Journal of Food Engineering 64 (2004) 63–79 www.elsevier.com/locate/jfoodeng Modified stainless steel surfaces targeted to reduce fouling––surface characterization Olga Santos a,*, Tommy Nylander b, Roxane Rosmaninho c, Gerhard Rizzo d, Stergios Yiantsios e, Nikolaos Andritsos e, Anastasios Karabelas e, Hans M€ uller-Steinhagen d, Luis Melo c, Laurence Boulang
e-Petermann f, Christelle Gabet f, Alan Braem b, Christian Tr€ ag ardh a, Marie Paulsson a a Department of Food Engineering, University of Lund, P.O. Box 124, S-221 00 Lund, Sweden Department of Physical Chemistry II, University of Lund, P.O. Box 124, S-221 00 Lund, Sweden Department of Chemical Engineering, Faculty of Engineering, University of Porto, Rua Dr. Roberto Frias, s/n, 4200-465 Porto, Portugal d Institute for Thermodynamics and Thermal Engineering, University of Stuttgart, Pfaffenwaldring 6, 70550 Stuttgart, Germany e Chemical Process Engineering Research Institute––C.E.R.T.H., 6 km Charilaou––Thermi Road, P.O. Box 361, GR-57001 Thermi, Thessaloniki, Greece f Ugine and Alz––Groupe Arcelor, Centre de Recherches d’Isbergues, BP 15, F-62330 Isbergues, France b c Received 30 May 2003; accepted 13 September 2003 This paper is dedicated to the memory of Dr. Hans Visser Abstract The surface properties of several modified stainless steel samples were characterized according to their chemical composition, 2þ roughness, topography and wettability. The modifications tested were SiFþ 3 and MoS2 ion implantation; diamond-like carbon (DLC) sputtering; DLC, DLC–Si–O and SiOx plasma enhanced chemical vapor deposition (PECVD); autocatalytic Ni–P–PTFE and silica coating. X-ray photoelectron spectroscopy (XPS) and X-ray microanalysis were applied to determine the surface chemical composition. Atomic force microscopy (AFM) and stylus-type instruments were used for roughness determination, and the surface topography was imaged with AFM and scanning electron microscopy (SEM). The contact angle and surface tension were measured with the Wilhelmy plate method and the sessile drop method. For thick modified layers, only the elements of the coating were detected at the surface, whereas for thin layers the surface composition determined was that of the stainless steel substrate. The roughness of the 2R (cold rolled and annealed in a protective atmosphere) surfaces was not altered by the modification techniques (except for the Ni–P–PTFE coating), while for the 2B (cold rolled, heat treated, pickled and skinpassed) surfaces an increase in roughness was observed. The silica coating produced surfaces with consistent roughness, independent of which steel substrate was used. DLC sputtering and Ni–P–PTFE coating produced surfaces with the highest roughness. All modified surfaces revealed a similar surface topography with the exception of the Ni–P– PTFE coating, for which the coating masked the underlying steel topography. In terms of wettability, the SiOx -plasmaCVD and Ni– P–PTFE coating techniques produced the most hydrophilic and hydrophobic surfaces, respectively.
2003 Elsevier Ltd. All rights reserved. Keywords: Modified stainless steel; Fouling; Chemical composition; Roughness; Surface energy 1. Introduction Fouling of process equipment in the dairy industry has been one of the main challenges for researchers within food engineering, starting with the early work of * Corresponding author. Tel.: +46-46-222-98-08; fax: +46-46-22246-22. E-mail address: olga.santos@livstek.lth.se (O. Santos). 0260-8774/$ - see front matter
2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.jfoodeng.2003.09.013 Harold Burton in the 60’s (Burton, 1968). Already, these early studies established that protein deposition is one of the key factors in fouling of heat exchanger surfaces in the dairy industry. Protein adsorption onto solid surfaces results from a competition between different types of interactions involving the protein, the surface, the solvent and any other solute present in the system (Haynes & Norde, 1994). These interactions are significantly influenced by the surface properties. The 64 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 challenge is to determine which property is the important factor. Surface free energy of the solid–liquid interfaces, as defined by the Young equation, is one factor that can give some indication of the degree of foulant adsorption. However, a quantitative prediction relating surface energy and adsorbed amount is still lacking. Several reports can be found in the literature dealing with the influence of the substrate surface energy on the protein adsorption. For instance, Janocha et al. (2001) observed that the adsorbed amount of albumin decreased with increasing surface energy. Yoon and Lund (1994) also found higher milk fouling rates for the PTFE-coated plates in plate heat exchangers than for the 304 stainless steel plates with a high surface energy. However, in a study by Addesso and Lund (1997), a similar rate and extent of protein (a-La, b-Lg) adsorption was reported for surfaces with varying surface energies. Regarding the influence of surface energy on adhesion strength, Britten, Gree, Boulet, and Paquin (1988) observed weaker protein (raw milk) and phosphate adhesion forces on the lower surface energy surfaces. Bornhorst, Steinhagen, and Zhao (1999) found a similar correlation between CaSO4 deposits and solid surfaces. A critical surface tension of wetting, approximately 30 mN m
1 , at which protein adsorption/deposition is minimal, was reported by Baier and Meyer (1992) and later by Zhao and M€ uller-Steinhagen (2003). Also, properties like surface roughness, charge and charge density play an important role for protein adsorption. Wahlgren and Arnebrant (1990) observed higher adsorbed amounts of b-Lg on polysulfone than on methylated silica surfaces, which they attributed to the higher surface roughness of the polysulfone surface. The same trend was found in other studies for b-Lg adsorbed on silicon surfaces and on steel surfaces with increased hydrophobicity (Krisdhasima, McGuire, & Sproull, 1992; Santos, Nylander, Paulsson, & Tr€ag ardh, 2003). The effect of surface charge and charge density was studied by Norde, Arai, and Shirahama (1991), who found that on hydrophilic surfaces, structurally stable proteins adsorb only if electrostatic interaction is favourable, e.g. if the net charge of the proteins is opposite to the surface. However, on hydrophobic surfaces they are also adsorbed on charged surfaces that carry the same (net) charge, but to a lesser extent. For less stable proteins, where structural re-arrangements contribute to the tendency to adsorb, the adsorption on surfaces with the same charge can occur. Surface topography has also been found to affect bacterial adhesion. Boyd, Verran, Jones, and Bhakoo (2002) observed higher bacterial adhesion forces and cell retention after scanning with an AFM tip on abraded surfaces (unidirectional topography) as compared with either polished or unpolished (grain structure) surfaces. Fouling does not only impair the safety of the final product but also accounts for higher investment, pro- duction and maintenance costs. In spite of the many studies devoted to finding ways to avoid or reduce fouling, no major break-through has been reported that can provide the food industry with a sustainable and