Document Type : Research Paper
Authors
1 Laboratoire de développement de nouveaux matériaux et de leurs caractérisations, Université de Sétif, Algérie
2 Ecole Normale Supérieure de Sétif-ENSS, Algérie
3 Département de physique, Faculté des Sciences, Université de Sétif 1 Algérie.
Abstract
Keywords
INTRODUCTION
The alloy preparation by physical or metallurgical methods presents a large advantage of a fine composition and smooth control but, sincerely, with a high production cost. Therefore, due to the low fabrication costs, great purity of the deposits, and large choice of the piece shapes, the Electrodeposition make often the best technique to prepare many alloys and systems. In last decades, the number of papers on the electrodeposition of alloys has increased indicating a promising use of this film preparation method. Moreover, the deposition of ferromagnetic coating is thus one of the most important development of electroplating because of ever-increasing use of thin films in microelectronic, computers and data handling technology and some other fields.
Ni-Fe alloys have been intensively investigated during the last decade for their unique physical properties such as mechanical [1] and magnetic [2] properties, which was known to form soft magnetic coating. The composition of soft magnetic alloys and their properties have been reported in the earlier works of Morika and Tanhaschi [3], Wolf [4, 5], Wolf and Katz [6], Bozart [7], Tsu and Sallo [8], Smith et al. [9] and Avdeeva [10].
Such NiFe alloys are electrodeposited from Chloride, Sulphate, mixed or sulphamate baths were reported to be among the most interest nowadays because they yield deposits with the lowest internal stress [11]. It is well known that Ni displays a corrosion resistance greater than iron in most environments and that NiFe alloys show intermediary behavior.
New applications of Invar (Fe64Ni36) have been proposed, such as structural reinforcement in monument restorations and glass roofs repairs in historical and cultural buildings. [12]. Because of Ni-Fe alloys similarity to the Invar, many attentions have been devoted to them. Report that for this atomic composition, a change of the crystal structure (from bcc for the Fe-rich alloys to fcc for the systems with higher Ni concentration) appears [13, 14, 15]. The deposition of Ni-Fe alloy is called abnormal [16], this is because of the famous raison given the fact that the less noble element (iron) is deposited preferentially before the noblest (nickel) element; as well the amount of iron in the deposit will not be connected to its portion in the electrolyte. From a sulphate bath, Dahms et al. [17] have been studied electrodeposited NiFe on a rotating electrode. They have found that increasing in deposited iron amount is directly related to a decrease in the velocity of nickel deposition. Krause et al. [18] have showen that abnormal co-electrodeposition plating NiFe alloy is due, essentially, to iron hydroxide Fe(OH)2 precipitation at the surface of the electrode.
In view of the previously cited papers, we are aware that NiFe alloy is deeply investigated since the few last decades. However, to the best of our knowledge, there are few experimental works exploring the dependence in electrodeposition time of microstructure, thickness and electrical resistivity of Ni1-xFex/ITO/Glass, abnormal alloy. On the other side, We had ambition to employ unusual potential range between -1.45 V and -1.48 V, since the majority of old published reports using the same method as ours, usually do not exceed -1.35 V during their electrodeposition. This can reveal hidden properties of Ni1-xFex if samples are successfully elaborated. Moreover, we aim to contribute towards the understanding of existing correlation between different properties (structural parameters, electrical resistivity and morphological properties).
MATERIALS AND METHODS
Ni1-xFex thin films were prepared by electrodeposition onto commercial polycrystalline indium tin oxide (ITO) covered glass substrate which possess about 10 Ω/cm; (measured in area of 1×2 cm2 exposed to the electrolyte), as square resistance. All the depositions were made in a three electrodes cell containing Ni as a counter electrode, silver-silver chloride (Ag/AgCl) as reference and ITO coated glass as a working electrode. We have chosen the ITO substrate as cathode because of its high transparence and its properties as inert material. Before electrodeposition, ITO glass was first cleaned with acetone for 30 min and then alcohol for 30 min, finally rinsed in distilled water. A potensiostat/galvanostat system”parstat 2253” was utilized for preparing our alloy samples. In Fig. 1. We illustrate a real photo of the electrodeposition equipment that we have used with two others photos insets for the preparation process of Ni1-xFex thin films. However, visualization of V=f(I) allure is obtained via PowerSuite software that is implemented in the parstat 2253 . The electrolyte [0.1] (see In Table 1) was freshly prepared. The used Boric acid (adjusted to 2.7 using diluted H2SO4) was added to the bath in order to control the pH of the solution also to improve the quality of the deposit. In addition, Saccharin has been used as an organic additive to reduce the internal stress of deposits and to refine the grain structure [19]. Ni1-xFex thin films were deposited at a cathode potential of -1.45 (vs. Ag/AgCl). After carful test, we have deduced that the above study potential was the best ones to reach higher efficiency electrochemical deposition. For then, all time depositions were obtained at this potential. Electrodeposition was performed at room temperature, without stirring. Unlike several previous report we have successfully electrodeposited Ni1-xFex thin films onto semiconductor ITO substrate.
