[nanoPost] COHESION AND ADHESION IN THE CLEANING OF FOOD FOULING DEPOSITS

University UK

Cleaning of process plants is costly and time consuming. It is important to understand the removal of food fouling deposits. Fouling forms as a result of adhesion to the surface and cohesion between elements of the material. We have used micromanipulation to measure directly the adhesive/cohesive strength of deposits on a stainless steel surface. The apparent adhesive strength between the fouling deposits and the substrate can be measured as the work required to remove the deposits per unit area from the surface. The method has been used to study a wide range of food deposits, including proteins and starch-based materials. A range of surfaces of different energy have been used to study the effect of surface energetics on adhesion and on the force required for removal. Measured values of the apparent adhesive/cohesive strength are a function of heating temperature and time, probe speed, cleaning agent concentration and exposure time. The work can be combined with visual monitoring of the cleaning process, and makes it possible to separate adhesive and cohesive effects: some deposits (such as tomato starches) are essentially cohesive, and can be removed in large chunks by hydration of the deposit-substrate interface.

Introduction
Extensive work has been done on the fouling from food deposits, especially dairy materials (see conference proceedings such as Fryer et al, 1996, and Wilson et al, 1999, 2002) and models have been produced which claim to reduce fouling in practice (such us de Jong et al, 2002 and Petermeier et al, 2002). Work on cleaning is less developed, and predictive mechanistic models are not available. The physical properties of the deposit and the surface will affect removal. Fouling deposits form as a result of adhesion of species to the surface and cohesion between elements of the material. The principal factors responsible for adhesion between surface and foulant include: (i) van der Waals forces, (ii) electrostatic forces, (iii) and contact area effects; the greater the area the greater the total attractive force (Bott, 1995). The forces between elements of the deposit depend on the nature of the material; deposits may be covalently bonded (for example reacted egg or milk proteins) or held together physically (such as gelled biopolymers).
An understanding of the interaction between deposits and surfaces is clearly critical in cleaning. On a smaller length scale, atomic force microscopy (AFM) has been used to characterise surfaces and fouling (such as Parbhu et al, 2002; Weiss et al, 2002). Low-adhesion coatings (Muller-Steinhagen and Zhao, 1997) have been shown to reduce fouling in some situations such as mineral scales. Zhao et al. (2004) demonstrate that biofouling can be reduced by changing surface energy, and link this to adhesive energy between surface and deposits; there is evidence for the minimum in terms of foulant attachment when the surface free energies cover the range 20-40 (mN/m). The theory gives that a minimum adhesion energy between deposit and surface exists, given by:
                            (1)
where  ,   and   are the Lifshitz-van der Waals (LW ) surface free energy of the surface, foulant and fluid (e.g. water), and can be determined experimentally from contact angle measurements (Zhao et al, 2004).  As the equation isolates the effects of surface free energy upon foulant adhesion from the numerous parameters in the DLVO theory, it appears relatively simple. In general, for a multi-component system, fouling formed on processing equipment may consist of various types of foulants (such as the mixtures of proteins and minerals deposited from milk systems) and    in equation (1) is the average LW surface energy of the fouling deposit, which can be determined by measuring contact angles on the deposit. If the surface energy of the stainless steel surface is reduced to the fouling-resistant value by some surface modification technique, foulant adhesion force to the surface could be decreased significantly, and the fouling deposit could be removed more easily. The effectiveness of this approach has been demonstrated by Zhao et al (2004) for microbial adhesion; here, we use materials science techniques to measure the effects of changing surface directly.

We have developed novel micromanipulation probes to remove layers of deposit at millimeter level (Liu et al, 2002). The force required to disrupt a surface film can be determined by drawing a probe across the film, and the effectiveness of removal followed both by filming the process and examining the surface afterward. The method was developed to study biofilms (Chen et al, 1998), but has been extended to study various food films. The methods give for the first time a direct measure of the forces required to remove deposits, and offer ways in which different surface and operating procedures can be compared. The aim of research work is to identify the forces needed to disrupt deposits, how they are affected by conventional (and novel) cleaning methods and thus how to optimise cleaning. Here, we used the equipment to observe process variables affecting the adhesive forces of the deposits, and separate adhesive and cohesive effects from the different samples. Experiments have also studied removal from a series of coated surfaces to observe the effect of surface energies on adhesive and removal forces. Classification of deposits into those that are primarily adhesive and those that are cohesive could be useful both for selection of cleaning protocols and of the correct type of surface.