economic anti-fouling coating. The challenge is that stainless steel, the preferred material in food, pharmaceutical and biotechnology industry due to its resistance to corrosion, cleanability, strength and durability, is hard to modify. However, recent developments in surface treatment technology in material science have opened up new possibilities. These possibilities are explored in the European MODSTEEL project to produce stainless steel surfaces that are less prone to milk fouling, and in this paper we will present and discuss some of our major achievements regarding surface modification. We have applied and tested four different approaches for surface treatment of stainless steel surfaces: (1) Ion implanted surfaces, obtained by introducing 2þ SiFþ 3 and MoS2 ions. The modifying material is directly implanted into the base material to form a surface alloy. This approach is expected to lower the surface energy. (2) Diamond-like carbon (DLC) surfaces obtained by sputtering and plasma enhanced chemical vapor deposition (PECVD). Using this method, the aim is to produce a thin film on the surface. (3) Silica surfaces obtained by (a) SiOx also manufactured by PECVD. This will give a hard glass-like surface. (b) Silica coating by the sol–gel process. This approach is applied to produce a hydrophilic and hydrated anionic surface. (4) Creating a Teflon surface by autocatalytic Ni–P– PTFE deposition to obtain a hydrophobic ‘‘nonsticky’’ surface. Some of these modifications, namely SiFþ 3 implantation and DLC sputter coating, have previously proved to be efficient in reducing CaSO4 scale formation during pool boiling and bacteria attachment (Bornhorst et al., 1999; M€ uller-Steinhagen & Zhao, 1997; M€ uller-Steinhagen, Zhao, & Reiss, 1997). Bornhorst et al. (1999) obtained on these two modified surfaces final heat transfer coefficients that were 2–3 times higher as compared to the unmodified surface. In addition, the deposit layer formed on the modified surfaces was thinner and easier to remove. These modified surfaces, SiFþ 3 implanted and DLC sputtering, have also shown promising results regarding reduction of b-Lg adsorption (Santos et al., 2003). A 19% reduction of the adsorbed amount of b-Lg on the DLC sputtering surface was determined in a previous study (Santos et al., 2003). The focus of this work is on the interactions involving surface and foulant (protein, salt, bacteria). In order to understand these interactions, it is important to have a 65 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 good knowledge of the different properties of the modified surfaces. Therefore, the aim of the work presented in this paper was to characterize the modified surfaces mentioned above in terms of surface energy, chemical composition, roughness, and topography, for subsequent studies on adsorption/deposition of organic and inorganic material. Several techniques were used for this purpose. X-ray photoelectron spectroscopy (XPS) and X-ray microanalysis gave information about the chemical composition of the samples. The measurement of the surface roughness was performed by atomic force microscopy (AFM) and with stylus-type instruments. The topography of the surfaces was recorded by the use of microscopic techniques, such as AFM and scanning electron microscope (SEM). The surface energy (contact angle and solid/liquid interfacial tension) was determined using either the sessile drop or the Wilhelmy plate method. 2. Materials and methods 2.1. Materials Alkaline detergent RBS35 (NFT 72151–72190), containing a mixture of non-ionic and anionic surfactants, was obtained from RBS Chemical Products, Brussels, Belgium. The water used was deionised and passed through a Milli-Q Gradient water purification system (Millipore S.A., Molsheim, France). The unmodified surfaces tested were of type 316L with surface finish 2R (cold rolled, bright annealed) and 2B (cold rolled, heat treated, pickled and skinpassed), received from a European manufacturer of stainless steel. Their chemical compositions (in percentages), given by the manufacturer, are presented in Table 1. The 2þ surface modifications consisted of SiFþ 3 and MoS2 ion implantation; a DLC sputter coating; DLC, SiOx , and DLC–Si–O plasmaCVD thin films; an autocatalytic Ni– P–PTFE coating, provided by the University of Stuttgart (Germany) and Silica coating provided by CPERI (Greece). Two different batches of SiFþ 3 implanted surfaces, SiF and SiF FZ, prepared in different plants under identical conditions, were used. mersed in a 2.0% w/v detergent (RBS35) solution in distilled water at 50 C for 10 min; (2) rinsed with distilled water at 50 C for 10 min and (3) rinsed with distilled water at 20 C. 2.3. Surface modification techniques 2.3.1. Sputter technique Sputter coating is a physical vapor deposition (PVD) process, which take place at sub-atmospheric pressures. The surface to be coated (the ‘‘substrate’’) and the coating material (the ‘‘target’’) are arranged in the vacuum, as depicted in Fig. 1, or they may be moving past each other in case of larger substrates. The high voltage applied between substrate and target ionises the gas (usually argon, but other gases are possible) in the chamber until a plasma is ignited. In Fig. 1, the reactive gas is indicated by big black circles and the ionised gas by small black circles. In practice, both direct current (DC) and radio frequency (RF) voltages are used, with the target usually being the cathode. Due to their positive charge, the plasma ions are accelerated towards the cathode. On hitting the cathode surface, the plasma ions dislocate atoms from the target surface (indicated by grey circles in Fig. 1). These atoms can in turn be deposited on other parts of the surface. If reactive gases, e.g. N2 , are available in the vacuum chamber in addition to the inert plasma gas, they may react with the available ions and atoms. In this case, the technique is referred to as ‘‘reactive sputtering’’. If the desired final surface film does not adhere well, the substrate can be precoated with another material. DLC thin films are produced in this way with a TiN and then a 2.2. Cleaning procedure All stainless steel samples were cleaned with the commercial detergent RBS35 before each experiment. The procedure was as follows: samples were (1) im- Fig. 1. Schematic representation of sputtering equipment. Table 1 Chemical composition of the stainless steel 316L finishes used in this work Steel 316L 2R 316L 2B Elements (wt.%) Fe Cr Ni Mo Mn Si Cu C Others 66.5 66.8 17.7 17.5 11.1 11.1 2.1 2.1 1.7 1.6 0.52 0.55 0.14 0.13 0.03 0.02 0.24 0.20 66 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 TiC film formed in presence of mixed N2 gas and acetylene on top of a metal substrate before the final DLC layer is added. The DLC coating itself can be produced by sputtering of graphite targets or through a plasma CVD process. The advantages of sputter coatings are the hard and generally very adhesive films. 