X-ray diffraction (XRD) analysis, using a Philips X’Pert diffractometer with CuKα (λ=1.54 Å) radiation is employed for determining structural properties. Veeco Dektak 150 surface profilometer using contact stylus profilometry measurement technique was employed to investigate film thicknesses, roughness and surface quality. Scanning Electron Microscope, high power optics Thermal Field Emission (SEM –TFE) JSM 7001F/ JSM 7001FA deliver surface images and film composition. Square (sheet) resistance and the electrical resistivity were inferred from a Van Der Paw four-point cell measurements set-up (using the ECOPIA HMS-3000 system from Bridge Technology).
RESULTS AND DISCUSSION
Structural analysis and electrodeposits composition
Structural analysis of electrodeposited Ni1-xFex/ITO/glass, films at room temperature was characterized by XRD technique. In first stage, for a fixed 600 s electrodeposition duration, four electrodeposits were prepared at -1.45, -1.46, -1.47 and -1.48 V electrochemical potentials. In second stage, we have chosen to fix electrochemical potentials at -1.45 V while changing electrodeposition duration each time as 600, 720, 840, 960, 1080, 1200 s to obtain six other Ni1-xFex/ITO/glass, films. Fig. 2.a shows the XRD patterns of our first stage study of Ni1-xFex films obtained for 600 s and deposited with -1.45, -1.46, -1.47 and -1.48 V, respectively. Indeed, the peaks marked with star (*) are assigned to the ITO coated conducting glass substrates. But, the four other peaks at 2θ values of 41.65°, 44.28°, 47.6°, and 51.50° might be attributed, respectively, to the textures <110> <200> <220> and <211> of the body centered crystal “bcc” polycrystalline adopted structure. Overall texture is dominated (110) reflexion with small co-existence of Invar alloy presented by (200) reflexion revealed in film elaborated particularly at 600 s (see Fig.2.b). This result is expected for our films with high Fe content. Our result about the overall texture is entirely consistent with that of Chang Su-wei et al [20]. The authors have suggested that the (110) orientation is preferred and dominant when the overall texture crystallizes in bcc structure while the (111) orientation is preferred and dominant when it is fcc crystallization [20-21].
As mentioned in Table 2, films composition take change in the range of 81–92 (wt.% Fe). According to I.Tabakovic et al [21] This compositional range show that dominant structures tilt into bcc Ni1-xFex alloys.
Indeed, changing time electrodeposition between 600s and 1200 s have not affected considerably Ni1-xFex films composition comparing to other effects cited in literature such as the precursor PH, annealing, mechanical alloying…etc. [22-23]. In Fig. 3. We have chosen EDS spectrum for film that was obtained at 600s as demonstrator sample which is characterized by the highest content of Fe.
Generally, crystallization tilt into the structure of dominating chemical element is such NiFe alloys. In addition, It was reported [21] that X-ray investigation of electrodeposited Ni1-xFex films at room temperature reveal the existence of fcc or γ phase in the range of 10-58 wt.% Fe, mixed fcc/bcc phase in the range of 59-63 wt.% Fe, and purely bcc or α phase in the range of 64-90 wt.% Fe.
A suddenly changing in crystalline texture (see Fig.2) at 1080 second from <110> into <211> is another fact which can be considered as a pointed result and might be connected with a new criterion of abnormal behaviour in our Ni1-xFex studied alloy. To the best of our knowledge, there is no report on Ni1-xFex alloy have mentioning this texture change before. That one paper of B. Ghebouli et al [24] which have reported texture tilt from <110> into <211> in pure evaporated Fe/glass layers. It is to be noted here that lattice parameter of pure iron in ICPDS file is 0.28664 nm close to our obtained ones since the content of Fe was greater than 80%. Xiaobai Chena et al [25] have discussed structural phase transformation and have found for sputter deposited NiFe thick films obtained with different composition that lattice parameter is 0.35866 nm for annealed Ni33Fe67 film at 753K and was found to be 0.28442 nm for the as-deposited Ni21Fe79 film.
