g )) Y1 =(m21 (1 − g )+ m22 (1 + g ))

That is to say X 1 =η1T0 +η 2 E n and Y1 =η 3T0 +η 4 E n

By replacing X1 and Y1 by its expressions in equation (4) one obtains:
(1 − b) e −σ 1 l1 (η1T0 +η 2 En ) −(1 + b) eσ 1 l1 (η3T0 +η 4 En ) − ( r1 −b) e −α1 l1 E1 = 0

What gives:

Finally T0 = ((1 + b) η 4 e σ 1 l1 − (1 − b)η 2 e −σ 1 l1 ) E n + (r1 −b) e −α1 l1 E1 With E i =

[

M AN U

((1 − b)η1 e −σ 1 l1 −(1 + b) η 3 e σ 1 l1 ) T0 = ((1 + b) η 4 e σ 1 l1 − (1 − b)η 2 e −σ 1 l1 ) E n + (r1 −b) e −α1 l1 E1

SC
] [(1 − b)η
1

RI PT e −σ 1 l1 −(1 + b) η 3 e σ 1 l1

T0 E + ( m21 (1 + rn ) + m 22 (1 − rn ) + 2 m23 ) n . 2 2

]

αi 2 K c i (α i2 − σ i2 )

AC C

EP

TE D

ACCEPTED MANUSCRIPT
3 Results and discussion 3.1 Characterization of the PVD coatings
According to the EDS+WDS analysis, in the case of multilayer coatings obtained with a two targets, above 20 % of nitrogen in the plasma, we obtain a constant atomic composition

on table 3. The other deposition parameters (working pressure, deposition time, and target voltage) had mentioned in table 2.

Alumina and chromium were used as the reference materials for quantitative analyses. In the table 3, the N/(Cr+Al) ratio varied from 0.95 to 1.2 for all the coatings, which reflects their stoichiometry. Furthermore, the Ra and Rq roughness of the films are low, and their variations do not exceed 4/100 on all surfaces. On the other hand the evolution of the thickness versus the deposition time is always linear.

Fig.5 shows XRD patterns of the synthesized Cr/CrN/CrAlN and CrN/CrAlN nanoscale multilayers with the same thicknesses (1500 nm). Both the multilayers coatings were

the films are unambiguously identified by the peak (220) to 76.68° and (200) to 51.2°, respectively. The coating has CrN/CrAlN main diffraction peaks [(111) and (200)

PCPDFWIN Version 2.3, JCPDS-ICDD (2002)] [27 - 29]. One can conclude from the spectra

AC C

obtained on multilayer coatings Cr/CrN/CrAlN and CrN/CrAlN that the CrN surface layer crystallizes according to the preferred orientation (311) when it crystallizes in volume according to the preferred orientation (200) [2, 29 - 31]. Moreover, the presence of preferred crystallographic orientation is related to the columnar morphology of monolayers Cr, CrN and CrAlN observed by SEM on cross section as shown in Fig. 6. The growth of the columnar structure perpendicular to the surface is little disturbed at the interfaces between layers. The morphologies of the traditional CrN and nanocomposite

EP

respectively at 43.76° and 51.2° Characteristics of CrN, compared to the PDF 76-2494,

TE D

crystallized into a rocksalt-type cubic structure. The layers of chromium and CrN coatings in

M AN U

SC

RI PT

of 50-52 % of N2, 4-5 % of Al, 41-43 % of Cr and 3-4 % of O2 (which is negligible) as shown

ACCEPTED MANUSCRIPT
CrAlN films (5% of Al content) are studied. It was found that the CrN film exhibited strong columnar structure and no significant difference was observed after alloying with Al. However, the CrAlN film had much smaller grain size (the mean radius of grain is about 12 nm) and denser structure when compared with CrN film deposited under the same conditions

crystallinity due to Al alloying and the formation of an amorphous/crystalline nanocomposite structure. Our results are comparable to those obtained in the work of Barshilia et al. [32], as well as Okumiya et al [33] made on the coating CrN/CrAlN.