2.3.2. Plasma enhanced chemical vapor deposition Plasma CVD––or plasma enhanced CVD––is a chemical vapor deposition process where the chemical reaction is not initiated through thermal energy (as in normal CVD), but through formation of a plasma. This plasma is usually induced through RF or microwaves. Contrary to sputter techniques, where the coating material is introduced into the process as a solid target, CVD processes use gaseous or vaporous precursors. Typical gases are Cx Hx for DLC films or Hexamethyldisiloxan (HMDSO) for SiOx films. As with other film techniques, argon and other process gases are added to the vacuum chamber. In addition to the easier handling, Plasma CVD also offers the possibility to implement additional atoms into the coating. By adding a siliconcontaining precursor and oxygen to the working gases, it is, for example, possible to obtain DLC–Si–O coatings. Disadvantages of plasma CVD coatings are their reduced density and weaker adherence to the substrate material. 2.3.3. Directed ion implantation During directed ion implantation (or ion beam implantation), the surface is bombarded with highly accelerated ions, with an average energy of several 100 keV. The modifying material is directly implanted into the base material to form a surface alloy. Hence, there are no adhesion problems such as in case of chemical vapor deposition (CVD) or galvanic coatings. Furthermore, it has been found that the corrosion resistance of the substrate material is considerably improved through the implanted ions and/or the implantation process (Zettler & M€ uller-Steinhagen, 2000). The drawback is the high capital and operating costs of directed ion implantation processes. The investigated sample surfaces have been implanted with SiFþ 3 ions with an implantation energy of 200 keV. The ion concentration was 5 · 1016 ions/cm2 . In order to achieve homogenous modification, the ion beam scanned the surface in a regular pattern. 2.3.4. Turbulent ion implantation Similar to sputtering, a plasma is ignited in the vacuum chamber during turbulent ion implantation, but in this case the substrate is a cathode. However, the atoms impacting on the surface of the substrate have such a high energy that they can penetrate into the interior of the material as in direct ion implantation. In fact, atoms hitting the surface can push already deposited atoms further inside, leading to implantation depths of up to 100 lm, for initial surface layers thickness of 2–5 lm. Similarly to reactive sputtering, it is possible to introduce additional reactive particles to the vacuum chamber by turbulent ion implantation. This leads to a high flexibility in the selection of the coating materials. MoS2 particles were chosen in this study. Another advantage with this technique is that it is possible to implant on the inside of an object, e.g. tubes or valves. 2.3.5. Autocatalytic Ni–P–PTFE coating The Ni–P–PTFE coating is produced by an autocatalytic plating process (Nasser, 2000). The process contains five steps, as can been seen in Fig. 2. The first step is an alkaline cleaning bath, followed by a pickling process. After this, the substrate must be activated. This takes place in a third step, the galvanic deposition of nickel. The fourth step is the beginning of an autocatNi-P-PTFE coated sample uncoated sample 30 s ... 1 min 5 ... 10 min alkaline cleaning 1 ... 3 min 1h PTFE particles pickling galvanic Ni plating (2 ... 3 A/dm2) autocatalytic Ni-P plating Fig. 2. The autocatalytic Ni–P–PTFE coating process. autocatalytic Ni-P-PTFE plating O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 alytic reaction in which a Ni–P plating is deposited on the sample. After the Ni–P plating has reached a certain thickness, PTFE particles are added to the coating bath, to be incorporated in the Ni–P matrix. To be effective in reducing fouling, the outer surface of the Ni–P–PTFE plating should contain a high percentage of homogenously distributed PTFE particles. Up to now, the maximum achievable amount of PTFE in the outer surface is 20%. 2.3.6. Silica coating The procedures adopted for silica coating are based on the sol–gel process. Commonly, the sol–gel method uses metal alkoxide compounds as raw ingredients, which are hydrolysed in the presence of water and condensed to form M–O–M bonds (M ¼ Si, Ti, Zr, etc). After hydrolysis and condensation have proceeded to some degree (sol formation), a coating can be applied on the substrates by, for example, dip- or spin-coating. Finally, the coating structure is consolidated by high temperature annealing, where residual organic molecules are oxidized and removed. Sols were prepared from methyl–triethoxysilane (MTES):H2 O (molar ratio 1:3) solutions in EtOH at various dilutions (1:10–1:4), and annealed at temperatures in the range 200–500 C. Under these conditions, crack-free coatings on both surface finishes, 2R and 2B, could be obtained. The mechanical properties of the films, as measured by advanced nano-indentation techniques, are superior to those of the stainless steel substrate, in terms of elasticity modulus (415 vs. 215 GPa for the substrate) and yield stress (270 vs. 65 GPA for the substrate). The thicknesses of the various modified films are presented in Table 2. 2.4. Chemical composition techniques The corrosion resistance of stainless steels depends on the protective properties of their passive layer, which are determined by their physical and chemical characteristics. Since the thickness of the passive layer is of the order of 1–10 nm, depending on the case, sensitive Table 2 Thickness of the modified layers Modifications Thickness (lm) SiF implantation SiF FZ implantation MoS2 implantation DLC sputtering DLC-plasmaCVD DLC–Si–O-plasmaCVD SiOx -plasmaCVD Ni–P–PTFE coating Silica-sol gel 0.2 0.2 0.2 2 1 2 2.5 10 0.1–0.2 67 techniques for physical surface analysis are required. The 316 2R and 2B surfaces, as received and after the cleaning procedure, were analysed by XPS. The chemical composition of the modified and unmodified surfaces was measured quantitatively by X-ray microanalyses. 2.4.1. X-ray photoelectron spectroscopy All samples were analysed after a plasma cleaning treatment in order to reduce the surface contamination and to exhibit photo-peaks coming from inner layers. The argon plasma treatments were carried out in a March I Instrument Plasmod unit. The radio-frequency power was 50 W and the treatment time was 10 min. After any surface treatment, the samples were transferred within 3 min to the XPS instruments. XPS analysis was performed using a Vacuum Generator XR3E2 spectrometer employing a Mg Ka (1253.6 eV) achromatic X-ray source operated at 15 kV with a power of 300 W. Survey scan spectra were obtained with a pass energy of 30 eV for all samples, to determine what elements were present in the top 5 nm of the surface. The value of the take off angle between the surface and direction of electron detection was 90. Typical operating pressures were 10