Addionaly, we can note, here, that in spite of changing potential deposition, the Ni1-xFex alloy textures remained unchanged and <200> texture did not yet disappear. Generally, the increase in lattice parameters in Ni1-xFex alloy has been found leading to the solid solution formation or the amorphous phase formation [26]. With increasing time deposition, it can be seen that the lattice parameter of Ni1-xFex samples deviates slightly by less than 1% (about 0.55 %) during all deposition times (see Table 1). Literally, lattice parameter have not detected the disordering induced by electrodeposition whether when we enhance time or potential during electrodeposition. R. Hamzaoui et al [27] have suggested that the value of the lattice parameter take a small increases from a=0.286599±0.00005 to 0.287049±0.00005 nm which was resulting from 96 h of milling for Fe0.9Ni0.1 alloy and from a= 0.286589±0.00005 to 0.287069±0.00005 nm for Fe0.8Ni0.2 alloy.
Surface quality and roughness
Using contact stylus profilometry measurement technique allow review of surface quality and curvatures as well as thickness measurement. As illustrated in Fig. 4. Bi-dimensional surface scan show curvatures in deep and high for all samples. More curvatures are revealed within potentials changes especially at -1.47V (Fig.4.a). One can remark that there is no considerable a deep grooves or defects that can reflect structural or electrical properties worsening (Fig.4.a and b). All films have a mirror behaviour of good metallic character, no peels or bubbles on the top of all films surface. Several tests that we have carrying out before adoption of the present samples allow us to observe extreme conditions of the present electrodeposition. Beyond -1.47V and/or 1200 second electrodeposition duration, surface appear peeled and bubbled that’s why we have limited our conditions. Indeed, film surface at 1080 s tilted overall texture appear the smoothest with the less curvatures (see Fig.4.b), thus may relate to best structural stability. Surface roughness values obtained for various time electrodeposition are mentioned in Table. 3. Values changes between 75 nm and 138 nm that seems, for first view, to have an aleatory evolution versus electrodeposition time.
Grain size evolution under dominant texture tilt
The average grains size D in our films take a change in the general range of 48-106 nm (see Table 3). The increase of grain size in NiFe alloy can be assigned to coalescence and agglomeration since smaller grains. S. Sam et al [28] have result to the average grain size of about 16 nm for Ni81Fe19 compositional film and about 23 nm for Ni83Fe17 film. Whereas, C. Cheung et al [29] have suggested that electrodeposited Ni1-xFex Permalloy grain size were calculated to be 10 nm. In the case of those authors which they have suggest a bigger size of grains, we found T.E. Buchheit et al [30]. These last authors have reported twice that electrodeposited Ni1-xFex Permalloy Ni80Fe20 grain size was found to be less than 100 nm.
Two distinct case depending on dominant over all texture are clearly revealed. In first stage, when dominant texture was <110> from 600s to 960s, average grain size decrease considerably until reach the half approximately. After dominant over all texture tilting into <211> from 1080s to 1200s, average grain size increase suddenly and preserve increasing between 99 nm and 106 nm respectively (see Table 3). To highlight the granular size change induced by texture tilt <110> into <211> we have join tow SEM images as inset in Fig. 5. We have revealed that dominant texture tilt affect considerably the average grain size and not alloy compositional. This result disagreed with Sam et al [30] report. The authors claim that Ni1-xFex with Ni domination (permalloy) compositional affect remarkably the average grain size. Thus, may our result validation is restricted for NiFe alloy with Fe domination.
Resistivity evolution under dominant texture tilt
Using the direct contact of Van Der Paw measurement metallic probes applied on Ni1-xFex/ITO/Glass, obtained values of square resistances cannot be attributed to the Ni1-xFex films solely since the commercial ITO covering glass can conduct electricity. For measurement accuracy, the ratio: vertical/horizontal should be respected to be near to the perfect value of 1. To purify Ni1-xFex films we could proceed as follow:
When Rsq, Rsq1 and Rsq2 are, respectively, the total, Ni1-xFex thin films and ITO films coating glass support square resistances (all in Ω).