The results of L.Wang et al. [34] suggested that the lattice structure of CrAlN films fabricated in this work formed a solid solution whereby Cr atoms were substituted by Al atoms since the covalent radius of Al (0.121 nm) is smaller than that of Cr (0.139 nm). This also caused the shift of the diffraction peaks of the films. In addition, the peak broadening phenomenon and the decrease of the diffraction peaks intensity was also observed with the incorporation of Al in CrN matrix, which could be attributed to the fact that the crystallinity of film

In order to investigate the micro-structural changes of CrAlN films, high resolution TEM observations were performed on CrAlN film as shown in Fig.7.

coating layer was well grown crystalline phase with relatively regular grain shapes and

AC C

defects. In addition, it was also found that fine grains about several nanometers were observed in the big grains of this film (Fig. 7B), and at the higher magnification small crystalline grains about some nm (black area distinguished by the lattice fringe contrast) were embedded in the amorphous (white area) matrix. The XDR and HRTEM investigations detailed above have proved that the incorporation of Al into the CrN matrix led to the formation of a unique nanocomposite micro-structures composed of nanocrystallites uniformly embedded in an amorphous matrix. Such formation is

EP

From the TEM images for the film CrAlN cross-section (Fig. 7A), it was observed that the

TE D

correspondingly decreased [34].

M AN U

SC

RI PT

[32-34]. The dissolving columnar agglomerates were considered resulted from reduced

ACCEPTED MANUSCRIPT based on a strong, thermodynamically driven and diffusion-rate-controlled (spinodal) phase segregation that leads to the formation of stable nanostructures by self-organization [34]. Based on the literature, this amorphous/crystalline nanocomposite structure may be favorable to the high performance of mechanical properties owing to the following concepts [34]:

toughness; and (ii) maintain nanocrystalline size at several nm levels to restrict crack and create a large volume of grain boundaries.

Nevertheless, the addition of a third element in the transition metal nitride coating generally increases the residual stress [35]. Therefore, the CrAlN coatings are expected to have higher residual stresses than CrN ones. High internal stresses affect the adhesion of the coatings critically. For this reason thick CrAlN coatings cannot be obtained. These factors led to the exploration of multilayers such as CrN/CrAlN, wherein the properties of CrN and CrAlN can be combined.

Fig. 8 shows that residual stresses in multilayers are lower than the monolayers ones. The

is also mainly due to their interface. The interlayers seem to be effective in absorbing residual stresses during film growth. It’s obvious that their values strongly depend on the residual

3.2 Thermal properties

AC C

The determination of the thermal properties of the layers has been made as follows: we

started to determine the thermal properties of a chromium layer, and then we studied a sample Cr/CrN coated and determine the thermal conductivity and thermal diffusivity of the CrN layer. After we have understood the thermal properties of the Cr and CrN layers, we could study the thermal properties of CrAlN from the Cr/CrN/CrAlN coated sample. The AISI4140 steel substrate and the layers have a great optical reflection coefficient that is why they should be covered with a thin graphite layer to absorb the incident light and

EP

stresses of the monolayers.

TE D

reduction of the residual stresses in multilayer coatings compared to those of the monolayers

M AN U

SC

RI PT

(i) combine crystalline and amorphous CrAlN matrix to achieve high hardness with good

ACCEPTED MANUSCRIPT therefore serve as a heat source. The graphite layer thickness for all samples is 2µm, its thermal properties and those of the substrate were determined in previous works [18, 26]. Fig. 9 shows the experimental and theoretical amplitude and phase variations with the square root modulation frequency of the photo-thermal signal, for the graphite/Cr layers deposited on

thermal conductivity K = 93.9 W.m-1.K-1 and thermal diffusivity D = 0.79×10-4m2.s-1 for the Cr layers.

We purchased our work for the sample composed of steel/Cr/CrN/graphite using the theoretical model for three layers (n=3) whose results are presented in Fig.10. The thermal conductivity and the thermal diffusivity values obtained are K = 11 W.m-1.K-1 and D = 0.23×10-4m2.s-1 respectively. Then we determine the thermal properties of the CrAlN layer deposited on the sample consisting of stacking steel/Cr/CrN/CrAlN/graphite, the curves of Fig. 11 show that the CrAlN layer presents the following values K = 2.8 W.m-1.K-1 and D = 0.052×10-4m2.s-1 of thermal conductivity and diffusivity respectively.

parameters used as input in the calculations. Especially, the coating to substrate thickness ratio is small; the calculated conductivity value will be extremely sensitive to errors in the

strongly depends on possible errors of seven input parameters whereas the accuracy of the

AC C

thermal diffusivity determined for one-layer is only influenced by errors in sample thickness of layer. Therefore, single layers are strongly preferred if high accuracy is needed. The table 4 resumes the thermal properties of the different layers taken separately and of the equivalent thermal conductivity of the stacked films given respectively by:

for the stacking Cr/CrN and

EP

input parameters. In general, the accuracy of the laser flash technique for two or three layers