10 torr and the analysed area was a 10 · 4 mm2 . All binding energies were referenced to the carbon C–H photo-peak at 285.0 eV. The atomic sensitivity factors used for quantitative determination were obtained from Scofield data. The background in the peak area computation was assumed to follow a Shirley behavior. The structure of the passive film is expressed by the thickness, the hydroxides/oxides ratio and the chromium/iron ratio. 2.4.2. X-ray microanalysis In SEM, an electron beam is scanned across a sample’s surface. When the electrons strike the sample, a variety of signals are generated and the detection of specific signals produces an image of the surface or a sample’s elemental analysis. Interaction of the electron beam with atoms in the sample causes shell transitions that result in the emission of an X-ray. The emitted X-ray has the energy characteristic of the parent element. Detection and measurement of that energy permits a quantitative analysis of the elemental composition of the sample with a sampling depth of 1–2 lm. Analysis of the constitutive elements on the sample surfaces was measured by energy dispersive spectroscopy (EDS) using an OXFORD ISIS 300 system. Quantitative microanalysis was performed at least at three different positions and the average composition was recorded. 2.5. Surface roughness techniques The surface roughness of the different samples was measured by AFM and by the stylus instruments, 68 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 Dektak3 ST profiler and Perthometer (micrometer scale). These techniques differ in probe sizes and consequently in their capability in resolving surface defects, with AFM being the one having the highest resolution (nanometer scale). 2.5.1. Atomic force microscopy The most commonly used roughness parameter, mean roughness Ra, was measured by AFM (Binning, Quate, & Gerber, 1986) for the different steel surfaces. Ra is defined as the mean value of the surface relative to the center plane and is calculated by Z Ly Z Lx 1 Ra ¼ jf ðx; yÞj dx dy LxLy 0 0 where f ðx; yÞ is the surface relative to the center plane and Lx and Ly are the dimensions of the surface. The surfaces 316 2R and 2B had the dimensions 1 · 1 cm2 and 0.6 and 1 mm thickness, respectively. The equipment was operated in contact mode in air. At least five readings were taken for each surface tested. The Nanoscope III AFM was equipped with an ultrasharp silicon cantilever (MikroMasch). The cantilever chip includes two triangular springs. The one used for most of the surfaces has a length of 290 lm, a width of 40 lm, a resonant frequency of 15 kHz and a force constant of 0.12 N/m. The reflective side is coated with aluminium. The tip has a radius of curvature less than 10 nm and a height between 15 and 20 lm. The force applied to the cantilever was kept constant at 6 nN (50 nm cantilever deflection). 2.5.2. Stylus-type instruments 2.5.2.1. Dektak3 ST profiler. The Dektak3 ST (VEECO, Santa Barbara, USA) profiler was used to characterize the surface roughness of steel samples. Measurements are made electromechanically by moving the sample (by moving the stage on which the sample is placed) beneath a diamond-tipped stylus, which is coupled with a differential transformer. The radius of the diamond stylus is 2.5 lm and the stylus force was selected at 0.3 mN (30 mgf) for most surfaces. This is the force or effective weight of the stylus suitable for hard surfaces, whereas for soft surfaces this force should be significantly lower. Measurements were also carried out with a stylus force of 10 mgf, without noticing any roughness increase. The Ra values reported are the mean value of at least five different traces of a scanning distance of 400 mm. The measurements were made with a scanning direction perpendicular to the texture lines. 2.5.2.2. Perthometer. A stylus instrument Perthometer rk5 (MAHR, Germany) was also used to measure the surface roughness of the different samples. Although the method is simple to apply, the roughness is registered only along one direction in the surface plane. Hence the direction of the instrument placement on the sample must be chosen carefully. Because of the rolling process used, the plates as well as the samples show a significant straight texture. In the present study the measurements were always performed perpendicular to the texture lines. For all surface texture measurements, the same tracing length of 5.6 mm with a cutoff of 0.80 mm was chosen. The radius of the diamond tip of the used perthometer is 5 lm and the stylus force was for all measurements 0.8 mN. The profile resolution is equal to 12 nm. 2.6. Topographic techniques 2.6.1. AFM and SEM The AFM (Nanoscope III) and SEM (NORANVOYAGER and JEOL 6300) were also used to image the topography of the unmodified and modified stainless steel surfaces. 2.7. Contact angle measurement techniques The contact angles of the different surfaces were measured by dynamic and static methods. The dynamic measurements were performed with the Wilhelmy plate technique and the static measurements by using the sessile drop technique. In the Wilhelmy plate method the measured thermodynamic property is the wetting tension, s, which is related to the contact angle, H, by s ¼ cl cos H, where cl is the liquid surface tension. In the sessile drop technique the contact angle is directly obtained. The total interfacial surface tension between a solid s and a liquid l ðcTOT Þ can be expressed in terms of sl TOT the total surface tension of the solid ðcTOT sv Þ, liquid ðclv Þ and the contact angle by the use of Young’s equation, cTOT ¼ cTOT
cTOT cos H. It should be noted that the sl sv lv phases must be in mutual equilibrium, and thus the solid surface must be in equilibrium with the saturation vapor pressure of the liquid. The value of cTOT therefore consv tains the pressure of the already adsorbed film of the liquid on the surface (Adamson & Gast, 1997). In the case of low energy materials, this pressure has been suggested to be negligible. On the contrary, for high energy surfaces like metals or oxides, the spreading pressure is important. In these cases, the use of the solid–liquid–vapor method can lead to errors in the calculation of surface tensions. The solid–liquid–liquid method offers an alternative way to evaluate the surface tension of high energy surfaces. The contact angle of the liquid drop is measured in a hydrocarbon medium. Also, in this case the liquids must be in equilibrium with each other. Several semi-empirical models to extend the Young’s equation have been proposed (Adamson & Gast, 1997). Two different approaches, the van Oss approach (van Oss, 1994) in which the total surface tension is expressed as the sum of an apolar Lifshitz–van 69 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 der Waals component ðcLW Þ and an acid–base polar component ðcAB Þ, and Owens and Wendt approach (Owens & Wendt, 1969) that considers the total surface tension as the sum of a dispersive component ðcd Þ and a non-dispersive, acid–base or polar component ðcp Þ, were used in this work. In both approaches, the solid–liquid– vapor method was employed in which the contact angle of the liquid drop is measured under a vapor medium. In the Owens and Wendt approach for the solid–liquid– liquid contact angle, the adsorption of the vapor phase on the solid surface is controlled by replacing this phase by a neutral liquid (commonly an alkane, a) with only dispersive contribution. The relevant equations used in each approach and methods are summarized in Table 3. The surface tension data of the liquids used are shown in Tables 4 and 5. 