The dependence of Resistivity ρ on grains size D under experimental conditions of times deposition of our studied Ni1-xFex alloy films is shown in Fig. 5. It can be seen that the D decrease from 97.5 nm to 48.73 nm while ρ increase from 21, 20 Ω.cm to 92, 78 Ω.cm with increasing time electrodeposition from 600 to 960 s. Resistivity ρ can reaches a minimum value of about 1,35 Ω.cm at 1200 s while D increase, disproportionally, till reach the value of 106.02 nm.
Indeed, several phenomena can contribute in thin films electrical resistivity such as: diffusion by the surface, diffusion by the grain boundaries, impurities, defects and magnetic disorder, high preferred orientation. As was suggested by Matthiessen, the resistivity ρf can be written as follow:
Where ρT is due to the electron phonon interaction and ρr is the residual resistivity which is the sum of the various contributions:
Here, ρi, ρd, ρm, ρg and ρs denote, respectively, the contribution of the impurities, the defects, and the magnetic disorder, the diffusion by the grain boundaries and the diffusion by surface. Besides, we can suggest that eq.(2) can be written separately for each preferred orientation when more than one can be occurring during experience.
In addition, thermal annealing is another important factor that may affect the resistivity of Ni1-xFex alloys. Xiaobai Chen et al [29] have found in the case of alloy Ni33Fe67 that the values of resistivity was about 8.7 at room temperature and reach 2.7 (10-6Ωm) at 753°K with enhancing vacuum annealing.
It is known that if grain size becomes smaller, grain boundaries number increases. Consequently, diffusion at grain boundaries enhance, thus explains why electrical resistivity has raised in our films. The texture tilt (see Fig.2) beyond 1080 second from <110> into <211> might be another fact directly, assigned to the brutal electrical resistivity fall down. One can say that <211> texture could be characterized by an easier electrical diffusion process at grain boundaries that <110>.
Thickness evolution under dominant texture tilt
In Fig. 6. we present thicknesses variation of our films with respect to times electrodeposition. That increase monotonously between 520 nm and 1400 nm for 600 s and 1200 s respectively (see Table 3). Two stages, before/after texture tilt, can be distinguished for that evolution of thickness. In first stage, when dominant texture was <110> from 600s to 960s, we reveal a thickness gain of 200 nm during 360s electrodeposition time gain. While, for <211> tilted texture, gained thickness in about 440 nm just for only a supplementary 120 s electrodeposition time. At probably, NiFe alloy textured <211> yield more thickness growth than NiFe alloy textured <110>. Mathematically, a slow evolution can be traduced by logarithmic equation. That is why we propose to fit thicknesses versus electrodeposition times to the standard following equation
Where e and t are films thicknesses and times deposition respectively. A= 0.001 (in S-1) which have the frequency dimension and B= 5.135 are the fit constants. Fit was done within minimal (relatively neglegible) errors, which can be quoted (see Fig. 6). Sumalatha Vasam et al [31] have reached for NiFe permalloy the thickness of 28 µm using an electrochemical analyzer of CH Instruments (Model: CHI6087E) in Galvanostatic mode. This is twenty time superior than our highest reached thickness of 1.4 µm. May NiFe Permalloy electrodeposition encourage the growth much more than NiFe abnormal alloy.
Surface morphology under texture tilt
Fig. 7 displays Scanning Electron Microscope SEM images of surfaces morphology of Ni1-xFex films obtained at different electrodeposition times at room temperature. SEM images show clearly a various polycrystalline textures as was suggested by XRD analysis. The surface morphology was found to be greatly influenced by changing time deposition. This is clear to observe by comparing SEM images in figures, which show that all films have grain like structure. It is clearly confirmed, now, that the morphologies of Ni1-xFex films are significantly influenced by dominant texture tilt to <211> (see Fig.7.d and e). As example, at 1080 s, grains size are clearly larger while grain boundaries number is, consequently, decreased. Ni0.14Fe0.86, Ni0.19Fe0.81 and Ni0.15Fe0.85 films have elongated grains Fig.7.d,e and f while all first other Ni0.08Fe0.92, Ni0.16Fe0.84 and Ni0.15Fe0.85 films have rounded grains (Fig.7.a,b and c). Fig.7.a,b shows that in both Ni0.08Fe0.92, Ni0.16Fe0.84 films, small grains have coalesced to form a smoother film cross section. All films show good adhesion with ITO coated glass substrates.