TE D

The thermal conductivity of the coating can be containing a large error due to inaccurate

e e 1 1 = ( Cr + CrN ) Ke eCr + eCrN KCr KCrN

M AN U

SC

RI PT

AISI4140 using the mathematical model of two layers (n=2). The best theoretical fit yields a

ACCEPTED MANUSCRIPT e e e 1 1 ( Cr + CrN + CrAlN ) = Ke eCr + eCrN + eCrAlN KCr KCrN KCrAlN for the stacking Cr/CrN/CrAlN. In a model developed by Ghrib et al. [25, 26] the overall coating conductivity results from the

zone model were modified in this work describing the total thermal resistance of a coating as a series combination of the thermal resistance of the inner zone and of the outer zone, providing a good agreement between model and measured data.

We note from the values summarised in table 4, that the CrAlN layer has lower thermal properties compared with the Cr and CrN layers, so it can act as a thermal insulator. The conductivity value of CrAlN is extremely low compared to those of the monolayer of Cr and CrN. This is well known and can be explained by the columnar microstructure which is characterized by elongated inter-columnar pores that are predominantly aligned perpendicular to the plane of the coating [36, 37]. In the CrAlN layer, the heat flow is mainly parallel to

columns exhibit a feathery structure with inter-columnar pore [38]. These fine inter-columnar gaps contribute to a moderate reduction in the thermal conductivity as they are generally inclined to the heat flow. For that reason, the thermal conductivity of CrN coatings is still significantly lower than the thermal conductivity of the fully dense Cr layer. This result

AC C

suggests that the structure of the Cr and CrN coating near the interface to the metallic substrate has a markedly different thermal conductivity. Structural differences of Cr/CrN/CrAlN and Cr/CrN PVD multilayers can be roughly divided into two zones, the fine grained inner zone (Cr/CrN film) and the coarse columnar outer zone (CrAlN coating). The thermal conductivity of the fine grain inner zone is much higher than the thermal conductivity of the outer zone. This effect is attributed to the higher grain boundary density as well as numerous oblique columnar boundaries in the outer zone [39]. Thus, the thermal conductivity is dominated by phonon scattering at defect/grain boundaries in this part of the coating, and

EP

TE D

these inter-columnar gaps, the heat transfer is not effectively reduced. However, the primary

M AN U

SC

RI PT

combined effect of these two zones (amorphous and crystallites). Essential parts of this two-

ACCEPTED MANUSCRIPT this results in a lower conductivity of around 2.5 W.m-1.K-1 at room temperature. Moreover, the decreased thermal conductivity of Cr/CrN/CrAlN multilayers is attributed to the increased total porosity resulting from an increase in the number of interface pores associated with the formation of non-uniform interfaces between layers, which causes increased phonon

Conclusion

The physic-chemical, mechanical and thermal properties of different nitride-based coatings obtained by PVD have been determined, taking into consideration their microstructure. CrN and CrAlN coatings deposited on AISI4140 substrates exhibited B1 NaCl structure with (200) reflection of cubic phase. The CrN/CrAlN and Cr/CrN/CrAlN multilayers also exhibited a prominent reflection along (200) planes.

Moreover, the present work reports experimental and theoretical results on the thermal properties (conductivity and diffusivity) and residual stress of Cr, CrN, CrAlN, Cr/CrN and Cr/CrN/CrAlN thin layers.

values determined on mono-layers and substrate are generally higher or equal than to the multilayer coatings.

Cr/CrN and Cr/CrN/CrAlN, it is clear that the latter coating has the lowest value.

AC C

3- The thermal conductivity of nanoscale thin film is remarkably lower than that of bulk materials because of its various size effects.

4- The obtained CrAlN coatings structure is porous and they present columnar grains containing many nanopores. The thermal conductivity of the coatings depended on the number and composition of the layers they consisted in. As a result the decrease of the thermal conductivity of multilayers is due to the increase in their total porosity resulting from the introduction of large numbers of pores between layers compared to a single layer,

EP

2- We compare the values of thermal conductivity obtained from the two multilayer

TE D

1- In light of various experimental investigation. It was found that the thermal conductivity

M AN U

SC

RI PT

scattering.