2.7.1. Wilhelmy plate method In the Wilhelmy plate method, the sample is held by an electrobalance and is then immersed and retracted, at a constant speed, into and out of the liquid contained in a beaker. During these cycles the force acting on the plate vs. depth of immersion are recorded. The meniscus formed at the solid–liquid interface is characterized by the contact angle. With this technique, two contact angles are measured. As the surface moves down, advancing into the liquid, an advancing contact angle is obtained, and as the surface moves up, receding from the liquid, a receding contact angle is obtained. The hysteresis is related to the roughness and heterogeneity of the surface. Neglecting the weight of the plate and the viscous force, the contact angle is obtained from the equation presented in Table 3, where the wetting tension, s, vs. immersion depth, x, is obtained from F ðxÞ=P , where F ðxÞ is the total force measured on the sample and P is the sample’s perimeter. The sample surfaces 316 2R and 2B had the dimensions of 3 · 1 cm2 and thickness of 0.6 and 1 mm, respectively, and were suspended from the microbalance of a DST 9005 Dynamic Surface Tensiometer (Nima Technology, Coventry, UK). The liquid used was MilliQ water. The plate was moved at a constant speed of 2 mm/min and three immersion–retraction cycles at a temperature of approximately 20 C were performed. Table 4 Surface tension values (in mN/m), for the liquids used in the different models Water Formamide a-Bromonaphtalene Diiodomethane cTOT l d cLW l =cl cþ l c
l p cAB l =cl 72.8 58 44.4 21.8 39 44.4 25.5 2.28 0 25.5 39.6 0 51 19 0 50.8 50.8 0 Table 5 Surface tension values (in mN/m), of the alkanes used in the Owens and Wendt model for the solid–liquid–liquid method Heptane Decane Hexadecane cTOT a cTOT la 20.6 23.4 26.7 51.6 51.0 51.2 2.7.2. Sessile drop method 2.7.2.1. Solid–liquid–vapor method––approach of van Oss. In this approach, as mentioned above, the surface tension is expressed as the sum of an apolar Lifshitz–van der Waals component ðcLW Þ and an acid–base polar component ðcAB Þ. The Lifshitz–van der Waals interactions arise due to three distinct interactions: induction (Debye), orientation (Keesom) and dispersion (London), the last one being the most significant term (van Oss, 1994). The acid base component ðcAB Þ consists of two non-additive parameters, one for the electron donor ðc
Þ and one for the electron acceptor ðcþ Þ contribution (van Oss, 1994). cAB ¼ 2ðcþ c
Þ 1=2 ð1Þ Combining the Young’s equation with the total interfacial tension equation (Table 3), a relation between the measured contact angle and the solid and liquid surface tension terms can be obtained: h
i


þ 1=2 LW LW 1=2 þ
1=2 ð1 þ cos HÞcTOT þ c c ¼ 2 c c c þ c s l l s l s l ð2Þ The contact angle values for water, formamide and abromonaphtalene (a-BR) on the different surfaces were Table 3 Summary of the equations used for contact angle and surface tension evaluation, according to the different methodsa Technique Method Approach Solid–liquid–vapor Van Oss Solid–liquid–liquid Owens and Wendt Wilhelmy plate Sessile drop Owens and Wendt a See the text for details. Equation s ¼ cl cos H h
i

þ 1=2
1=2 LW 1=2 þ c

2 cþ
2 cLW þ cTOT ¼ cTOT cTOT s cl s cl s cl l s sl
1=2
1=2
2 cps cpl
2 cds cdl þ cTOT ¼ cTOT cTOT l s sl

1=2 1=2
2 cps cpl
2 cds cdl þ cTOT ¼ cTOT cTOT l s sl

p p 1=2 d d 1=2 TOT TOT TOT
2 cs ca csa ¼ cs þ ca
2 cs ca 70 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 measured in a contact angle meter (DataPhysics OCA15 Plus) using an image analysing system. 2.7.2.2. Solid–liquid–vapor method––approach of Owens and Wendt. In the Owens and Wendt model, the polar contribution is expressed differently and the equivalent of Eq. (2) becomes
1=2  TOT  1=2 1=2 1=2 cl ð1 þ cos HÞ 2ðcdl Þ ¼ ðcps Þ cpl =cdl þ ðcds Þ ð3Þ For contact angle measurements using refer  several d 1=2 ence liquids, a plot of cTOT ð1 þ cos HÞ =2ðc Þ against l l
1=2 according to this model should give a straight cpl =cdl 1=2 1=2 line with slope ðcps Þ and intercept ðcds Þ . 2.7.2.3. Solid–liquid–liquid method––approach of Owens and Wendt. In this method, the total interfacial surface tension between the solid and the alkane ðaÞ must also be considered. Combining the equations in Table 3 with Young’s equation, and assuming that the polar contribution to the surface tension of the alkane is negligible, ðcpa ¼ 0Þ gives
cTOT þ cTOT cos H cTOT l a la i h
1=2 1=2 1=2 1=2 Þ þ 2 cps cpl ¼ 2ðcds Þ ðcdl Þ
ðcTOT a ð4Þ For contact angle measurements using several n-alkanes, a plot of cTOT
cTOT þ cTOT cos HðyÞ against l a la d 1=2 TOT 1=2 ðcl Þ
ðca Þ ðxÞ gives according to this model a d 1=2 straight line, y ¼ ax
þ b, with slope a ¼ 2ðcs Þ and inp p 1=2 . tercept b ¼ 2 cs cl ¼ cds þ Consequently, cds ¼ a2 =4, cps ¼ b2 =4cpl and cTOT s p cs can be calculated. 3. Results and discussion 3.1. Surface chemical composition The results of XPS analyses of the unmodified surfaces are reported in Table 6, and show that the passive film on the 2B stainless steel surface is thicker and contains more chromium than the passive film on the 2R finish. This can be caused by the pickling operation on the 2B surface. It should be born in mind that a pure metal/metaloxide surface is very susceptible to contamination from the atmospheric air during processing and handling. Both 2B and 2R finish surfaces are therefore contaminated, as determined from the carbon signal from XPS analyses. These analyses could detect the presence of carbon down to an apparent depth of about 2.1 nm for both surface finishes. Here it is noted that the roughness of these surfaces are 30 and 70 nm for 2R and 2B finish respectively (see Section 3.2). Therefore it is difficult to determine the extent of the contamination. The alkaline detergent cleaning leads to a decrease in both the iron/chromium ratio and the passive film and the apparent contamination film thickness (for instance, the apparent thickness of the contamination film on the 316 2B sample is 2.1 nm before cleaning and 1.4 nm after cleaning). X-ray microanalyses were performed on both unmodified and modified samples and the EDS spectra obtained are presented in Figs. 3 and 4. The layer analysed by this technique is about 1 lm deep, which allows some conclusions to be made about the thickness of the modified layer. For thick coatings (of the order of 1 lm and thicker) only the elemental composition of the coating layers was measured (Figs. 3(d, g and h) and 4(d, e and f)). The elemental composition of 316L stainless steel as well as the surface layer composition is apparent in the spectra for thin coated or ion implanted surfaces (Figs. 3(a–f) and 4(a–c)). The presence of argon was observed on the spectrum of the 2R DLC–Si–O–plasmaCVD surface (Fig. 3(h)), apparently due to the presence of this gas in the coating process, although its content was rather high (Table 7). The analysis on both DLC sputtered surfaces, 2R and 2B, revealed only the presence of titanium (Figs. 3(g) and 4(f); Table 7). This can be related to the sampling depth of the technique (about 1–2 lm) and the composition of this modification, which consists of several layers (total thickness of 2 lm), the first one being titanium. The molybdenum content of the 2R and 2B MoS2þ 2 surfaces was similar to the content of the unmodified surface. This can be explained by the low concentration of molybdenum used in the modification process together with the low resolution of the technique making its measurement difficult. 