Discussion of correlations and relationships
In Table 3 we have summarized results of roughness of our Ni1-xFex alloys films that take a change between 75-183 nm. Penny Tsay et al [32] have suggested for prepared films using Pulse-reverse electroplating technique, that when Fe becomes the main component, the roughness factor is generally increased and reaches the maximum value in case of pure Fe deposit. M. Saitou et al [33] have reported that preferred orientation in NiFe/ITO is strongly influenced by preferred orientation. We have remarked that increase in Fe contents does not meet an agreement with roughness values in our films. In Fig.8, it is clear that for all studied samples, roughness allure correlate well with electrical resistivity, as well with average grain size. This can be explained as follow: when grains become larger, the roughness must be lowered and electrical current density opposing alloy resistivity will traveling less grain boundaries then less obstacles during diffusion. Both of them, electrical resistivity and average grain size obey texture tilt from <110> onto <211>; thus, roughness also is indirectly depend on it. For the best of our knowledge, the roughness depends on the films growth mechanism not the composition of alloy. In similar context, we have remarked that compositional does not influence electrical resistivity and average grain size. Thickness values were depending on texture tilt. Their slow down increase allowed logarithmic fit that was established successfully. In Fig.8, we can confirm the close relationship between of resistivity and roughness. One can see that all details of roughness allure meet those of resistivity. Average grains size, resistivity, thichness and surface roughness was clearly justified by our observed SEM surface micrographs. But NiFe composition seems to be independent in this work that we sincerely think that can be managed by only one fact which, is the abnormal behavior of Ni1-xFex alloy.
CONCLUSIONS
In this work, we have using electrodeposition from a concentrated Fe and Ni sulfamate bath of 0.1 M to electrodeposit Ni1-xFex/ITO/Glass. Electrodeposition of samples was carrying out using various potentials and times. As changing potential between -1.45 and -1.48 V, we have electrodeposited about four successful samples of Ni1-xFex/ITO/Glass. Their investigation can be qualified as a routine study, that we have not focus a lot on them. Although we have used an unusual potential range between -1.45 V and -1.48 V. Whereas, changing electrodeposition time seems to be very important since we reveal an overall texture tilt from <110> onto <211>. Hence, microstructure, morphology, electrical resistivity, thickness and roughness are all remarkably affected. Our main results about the influence of texture change on the Ni1-xFex/ITO/Glass, structural, electrical and morphological properties are also summarized. Thus, the X-ray patterns studies of electrodeposited Ni1-xFex/ITO/Glass for various potentials and times at room temperature revealed the existence of bcc phase as the dominant structure. Enhancing electrodeposition potential did not affect Ni1-xFex global textures or the dominant one except some curvatures that may affect surface smooth of final samples. Whereas, enhancing electrodeposition time shows texture tilt from (110) to (211) revealed for first time at 1080s. This can make part of the general behaviour of Ni1-xFex abnormal alloy. After texture tilt, the rest of study was built on their influence that we have distinguished the cases before/after texture tilt. NiFe alloy textured <211> yield thicker NiFe films (960 to1400 nm) than the one textured <110> and average grain size become almost larger reached 99 to 106 nm. Diffusion at the grain boundaries seems to be the predominant factor controlling electrical resistivity and not directly the grain size. Lowest Resistivity ρ of about 1,35 Ω.cm at 1200 s was reached with tilted texture <211> making NiFe <211> more conductor that NiFe <110>. Indeed overall texture tilt have made NiFe alloy more performant and more interesting. Fe content (%wt) in electrodeposited Ni1-xFex films compositional have changed between about 82-92%. Average grain size, electrical resistivity, thickness and roughness seem to not correlate with electrodeposited Ni1-xFex films compositional. We prove that resistivity, grain size and roughness are highly correlated between each other’s and some relationships was presented and judged within available SEM images.
Our main perspective after carrying out the present work is to going through electrodeposition of Invar Fe64Ni36 with <200> overall texture using the same sulphate bath. This will be possible with optimizing and mastery some electrochemical parameters (such as precursor PH and concentration), since we have remarked co-existence of Invar alloy presented by <200> texture in films electrodeposited at 600 s wherever the used potential (see Fig.2.b).
ACKNOWLEDGMENTS
We would like to acknowledge La Direction Générale de la Recherche Scientifique et du Développement Technologique (DGRSDT), Algeria. Moreover, we send sincere thanks to Laboratory for Developing New Materials and their Characterizations, University of Setif1, Algeria.
CONFLICT OF INTEREST
The authors declare that there is no conflict of interests regarding the publication of this manuscript.