The Early Stages of Deposition of Magnetite Particles onto Alloy-800 Heat Exchange Surfaces under Subcooled Boiling Conditions

The data collected during experiments in water loops on the fouling of heat exchange surfaces with magnetite particles often display a dip in the deposit accumulation curve at about 10 – 20 hours into the experiment. Generally the dips are small enough to be considered part of the scatter and may not even be discernible in long runs with considerable amounts of deposit. We have examined this phenomenon in a series of short experiments lasting up to a maximum of 40 hours. The test section of the recirculating water loop used for the study contained a heat exchanger tube of Alloy-800. Synthetic magnetite colloids were suspended in the coolant, which was usually maintained at 60C or 90C, and their deposition onto the tube was monitored for different conditions of pH, flow rate and heat flux. For a pH where deposition should be dominated by transport processes, deposition curves all displayed a “knee” at about 5 – 10 hours, regardless of whether the tube was heated or not. The curve shape can be explained in terms of a layered growth of the deposit, the first layer conditioning the surface for the growth of subsequent layers. First-order phenomenological coefficients in a simple model reproduce the curves quite well. INTRODUCTION In previous experiments in this series of studies of the deposition of magnetite particles from suspension in water onto heat exchange surfaces, deposition patterns were usually monitored for periods longer than about 50 hours (Callamand, Basset and Lister, 1999; Carpentier, McCrea and Lister, 2001) – in some cases up to about 700 hours (Basset et al., 2000). Thus, the long-term kinetics of deposition were established and the means of predicting fouling as a function of the operating parameters of the system were provided. For understanding the mechanisms, however, the initial period of fouling can be particularly important, since the first layers of particles to deposit affect the surface markedly, modifying it for the deposition of subsequent layers. The first few layers are therefore expected to deposit with kinetics different from those of the rest of the deposit. In water systems, a prime example arises from the change in electric charge on the surface of a heat exchanger tube at a certain pH as particles of a different point of zero charge (PZC) deposit. Initially, charge differences would create attractive forces and deposition would be promoted, in the extreme case limited only by transport in the fluid. Later, a thicker deposit would present similar charges to the particles and repulsive forces would impede deposition, in the extreme case limiting the process entirely through the particle-surface attachment term. The scatter of the data in the deposition results from the early experiments often seems to contain a dip in the patterns that arise from a reduction of the deposition rate a few hours into the experiments (Carpentier, McCrea and Lister, 2001); Basset et al., 2000). As described later, we have reproduced this in a series of short experiments measuring the deposition of synthetic magnetite particles onto Alloy-800 heat exchanger tubing. The distinct knee in the accumulation curves that occurred at 5 – 10 hours can be described by the mechanism of an initial layer’s modifying the surface for the deposition of subsequent layers. EXPERIMENTAL METHODS AND APPARATUS The magnetite particles used in the experiments were synthesised using a sol-gel technique first described by Sugimoto and Matijevic (1980). Monodispersed spheres were obtained; at about 0.4 μm diameter these were somewhat smaller than the 0.6 μm diameter particles of Basset et al. (2000). The magnetite was added to the coolant of a recirculating water loop that operated at about atmospheric pressure and at temperatures up to 90C. Loop construction is mostly of stainless steel, comprising a 170 L reservoir equipped with a stirrer and electric heater, a centrifugal pump and a cooler. The test section is a vertical glass column, 1.5 m long and 9.4 cm I.D., with two outlet ports at the top. The Alloy-800 heat exchanger tube has a 1.6 cm OD and is cut to a length of 30 cm. It is inserted into the closure seal at the top of the test section. Heat fluxes up to 240 kW/m can be generated in the tube by means of an internal electric heater. During an experiment, samples of coolant were taken regularly and tested for pH and magnetite concentration, the latter via an atomic absorption spectrometer. Most of the experiments were done at neutral conditions but, when required, adjustments of pH were made with potassium hydroxide or nitric acid. Air was excluded from the coolant by continuous purging with nitrogen and magnetite concentration was adjusted by dilution or colloid addition. 