3.2. Surface roughness The average roughness for the different modified surfaces was measured by AFM and by the stylus in- Table 6 Characteristics of the passive and contamination films, before and after the cleaning procedure, on 316 2B and 2R finishes, as determined by XPS Samples 316 316 316 316 2B 2B + cleaning 2R 2R + cleaning Passive film Contamination film Thickness (nm) Fe/Cr Hydroxide/oxide Apparent thickness (nm) 3.2 2.8 2.4 1.6 0.7 0.1 1.1 0.6 1.2 1.3 1.2 0.55 2.1 1.4 2.1 1.4 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 71 Fig. 3. EDS spectra of modified and unmodified 2R surfaces: (a) unmodified; (b) SiF ion implanted; (c) MoS2 ion implanted; (d) Ni–P–PTFE coated; (e) SiOx -plasmaCVD; (f) silica coated; (g) DLC sputtered; and (h) DLC–Si–O-plasmaCVD. struments, Dektak3 ST and Perthometer. The results are presented in Table 8. As these methods operate on different length scales, it is interesting to note that the Ra values give similar trends although they are quantitatively different. The Ra values (in the nanometer scale) obtained from AFM, having the highest resolution, are also the lowest. Due to strong adhesive forces, arising from van der Waals attractive force between the silicon tip and surface, measurements of surfaces modified with silicon ions posed some problems. A smaller cantilever with a higher spring constant (2 N/m) was therefore used for these surfaces. No such problem occurred with the Dektak3 ST profiler, where the dimension of the stylus is about 100 times larger than the AFM tip. Modified and unmodified 316 2R surfaces have a very similar Ra, approximately 30 nm, when measured by AFM and Dektak3 ST profiler, or 40 nm when using the stylus instrument Perthometer. The exception is the Ni–P–PTFE coating, which gave a significantly rougher surface. The modified 316 2B surfaces exhibit a higher 72 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 Fig. 4. EDS spectra of modified and unmodified 2B surfaces: (a) unmodified; (b) SiF ion implanted; (c) MoS2 ion implanted; (d) Ni–P–PTFE coated; (e) SiOx -plasmaCVD; (f) silica coated; and (g) DLC sputtered. roughness as compared to the unmodified one. The exceptions are the silica surface, which had a significantly lower roughness, and the MoS2þ 2 surface, with a similar roughness. All investigated surface modifications prepared on 2B finish stainless steel samples were rougher than the ones produced on the 2R finish. Depending on the technique used, the 2B unmodified surface has a roughness 2–4 times higher than the 2R unmodified surface. For the modified surfaces, 2B has a roughness of one order of magnitude higher than the corresponding 2R surfaces. This is consistent with the pickling treatment that the 2B surface had been exposed to, which is expected to give a rougher surface. The Ra values obtained by AFM and Dektak3 ST profiler for the 2B surfaces (Table 8) show that the modification increase the roughness of the surface 316 2B by a factor of 3, while 2R is unaffected in this respect. 73 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 Table 7 Chemical composition of the unmodified and modified steel surfaces 316 2R and 2B, as determined by energy dispersive spectroscopy Samples Elements (wt.%) Fe Cr Ni Mo Si Ti P Ar Others 2R Unmodified SiF ion implanted MoS2 implanted Ni–P–PTFE coated SiOx -plasmaCVD Silica coated DLC sputtered DLC–Si–O-plasmaCVD 67.4 ± 0.7 68.2 ± 0.6 67.5 ± 0.6 0 46.6 ± 0.4 69.3 ± 0.7 0 54.9 ± 0.4 18.4 ± 0.3 17.6 ± 0.3 18.1 ± 0.2 0 13.2 ± 0.4 15.7 ± 0.3 0 16.4 ± 0.2 11.2 ± 0.3 10.8 ± 0.3 11.2 ± 0.2 90.5 ± 0.5 7.3 ± 0.3 11.5 ± 0.5 0 8.8 ± 0.3 2.4 ± 0.2 2.2 ± 0.2 2.5 ± 0.3 0 1.5 ± 0.2 2.4 ± 0.2 0 2.2 ± 0.2 0.6 ± 0.1 1.2 ± 0.2 0.7 ± 0.1 0 31.4 ± 0.6 1.1 ± 0.2 0 14.9 ± 0.2 0 0 0 0 0 0 100 0 0 0 0 9.5 ± 0.5 0 0 0 0 0 0 0 0 0 0 0 2.8 ± 0.1 C C, C, C, O C, C 2B Unmodified SiF ion implanted MoS2 implanted Ni–P–PTFE coated SiOx -plasmaCVD Silica coated DLC sputtered 69.7 ± 0.6 68.3 ± 0.7 68.5 ± 0.3 0 47.1 ± 0.3 69.2 ± 0.6 0 16 ± 0.5 17.5 ± 0.3 17.5 ± 0.3 0 13.1 ± 0.3 15.6 ± 0.3 0 11.5 ± 0.2 10.7 ± 0.4 11.0 ± 0.2 90.9 ± 0.7 7.5 ± 0.2 11.2 ± 0.5 0 2.4 ± 0.1 2.1 ± 0.2 2.4 ± 0.2 0 1.8 ± 0.1 2.3 ± 0.2 0 0.4 ± 0.05 1.4 ± 0.2 0.6 ± 0.1 0 30.6 ± 0.3 1.7 ± 0.3 0 0 0 0 0 0 0 100 0 0 0 9.1 ± 0.5 0 0 0 0 0 0 0 0 0 0 C C, F C, O, S C, F Table 8 Ra values (in nm) of the unmodified and modified steel surfaces 316 2R and 2B, as determined by AFM and by the stylus instruments, Dektak3 ST profiler and Perthometer Samples 2R Unmodified SiF implanted MoS2 implanted Ni–P–PTFE coated SiOx -plasmaCVD Silica-sol gel DLC sputtered DLC-plasmaCVD DLC–Si–OplasmaCVD SiF Fz implanted 2B Unmodified SiF implanted MoS2 implanted Ni–P-PTFE coated SiOx -plasmaCVD Silica-sol gel DLC sputtered AFM 30 ± 2 24 ± 3 25 ± 3 57 ± 4 23 ± 3 35 ± 18 30 ± 1 28 ± 5 27 ± 2 26 ± 4 67 ± 9 247 ± 18 75 ± 4 150 ± 8 206 ± 48 37 ± 6 267 ± 23 Dektak3 ST 30 ± 5 30 ± 5 45 ± 5 75 ± 25 58 ± 23 30 ± 5 45 ± 10 – 26 ± 4 – 88 ± 13 225 ± 25 175 ± 25 230 ± 50 200 ± 20 75 ± 5 450 ± 100 Perthometer 40 70 50 200 50 – 90 50 – – 160 210 160 220 150 – 480 The Ra values obtained for the unmodified and modified surfaces with the stylus instrument Perthometer were similar, except for the 2R Ni–P–PTFE coated and 2B DLC sputtered surfaces, which were rougher. Film thickness measurements were also performed by profilometry on silica coated microscope glass slides, prepared under identical conditions as the steel samples coated with the silica sol–gel technique. In general the silica sol–gel coating technique seemed to give similar roughness and type of modifications for F S F O C, O C the 2B as for the 2R stainless steel surface treatment. DLC sputtering and autocatalytic Ni–P–PTFE coating led to the roughest. This can be explained by the modification procedure. The DLC sputter coating requires a multilayer adhesive film, which can cause an increase of the roughness. For the Ni–P–PTFE coating the surface roughness is influenced almost only by the process parameters of the autocatalytic baths. 3.3. Topography The unmodified 2R sample exhibit a unidirectional surface topography in which some holes (defaults) are observed, while the 2B sample reveals grain boundaries that appear as valleys or cracks (Figs. 5–8(a)). This different topography results from the different production method. 2R stainless steel type is produced in a non-oxidizing annealing atmosphere, which gives a bright surface with a more homogeneous appearance. On the other hand, the 2B type is produced in an annealing oxidizing atmosphere, which gives thick oxide layer. This layer is removed in the pickling operation, which attacks in the grain boundaries and in turn gives a heterogeneous grainy superficial structure. The DLC-plasmaCVD; DLC–Si–O-plasmaCVD and 2þ SiFþ 3 , FZ and MoS2 implantation techniques did not affect the steel topography for either surface finish. On the other hand, the Ni–P–PTFE coating masked all surface characteristics of the steel support (Figs. 5–7 and 8(c)). The images obtained for the 2R SiFþ 3 , SiOx -plasmaCVD and DLC sputtered surfaces (Figs. 5(b, d, f) and 7(d)), revealed a more heterogeneous topography with some cracks present along the surface, as well as 74 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 Fig. 5. AFM height images and profiles of the surfaces 316 2R: (a) unmodified; (b) SiF ion implantation; (c) Ni–P–PTFE coated; (d) SiOx plasmaCVD; (e) silica coated; and (f) DLC sputtered. particles with a diameter larger than 100 nm. The profiles obtained for the silica coated surface (Figs. 5 and 6(e)), indicate a much smoother surface than the untreated surface. This is in accordance with the roughness values obtained in Section 3.2. The topography of 2B SiOx -plasmaCVD and DLC sputtered surfaces, as imaged by AFM, was different compared to the unmodified surface and featured particles with a diameter larger than 500 nm (Fig. 6(d, f)). The SEM images acquired for the 2B surfaces (Fig. 8) reveal that with the exception of the Ni–P–PTFE coating, all other modifications left the topography unchanged, although the grain boundaries seemed to be more pronounced. In conclusion, the only modification that significantly changes the morphology of the steel surface is the Ni–P– PTFE coating. O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 75 Fig. 6. AFM height images and profiles of the surfaces 316 2B: (a) unmodified; (b) SiF ion implanted; (c) Ni–P–PTFE coated; (d) SiOx -plasmaCVD; (e) silica coated; and (f) DLC sputtered. 