Experimental
Fouling deposits
Four types of fouling deposit have been studied: (i) Tomato paste baked and unbaked (Liu et al, 2002); (ii) milk protein deposits (Liu et al, 2004); (iii) chicken egg albumin deposit from powder (Sigma, grade II, code A-5253), prepared by dissolving powder in distilled water, then heating disks in contact with the fluid in a water bath. The discs were fouled with albumin deposits (composition: 42.81% protein, 1.38% carbohydrate and 0.27% phosphorus) (Analytichem, Birmingham, UK); the mean value of the sample mass was 1.06 g (wet) and thickness was 2.2 mm; (iv) a model bread dough, composition 60.6 % plain flour, 1.8 % wet yeast, 1.2% salt and 36.4 % water.

Modified surfaces
In this investigation, the surface energy of stainless steel disks was modified by Ni-P-PTFE composite coatings. The composition and the plating conditions for the electroless Ni-P-PTFE solutions used in the present investigation are listed in Table 1.  A 60% PTFE emulsion from Aldrich with particle size in the range 0.05 to 0.5 m and a cationic surfactant C20H20F23N2O4I (FC-4) were used.  Both the PTFE emulsion and the surfactant were diluted with demineralised water and stirred for one hour. Then the solution was filtered with a filter of pore size 0.2 µm before use. AFM images of some of the surfaces are shown in Figure 1; the nanostructure of the surfaces is very different.

Contact angles of the Ni-P-PTFE composite coatings and tomato paste (baked and unbaked) were obtained using a sessile drop method with a Dataphysics OCA-20 contact angle analyser. The surface energies of the coatings and tomato paste were then calculated, based on the values of contact angles using van Oss acid-base approach (van Oss et al, 1986).

The average LW surface energies of tomato paste (baked and unbaked),   were approximately 30.6 and 29.2 mN/m, respectively. The average LW surface energy of water,   is 21.8 mN/m. Equation (1) then produces theoretical values of

•   = 26 mN/m, for which the adhesion of baked tomato paste is minimal;
•   = 25 mN/m, for which the adhesion of unbaked  tomato paste is minimal.

Typical experimental results
The force required to remove the deposit is measured by drawing the micromanipulation probe across the surface of the deposit, The total work, W (J) done by the applied force, F(t) to remove the deposit may be calculated as the integral:

where d is the diameter of the circular disc, and tA and tC the first and last times at which the probe touched the fouled surface. The apparent adhesive strength of a fouling sample,  (J/m2), defined as the work required to remove the sample per unit area from the surface to which it is attached, is then given by 

where A (m2) is the disc surface area, and  is the fraction covered by the sample measured by image analysis as described above. The relationship between  and the actual adhesive strength between the surface and the deposit is not clear, as the measured force not only removes the deposit from the surface but also deforms it.

Results
Adhesive/cohesive strength
Two types of measurement are possible: (i) total removal of deposit, in which both adhesive and cohesive forces are overcome, and (ii) partial removal of deposit to measure cohesive strength. This has been done for all the four deposits. Figure 2 compares the results for full removal (adhesive) and partial removal (cohesive) of the four samples. For tomato paste (a), the force required for partial removal of the deposit exceeds that for the total removal, showing clearly   that   the   cohesive   forces   between   the deposits exceed those of adhesion between surface and deposit, the gap between the probe and substrate is 10 µm for total removal and 50 µm for partial removal; For milk, the initial thickness of the deposit layer was around 1300 µm. Force measurements were taken after leaving the gap between the probe and substrate to 900, 600, 100µm（partial removal）and 20 µm (total removal) respectively. Here, the cohesive forces between elements of the deposit  are  weaker than those of adhesion between surface and protein deposits. This is opposite to the behavior of tomato starch shown above, in which it is easier to remove the whole of the deposit than it is to remove a surface layer. For bread dough, the cohesive force was measured by leaving a 800 µm gap between the probe and the substrate during removal of the dough. The result shows that the adhesive strength is the same as the cohesive strength at the first 5 mins of dehydration, but, beyond this time, the adhesive strength exceeds the cohesive strength and the difference between the adhesive and cohesive strengths are virtually constant while the dehydration time increases.  For albumin, the cohesive strength was measured by leaving a 800um gap between the probe and the substrate during removal, the result shows that the cohesive strength is greater than the adhesive strength. These measurements show clearly that different materials have different balances between deposition and removal. This may enable us to select  different cleaning protocols for different deposits.