1 Arbeau et al.: The Early Stages of Deposition of Magnetite Particles Published by ECI Digital Archives, 2003 The amount of deposit was measured by removing the tube from the glass column, dissolving the magnetite in dilute hydrochloric acid and measuring the iron concentration with atomic absorption. This provided an average surface concentration of magnetite on the Alloy-800. Experiments were mostly carried out at 90C bulk temperature in the column with a few runs at 60C. Scoping tests to gauge the effects of system parameters on magnetite deposition measured the deposition velocity as the amount deposited after six hours exposure divided by the time (assumed to be the initial deposition rate) and by the bulk concentration of magnetite. The magnetite concentration in the loop was usually controlled between 4.0 and 4.5 μg/cm. RESULTS Many of the runs were carried out with a heat flux of 155 kW/m. At the highest coolant flow rate in the loop (13.5 L/min), giving a Reynolds number of 8,037 in the annular flow region of the test section, sub-cooled boiling was exhibited over most of the tube surface. The amount of boiling could be adjusted on the tube by increasing the heat flux or by lowering the flow rate while maintaining the heat flux at the original setting. Deposition during boiling produced 0.1 mm – 0.2 mm diameter rings or spots of magnetite around the steambubble nucleation sites. These were similar to those described earlier (Basset et al., 2000) and are attributed to particle trapping and agglomeration on the growing bubble surfaces and to evaporation of the coolant micro-layers below the bubbles. Figure 1 is a composite scanning electron micrograph of a section through a typical ring along with its measured profile shown to scale. The first set of experiments gauged the effect of bulk magnetite concentration on the deposition velocity. The results showed that the deposition velocity is approximately constant over the range of concentrations explored (0.112 – 48.5 μg/cm). This is consistent with previous findings, which found deposition rate to be approximately linear with concentration (Carpentier, McCrea and Lister, 2001); Basset et al., 2000). Figure 2 presents the results of two sets of runs showing how deposition velocity varies over similar ranges of pH. The sets are in good agreement, both indicating a maximum deposition rate at about pH 7.5. This corresponds to the result of Basset et al (2000), who found the maximum deposition rate under similar sub-cooled boiling conditions to occur at pH 7.5 – 8.0. In a previous study under bulk boiling conditions, however, maximum deposition rates occurred at lower pH values, presumably because of local changes in the PZC of both the particles and the surface caused by local concentration of the chemistry additive (Carpentier, McCrea and Lister, 2001). Figure 3 presents deposition velocity as a function of heat flux for otherwise constant conditions. As was found before (Basset et al., 2000), deposition was fairly constant below a heat flux of 40 kW/m, close to the value at which bubble nucleation began, and increased rapidly thereafter. Unlike the previous results, however, the variation with heat flux was more or less linear rather than parabolic. This may be due to the scatter of the data over the lower range of 0 – 194 kW/m, in contrast to repeated results over the range 0 – 240 kW/m in the previous experiment. In any case, the curve is much flatter than expected from observations in previous experiments, which is probably the result of data scatter. Figure 1. Composite scanning electron micrograph of a section through a magnetite ring formed at a bubble nucleation site. 2 Heat Exchanger Fouling and Cleaning: Fundamentals and Applications, Art. 35 [2003] http://dc.engconfintl.org/heatexchanger/35 3 4 5 6 7 8 9 10 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 D ep os iti on V el oc ity (μ m /s ) pH @ 25C Trial1 Trial2 Re = 8037; 155 kW/m, Tb = 90 C; Cb = 4.3±0.1 μg/cm 3 Figure 2. Variation of deposition velocity with pH25oC. -50 0 50 100 150 200 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 90C 60C D ep os iti on V el oc ity (μ m /s ) Heat Flux (kW/m) Re = 8037; pH = 6.4±0.1; Cb = 4.3±0.1 μg/cm 3 Figure 3. Variation of deposition velocity with heat flux. 4000 5000 6000 7000 8000 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 D ep os iti on V el oc ity (μ m /s )

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.