ACCEPTED MANUSCRIPT which is related to the nature of the non-uniform interface and inter-phases between layers. The increased porosity results in an increase in phonon scattering, which decreases the mean free path for phonon transport through the coating. As a conclusion, the CrAlN and Cr/CrN/CrAlN coatings with their low thermal conductivity

alternating stresses.

Acknowledgements

The authors wish to thank the Regional Council of Burgundy and EGIDE for their financial support, and also the technical staff of the Arts et Métiers ParisTech of Cluny: especially Romaric Masset and Pierre-Michel Barbier for the samples preparation.

References

[1] Y Chiba, T Omura, H Ichimura, J. Mater. Res. 8 (1993) 1109

[2] T Bin, Z Xiaodong, H Naisai, H Jiawen, Surf. Coat. Technol. 131 (2000) 391 [3] J A Sue, T P Chang, Surf. Coat. Technol. 76-77 (1995) 61

[5] H C Barshilia, A Jain, K S Rajam, Vacuum 72 (2004) 241 [6] H C Barshilia, M S Prakash, A Jain, K S Rajam, Vacuum 77 (2005) 169

[8] H C Barshilia, K S Rajam, Surf. Coat. Technol. 183 (2004) 174

AC C

[9] H C Barshilia, M S Prakash, A Poojari, K S Rajam, Thin Solid Films 460 (2004) 133 [10] H.C. Barshilia, K.S. Rajam, D.V. Sridhara Rao, Surf. Coat. Technol. 200 (2006) 4586 [11] H C Barshilia, K S Rajam, A Jain, K Gopinadhan, S Chaudhary, Thin Solid Films 503 (2006) 158 [12] M Okumiya, M Griepentrog, Surf. Coat. Technol. 112 (1999) 123 [13] P H Mayrhofer, H Willmann, C Mitterer, Surf. Coat. Technol. 146-147 (2001) 222 [14] R Wuhrer, W Y Yeung, Scripta Mater. 50 (2004) 1461

EP

[7] V K W Grips, V Ezhil Selvi, H C Barshilia, K S Rajam, Electrochem. Acta 51 (2006) 3461

TE D

[4] H C Barshilia, K S Rajam, Bull. Mater. Sci. 26 (2003) 233

M AN U

SC

RI PT

can be used as thermal barrier on a tool surface and therefore minimize crack-causing

ACCEPTED MANUSCRIPT
[15] M Kawate, A K Hashimoto, T Suzuki, Surf. Coat. Technol. 165 (2003) 163. [16] M Brizuela, A G Luis, I Braceras, J I Onate, J C S Lopez, D M Martinez, C L Cartes, A Fernandez, Surf. Coat. Technol. 200 (2005) 192 [17] M Uchida, N Nihira, A Mitsuo, K Toyoda, K Kubota, T Aizawa, Surf. Coat. Technol. 177-178 (2004) 627

[19] S Baragetti, G M La Vecchia, A Terranova, International Journal of Fatigue 27 (2005) 1541 [20] M Nordin, M Larsson, S Hogmark, Surf. Coat. Technol. 106 (1998) 234.

[21] Q. Wang, H. Ishikawa, S. Nakano , H. Ogiso, J. Akedo, Vacuum 75 (2004) 225–229. [22] T. Ghrib, B. Tlili, C. Nouveau,Y. Benlatreche, M. Lambertin, N. Yacoubi, M. Ennasri, Physics Procedia 2 (2009) 1327–1336 [23] J C Murphy, L C Aamodt, J. Appl. Phys. 51 (1980) 4580

[24] W B Jackson, N M Amer, A C Boccara, D Fournier, Appl. Opt. 20 (1981) 1333 [25] T Ghrib, S B Salem, Y Noureddine., Tribology International 42(3) (2009) 391 [26] T Ghrib, N Yacoubi, F Saadallah., Sensors and Actuators A 135 (2007) 346

[28] M Uchida, N Nihira, A Mitsuo, K Toyoda , K Kubota , T Aizawa, Surf. Coat. Technol. 177-178 (2004) 627

[30] Y J Kim, T J Byun, J G Han., Superlattices and Microstructures 45 (2009) 73

AC C

[31] H C Barshilia, K S Rajam, J. Appl. Phys. 98 (2005) 014311 [32] C Harish. Barshilia, B Deepthi, N Selvakumer, A Jain, K S Rajam., Applied Surface Science 253 (2007) 5076 [33] M Okumiyak, M Griepentrog., Surf. Coat. Technol. 112 (1999) 123 [34] L Wang, Z Guangan, R J K. Wood, S C. Wang, Q Xue., Surf. Coat. Technol. 204 (2010) 3517-3524 [35] O Vingsbo, S Hogmark, B Jönsson, A Ingemarson, American Society for Testing and materials, Philadelphia (1986) 257