3.4. Wettability 3.4.1. Contact angle The contact angle values obtained with the Wilhelmy plate and sessile drop techniques are presented in Tables 9 and 10, respectively. In general, the hydrophobicity of the different modified surfaces increases according to their water contact angle in the order: SiOx
plasmaCVD < SiFþ 3  DLC-plasmaCVD < Silica-sol gel  DLC sputtered  DLC–Si–O-plasmaCVD < Ni–P–PTFE The water contact angle of the MoS2þ 2 ion implanted surface was dependent on the techniques used. When 76 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 Fig. 7. SEM images of some modified and unmodified 316 2R samples: (a) unmodified; (b) SiF ion implanted; (c) Ni–P–PTFE coated; (d) SiOx plasmaCVD; and (e) silica coated. Fig. 8. SEM images of some modified and unmodified 316 2B samples: (a) unmodified; (b) SiF ion implanted; (c) Ni–P–PTFE coated; (d) SiOx plasmaCVD; and (e) silica coated. 77 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 Table 9 Water contact angle (advancing and receding) and hysteresis values (in degrees), for the different steel samples as measured by the Wilhelmy plate technique Modifications Unmodified SiF implantation DLC sputtering DLC-plasmaCVD Ni–P–PTFE coating SiOx -plasmaCVD MoS2 implantation 316 2R 316 2B Adv. Rec. Hyst. 69 ± 1 61 ± 10 84 ± 3 65 ± 1 107 ± 5 42 ± 3 49 ± 9 14 ± 1 22 ± 5 19 ± 5 17 ± 1 31 ± 1 13 ± 1 12 ± 5 55 39 65 48 76 29 37 Table 10 Contact angle values (in degrees), for the different steel samples as measured by the sessile drop technique Samples HH2 O HFormamide Ha-BrNa 2R Unmodified SiOx -plasmaCVD SiF implanted DLC sputtered Ni–P–PTFE coated Silica-sol gel MoS2 Implanted DLC–Si–O-plasmaCVD SiF FZ implanted 24 ± 4 37 ± 3 31 ± 7 67 ± 4 114 ± 1 65 ± 6 65 ± 6 70 ± 1 42 ± 4 20 ± 5 25 ± 8 25 ± 6 50 ± 2 92 ± 2 49 ± 2 53 ± 4 44 ± 3 31 ± 2 13 ± 3 22 ± 3 14 ± 5 9±3 61 ± 1 39 ± 1 37 ± 1 16 ± 1 29 ± 2 2B Unmodified SiOx plasma-CVD SiF implanted DLC sputtered Ni–P-PTFE coated Silica-sol gel MoS2 implanted 83 ± 1 15 ± 3 49 ± 4 59 ± 4 111 ± 3 61 ± 1 64 ± 3 74 ± 1 12 ± 1 39 ± 2 33 ± 5 89 ± 7 44 ± 3 50 ± 3 15 ± 2 14 ± 2 21 ± 4 12 ± 1 69 ± 2 37 ± 1 18 ± 2 measured with the Wilhelmy plate its water contact angle was similar to the one obtained for the SiOx plasmaCVD surface, and when using the sessile drop it was similar to the one obtained for the silica surface. It is noteworthy that the same modification technique produces different surface characteristics depending on the type of stainless steel used (Tables 9 and 10), where the unmodified 2R finish surface is more hydrophilic than the unmodified 2B finish surface. The fact that the same modification technique gives rise to different surface properties depending on the base substrate was already observed in terms of surface roughness (Section 3.2). For this reason, to analyse the effect of the modification treatments, the samples were distinguished according to their stainless steel type substrate. For the 2R based surfaces, all modification techniques, except DLC sputtering and Ni–P–PTFE coating, gave a decrease in the advancing water contact angle values (Table 9). This is in contrast with what was determined by the sessile drop technique where all modified 2R surfaces exhibited a higher water contact angle value compared to the Adv. 99 ± 2 79 ± 6 78 ± 3 – 104 ± 4 45 ± 5 46 ± 9 Rec. Hyst. 21 ± 7 12 ± 1 0 – 34 ± 2 13 6±1 78 67 78 – 70 32 40 unmodified surface (Table 10). For the 2B based surfaces, all techniques, except the Ni–P–PTFE coating, resulted in a decrease of the water contact angle. For the Ni–P–PTFE coated surfaces both techniques gave similar water contact angle values, but for the other modified surfaces, the values obtained with the sessile drop technique (Table 10) are higher (or lower for the MoS2þ 2 surface) than the ones obtained with the Wilhelmy plate technique (Table 9). In conclusion, it can be stated that, independent of the steel substrate, the most hydrophilic and hydrophobic surfaces were, respectively, the SiOx -plasmaCVD and Ni–P–PTFE coated. 3.4.2. Contact angle hysteresis As mentioned in Section 2.4.1, the Wilhelmy plate technique measured both advancing and receding contact angle. Chemical heterogeneity and roughness of the surfaces are believed to contribute to the contact angle hysteresis (Adamson & Gast, 1997). Heterogeneity may arise from impurities concentrated at the surface, from crystal imperfections, or from differences in the properties of different crystals faces. Almost all of the modifications led to a lower hysteresis (Table 9). The higher values found for the 2R DLC sputtered and Ni–P– PTFE samples can be related to the more heterogeneous topography observed on the DLC surfaces (Section 3.3) and the increase in roughness on the Ni–P–PTFE samples (Section 3.2) compared to the untreated surface. 3.4.3. Surface energy The total surface energy value as well as the separation in corresponding non-polar and polar contributions as determined by the van Oss and Owens approaches are presented in Tables 11 and 12. The apolar component ðcLW Þ has approximately the same value for all samples, except for the Ni–P–PTFE coated, being around 40 mN/m for the 2R based samples and around 41 mN/m for the 2B (Table 11). The Owens method for solid–liquid–vapor interface also gives constant but lower values for the apolar contribution of about 30 mN/m, with the exception of the 78 O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 Table 11 Surface tension values (in mN/m) for the different steel samples according to the approach of Van Oss Samples cLW s cþ s c
s cTOT s 43.1 ± 0.7 41.3 ± 1.3 43.3 ± 0.7 43.9 ± 0.3 24.6 ± 0.4 0.6 ± 0.2 0.9 ± 0.4 0.5 ± 0.2 0.1 ± 0.0 0.3 ± 0.2 48.2 ± 4.2 37.2 ± 1.6 44.6 ± 4.9 14.3 ± 4.9 0.1 ± 0.2 53.3 ± 2.4 52.3 ± 3.6 52.7 ± 2.6 45.4 ± 0.7 25.0 ± 0.6 4. Conclusions 2R Unmodified SiOx -plasmaCVD SiF implanted DLC sputtered Ni–P-PTFE coated Silica-sol gel MoS2 implanted DLC–Si–O-plasmaCVD SiF FZ implanted 34.9 ± 1.3 0.7 ± 0.3 35.9 ± 0.5 0.2 ± 0.2 42.7 ± 0.2 0.6 ± 0.3 15.3 ± 3.4 41.6 ± 1.0 18.4 ± 5.0 39.8 ± 1.7 8.1 ± 0.8 47.0 ± 0.7 38.8 ± 1.0 1.0 ± 0.7 33.8 ± 0.6 50.1 ± 3.5 2B Unmodified SiOx -plasmaCVD SiF implanted DLC sputtered Ni–P-PTFE coated Silica-sol gel MoS2 implanted 42.8 ± 0.4 43.0 ± 0.3 41.5 ± 1.1 43.4 ± 0.2 20.5 ± 0.6 35.8 ± 0.7 42.1 ± 0.5 11.5 ± 1.0 52.1 ± 1.6 29.6 ± 3.9 14.5 ± 3.4 0.2 ± 0.1 17.7 ± 1.7 17.9 ± 2.0 2.0 ± 0.2 0.8 ± 0.1 0.3 ± 0.1 1.1 ± 0.5 0.0 ± 0.0 0.9 ± 0.1 0.2 ± 0.2 52.5 ± 0.3 55.6 ± 0.1 47.0 ± 0.8 51.0 ± 1.9 20.6 ± 0.6 44.0 ± 1.4 45.1 ± 2.1 Table 12 Surface tension values (in mN/m) for some 2R modified surfaces, calculated by the approach of Owens and Wendt with both solid– liquid–vapor (s-l-v) and solid–liquid–liquid (s-l-l) methods Samples Ni–P-PTFE coated DLC sputtered SiOx -plasmaCVD SiF implanted cTOT s cps cds most apolar surface, independent of the steel type, was the Ni–P–PTFE coated sample. s-l-v s-l-l s-l-v s-l-l s-l-v s-l-l 15 33 31 27 13 64 38 18 2 8 34 26 0 5 26 36 17 41 65 53 13 69 64 54 Ni–P–PTFE surface. The same holds for the electron acceptor component ðcþ Þ, which is in all cases close to zero. The biggest differences among samples were found in the electron donor component ðc