Consequences for cleaning
It is reasonable to suppose that the removal of deposit from a surface will occur when the forced required break either the deposit-surface or the deposit-deposit bonds are exceeded by the forces provided by fluid shear. The cleaning behaviour of the deposits was studied visually, and also in term of heat transfer recovery by using a parallel flowcell set-up. Figure 3 shows a result for cleaning of a tomato deposit. The figure shows the area covered by the deposits, and the heat transfer coefficient through the material, as a function of time. When the deposits contacted by a flow of water, the surface becomes clean in one go: all of the tomato is removed from the surface in about 525 seconds after start of the experiment. This reflects that the forces holding the deposits together exceed those that bind it to the surface. When the diffusion of water to the interface has lowered the adhesive forces sufficiently, all of the deposit is removed. The side pictures show individual stills from the camera system used to determine the area covered: the material swells and changes colour from the start before being rapidly removed. Four pictures were taken at 100 seconds (upper left hand picture), 300 seconds, 500 seconds and 525 seconds from the start respectively,

By way of contrast , the removal of milk protein deposits are more commonly removed in patches and gradually in small chunks for the surface. The morphology of the deposits changed after the deposits were contacted with cleaning chemicals from a matrix of aggregated particles to a more viscoelastic gel-like appearance (Liu et al, 2004)

Effect of modified surface energies on adhesive and cohesive strength.
Experiments have studied the removal of tomato paste, baked and unbaked onto the modified surfaces which had surface energy values ranging from 15 – 40 mN/m. The modified surfaces were made by coating the stainless steel disc (26mm in diameter) with electroless plating. The advantage of using unbaked paste is that homogeneous layers can be deposited. Details of the surfaces are given in Zhao et al (2002, 2004): as shown in Figure 1, the topography and roughnesses can be significantly different.

Figure 4 shows data for baked and unbaked pastes; in both cases there is a minimum adhesive strength between the surface energies of 20 and 25 mN/m, with an increase in the adhesive strength on either side. The data is scattered, but the minimum is clear in both cases; it is more obvious for the unbaked material in which the change in adhesive strength with free energy is smoother. The minimum is in the region predicted by the theory (see above); at either side of this minimum the force increases, an effect especially marked for the baked-on sample, as would be expected, as that has been cooked onto the surface and would be expected to bond more strongly.

The force measured by the micromanipulation probe is a composite of the cohesive forces between deposit elements and the adhesion to the surface. This is shown explicitly by Figure 5, which shows data for different thicknesses of unbaked paste. It can be seen:

• As the thickness increases, the total force required to remove the deposit increases. The increase reflects the need to overcome the cohesive forces between elements of the deposit and to force the deposit to break and flow with the probe away from the sample surface; as the thickness increases, so does this force
• A similar minimum to that found in Figure 4, in similar regions of surface energy. However, the minimum becomes more difficult to identify as the thickness of deposit increases, reflecting the decreased contribution of the surface forces to the whole. At the highest thicknesses used, there is no measurable minimum; the curve simply flattens out at the lowest surface energies.

Conclusions
Micromanipulation equipment makes it possible to separate adhesive and cohesive effects in the cleaning of food fouling deposits: some deposits (such as tomato paste) are essentially cohesive, and can be removed in large chunks by hydration of the deposit-substrate interface. On the other hand, protein deposits (such us milk) are adhesive; cleaning involves reaction of the deposit into a viscoelastic gel that adheres to the surface. This classification of deposits is useful as it allows decisions to be made about selection of cleaning protocols.

The use of modified surfaces have demonstrated that (i) the theoretical minimum adhesion condition  can be found in practice for tomato deposits; and that (ii) the effect of the surface decreases with increasing deposit thickness, as would be expected. Micromanipulation allows quantification of the minimum adhesion force for the first time; the effect of surface and bulk behaviour can thus be identified. Work is underway to quantify the relationship between the measured apparent adhesive strength and theoretical adhesive and cohesive strength. This could enable understanding in more details of the relationship between adhesion and deposit chemistry, and how to clean most effectively.