EP

[29] J Musil, H Hrubý., Thin Solid Films 365 (2000) 104

TE D

[27] C Nouveau, M Djouadi, Rémy Marchal, Michel Lambertin., Mécanique & Industries 3 (2002) 333

M AN U

SC

RI PT

[18] T Ghrib, F Bejaoui, A Hamdi, N Yacoubi, Thermochimica Acta, 473(1-2) (2008) 86-91.

ACCEPTED MANUSCRIPT
[36] L Chen, Y Du, S Q Wang, J Li, International Journal of Refractory Metals & Hard Materials 25 (2007) 400 [37] Y Chunyan, T Linhai, W Yinghui, W Shebin, L Tianbao, X Bingshe, Applied Surface Science 255 (2009) 4033 [38] J Lin, B Mishra, J J Moore, W D Sproul, J A Rees, Surf. Coat. Technol. 201 (2007) 6960

Table captions
Table 1. Chemical composition of AISI4140 steel Table 2. Deposition conditions

Table 3. Composition, atomic ratios, total thicknesses and roughnesses of the mono and multilayers

Figure captions

Fig.1. Plasmas operated in the magnetron.

Fig.2. Schematic experimental setup of the PTD technique

Fig.3. Schematic representation of the probe beam deflection Fig. 4. Different regions crossed by the heat

Fig.5. XRD patterns of uncoated AISI4140 steel and multilayers such as CrN/CrAlN and Cr/CrN/CrAlN on AISI4140 samples

Fig.7. (A) Cross-section TEM image and (B) HRTEM image of a CrAlN film

AC C

Fig. 8. Residual stress of the films Fig.9. Normalized amplitude and phase evolution of the photo thermal signal vs the square root frequency of the deposited Cr layers Fig.10. Normalized amplitude and phase evolution of the photo thermal signal vs the square root frequency of the deposited Cr/CrN layers Fig.11. Normalized amplitude and phase evolution of the photo thermal signal vs the square root frequency of the deposited Cr/CrN/CrAlN layers.

EP

Fig.6. SEM cross section observation of a Cr/CrN/CrAlN coated sample

TE D

M AN U

Table 4. Thermal properties of the layers

SC

RI PT

[39] H-J Rätzer-Scheibe, U Schulz, T Krell., Surf. Coat. Technol. 200 (2006) 5636

ACCEPTED MANUSCRIPT

Table 1
C Mn Si S Cr P Mo 0.410 0.770 0.280 0.026 1.020 0.019 0.160

Table 2
Coatings Time of deposition (min) Bias Voltages (-V) Al Cr

M AN U
0 0 -300 // // // -900 0 -900 // // // 4 // // // // //

SC
Target power (KW) Gas mixture (4µbar) Ar (%) 100 80 // // // // N2 (%) 0 20 // // // //

AC C

EP

TE D

Cr CrN CrAlN Cr/CrN CrN/CrAlN Cr/CrN/CrAlN

60 60 60 51 105 99

RI PT

ACCEPTED MANUSCRIPT

Table 3
Coatings N Content (at. %) Al Content (at. %) Cr Content (at. %) Atomic ratio N/(Cr+Al) Total thickness (nm) Roughness ( µm) Roughness Arithmetic roughness average (Ra) quadratic (Rq) 0.1 0.12 0.1 0.12

Cr CrN CrAlN CrAlN/CrN/Cr CrAlN/CrN

0 48 48.2 51.8 52

0 0 4 4.2 5

98 51 48 42.9 41.3

0 0.95 1.04 1.2 1.2

860 690 980 1500 1500

Layer and substrate AISI4140 Cr CrN CrAlN CrN/Cr CrAlN/CrN/Cr

Thermal conductivity (W.m-1.K-1) 26 ± 0.01 93.9 ± 0.01 11 ± 0.015 2.8 ± 0.012

M AN U
Table 4
Thermal diffusivity (10-6.m2.s-1) 9 ± 0.012 0.79 ± 0.01 0.23 ± 0.013 0.052 ± 0.001