Þ. The values for the c
component vary between 0.1 (Ni–P–PTFE coating) and 48.2 mN/m (unmodified surface) for the 2R based samples and between 0.2 (Ni–P–PTFE coating) and 52.1 mN/m (SiOx -CVD coating) for the 2B samples. Similar total surface tension values, cTOT , were obtained sl by the Van Oss and Owens and Wendt approaches for the 2R samples. Regarding both methods used, solid–liquid–vapor and solid–liquid–liquid, the calculated surface tension by the Owens and Wendt approach (Table 12) were quite similar, except for the DLC coating. The reason might be that the spreading pressure is high. DLC and Ni–P–PTFE coatings are nearly apolar, unlike the surfaces coated with SiOx and implanted with SiFþ 3 , which are very polar, like glass. In conclusion, all techniques reduce the apolar component ðcLW Þ, increase the c
component for the 2R samples and decrease this parameter for the 2B. The The surface properties of different modified surfaces were assessed in terms of chemical composition, roughness, topography and wettability. Analyses of the chemical composition on surfaces where the modified layer had a thickness higher than 1 lm (see Table 2) revealed mainly the chemical elements of the coating. For surfaces with a thinner coating and the ion implanted surfaces, the implanted ions as well as the stainless steel composition could be determined. The modification techniques affected the roughness of the rougher 2B surfaces more than the 2R surfaces, which were largely unaffected in this respect. The silica coating, using the sol–gel technique, gave surfaces with similar roughness independently of the steel substrate. The DLC sputtering and Ni–P–PTFE coating produced the highest increase in surface roughness. All the modified surfaces presented have similar topography as the unmodified samples, with the exception of the surface coated with Ni–P–PTFE where the topography of the unmodified steel was no longer visible. In general, the surface modification produced more hydrophilic surfaces, with the exception of the Ni–P– PTFE, which gave hydrophobic surfaces. The SiOx plasmaCVD coating produced the most hydrophilic surface. The great potential of some of the surface modification (SiFþ 3 implanted and DLC sputtering) studied in this work has been shown in previous studies (Bornhorst et al., 1999; M€ uller-Steinhagen & Zhao, 1997; M€ ullerSteinhagen et al., 1997; Santos et al., 2003) which revealed that surface modification can lead to reduction in: (1) CaSO4 scale formation during pool boiling, (2) bacteria attachment and (3) protein adsorption. Presently we are investigating the milk protein and mineral fouling on all the modified surfaces presented in this paper. Acknowledgements The authors would like to deeply acknowledge Dr. Hans Visser, the initiator and co-ordinator of this work within the MODSTEEL project. Financial support was obtained from the European Community under the ‘‘Competitive and Sustainable Growth’’ Programme (MODSTEEL, Contract No. G5RD-CT-1999-00066, Project No. GRD1-1999-10856). References Adamson, A. W., & Gast, A. P. (1997). Physical chemistry of surfaces. New York: John Willey and Sons. O. Santos et al. / Journal of Food Engineering 64 (2004) 63–79 Addesso, A., & Lund, D. B. (1997). Influence of solid surface energy on protein adsorption. Journal of Food Processing and Preservation, 21, 319–333. Baier, R. E., & Meyer, A. E. (1992). Surface analysis of foulingresistant marine coatings. Biofouling, 6, 165–180. Binning, G., Quate, C. F., & Gerber, C. (1986). Atomic force microscopy. Physical Review Letters, 56, 930–933. Bornhorst, A., Steinhagen, H. M., & Zhao, Q. (1999). Reduction of scale formation under pool boiling conditions by ion implantation and magnetron sputtering on heat transfer surfaces. Heat Transfer Engineering, 20(2), 6–14. Boyd, R. D., Verran, J., Jones, M. V., & Bhakoo, M. (2002). Use of atomic force microscope to determine the effect of substratum surface topography on bacterial adhesion. Langmuir, 18, 2343– 2346. Britten, M., Gree, M. L., Boulet, M., & Paquin, P. (1988). Deposit formation on heated surfaces––effect of interface energetics. Journal of Dairy Research, 55(4), 551–562. Burton, H. (1968). Reviews of the progress of dairy science. Section G. Deposits from whole milk in heat treatment plant––a review and discussion. Journal of Dairy Research, 35, 317–330. Haynes, A., & Norde, W. (1994). Globular protein at solid/liquid interfaces. Colloids and Surfaces B: Biointerfaces, 2(6), 517–566. Janocha, A., Hegemann, D., Oehr, C., Brunner, H., Rupp, F., & GeisGerstorfer, J. (2001). Adsorption of protein on plasma-polysiloxane layers of different surface energies. Surface and Coatings Technology, 142–144, 1051–1055. Krisdhasima, V., McGuire, J., & Sproull, R. (1992). Surface hydrophobic influences on b-Lactoglobulin adsorption kinetics. Journal of Colloid and Interface Science, 154(2), 337–350. 79 M€ uller-Steinhagen, H., & Zhao, Q. (1997). Investigation of low fouling surface alloys made by ion implantation technology. Chemical Engineering Science, 52(19), 3321–3332. M€ uller-Steinhagen, H., Zhao, Q., & Reiss, M. (1997). A novel low fouling metal heat transfer surface. In Proceedings of the 5th UK National Heat Transfer Conference. Nasser, K. (2000). Galvanotechnik. Hanser M€ unchen. Norde, W., Arai, T., & Shirahama, H. (1991). Protein adsorption in model systems. Biofouling, 4, 37–51. Owens, D. K., & Wendt, R. C. (1969). Estimation of the surface free energy of polymers. Journal of Applied Polymer Science, 13, 1741–1747. Santos, O., Nylander, T., Paulsson, M., & Tr€ag ardh, C. (2003). Adsorption behaviour of b-Lactoglobulin on modified stainless steel surfaces. In Proceedings of the heat exchanger fouling, fundamental approaches and technical solutions conference, Davos, Switzerland. van Oss, C. J. (1994). Interfacial forces in aqueous media. New York: Marcel Dekker. Wahlgren, M., & Arnebrant, T. (1990). Adsorption of b-Lactoglobulin onto silica, methylated silica and polysulfone. Journal of Colloid and Interface Science, 136(1), 259–265. Yoon, J., & Lund, D. B. (1994). Magnetic treatment of milk and surface treatment of plate heat exchangers: effects on milk fouling. Journal of Food Science, 59(5), 964–980. Zettler, H., & M€ uller-Steinhagen, H. (2000). Confidential Report for BP Exploration. Zhao, Q., & M€ uller-Steinhagen, H. (2003). Intermolecular and adhesion forces of deposits on modified heat transfer surfaces. In Proceedings of the heat exchanger fouling, fundamental approaches and technical solutions conference, Davos, Switzerland.