TE D

SC
Equivalent thermal conductivity (W.m-1.K-1) 13.1± 0.012 4.6 ± 0.013

AC C

EP

RI PT
0.1 0.12 0.1 0.13 0.12 0.11

ACCEPTED MANUSCRIPT

1

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Laser probe beam Heating Laser pump Mechanical Chopper Specimen Look-in amplifier
1,0

40 35 C r T heoretical curve
K1=93 D = ,9; 1 0,79.10 m .s
-4 2 -1

0,8

30 Phase (degree)
Cr Theo retical curve
K1=9 D1=0,79 m .s 3,9; .10
-4 2 -1

25 20 15 10 5

0,6

0,4

2

4 6 8 10 12 14 16 18 1/2 Squareroot m ulation freq od uency(H ) z

2

4 6 8 10 12 14 16 18 1/2 S ar ro m dulatio frequ qu e ot o n ency (Hz )

2

AC C

EP

TE D

M AN U

SC

Detector

RI PT

Normalized amplitude

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U
3

SC

RI PT

ACCEPTED MANUSCRIPT

Backing

Substrate

Layer (1)

Layer (2)

Layer (i)

Layer (n)

Fluid

4

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

600

Cr/CrN/CrAlN CrN/CrAlN AISI4140

(Intensity counts))

200

0 40 50

M AN U
60 70 80

2θ

(° )

5

AC C

EP

TE D

SC
90



400

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U
6

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U
7

SC

RI PT

ACCEPTED MANUSCRIPT

0,6 0,5 0,4

0,3 0,2 0,1 0,0
Cr CrN CrAlN

AC C

EP

TE D

M AN U
8

CrAlN/CrN CrAlN/CrN/Cr

SC

RI PT

σ r (-GPa)

ACCEPTED MANUSCRIPT

1,0

40 35

Cr Experim ental Theoretical curve
K1=93,9 W.m .K ; D1=0,79.10 m .s
-1 -1 -4 2 -1

Normalized amplitude

0,8

30 Phase (degree) 25 20 15 10
-4 2 -1

0,6 Cr Experim ental Theoretical curve
K1=93,9 W.m .K ; D1=0,79.10 m .s
-1 -1

0,4

5 2 4 6 8

2

16 18 1/2 Square root m odulation frequency (Hz )

4

6

8

10

12

14

16 18 1/2 Square root m odulation frequency (Hz )

SC
10 12

AC C

EP

TE D

M AN U
9

RI PT
14

ACCEPTED MANUSCRIPT

1,0

35
0,9 Normalized amplitude

30
0,8 0,7 0,6 0,5 0,4 0,3 4 16 18 1/2 Square root modulation frequency (Hz ) 6 8 10 12 14 CrN/Cr Experimental Theoretical curve
K1=93,9 W.m .K ; D1=0,79.10 m .s K2=11 W.m .K ; D2=0,23.10 m .s
-1 -1 -4 2 -1 -1 -1 -4 2 -1

CrN/Cr Experimental Theoretical curve
K1=93,9 W.m .K ; D1=0,79.10 m .s K2=11 W.m .K ; D2=0,23.10 m .s
-1 -1 -4 2 -1 -1 -1 -4 2 -1

Phase (degree)

25 20 15 10 5 4 6

16 18 1/2 Square root modulation frequency (Hz )

SC
8 10

AC C

EP

TE D

M AN U
10

RI PT
12 14

ACCEPTED MANUSCRIPT

1,0 0,9
Normalized amplitude

30 25 20 15 10 5
-1 -1

Phase (degree)

0,8 0,7 0,6 0,5 0,4
CrAlN/CrN/Cr Experimental Theoretical curve
K1=93,9 W .K ; D1=0,79.10 m .s .m -1 -1 -4 2 -1 K2=11 W .K ; D2=0,23.10 m .s .m K3=2.8 W .K ; D3=0,052.10 m .s .m
-1 -1 -4 -1 -1 -4 2 -1

Experimental Theoretical curve
-1 -1

K1=93,9W .K ; D1=0,79.10 m .s .m K2=11W .K ; D2=0,23.10 m .s .m
-1 -1 -4 -4 2 -1

2 -1

K3=2.8W .K ; D3=0,052.10 m .s .m

4 6 8 10 12 14 16 18 1/2 Square root modulation frequency (Hz )

4 6 8 10 12 14 16 18 1/2 Square root modulation frequency (Hz )

AC C

EP

TE D

M AN U
11

SC

RI PT
-4 2 -1 2 -1

CrAlN/CrN/Cr