Thermal Decoating of Aluminium Scrap
MetadataShow full item record
A major problem in the recycling of aluminium is the presence of contaminants in the purchased scrap. Future plants may handle part of this challenge by installing a delacquering/decoating unit. This thesis attempts to lay the foundation for, and provide a practical recipe for, the thermal removal of lacquer. Such decoating of aluminium scrap prior to melting makes it possible to utilize the energy released by the combustion in a closed and controlled manner. The resulting benefits include lower emissions, improved control of the remelting process, reduced dross generation and greater melt cleanliness. The investigations described here were conducted on samples cut from an aluminium sheet (420 µm in thickness) produced by Holmestrand Rolling Mill ASA and coil-coated with a 5 µm thick polyester by Hydro Aluminium AluCoat AS. The coating was decomposed in a thermogravimetric / differential thermal analysis (TG/DTA) furnace, and the evolved gases were measured using a quadrupole mass spectroscopy (MS). The residues left on the aluminium sheet after decoating were characterised using a scanning electron microscopy (SEM) with Energy-Dispersive as well as X-ray spectrometry. The degradation was also monitored visually using a hot stage light microscopy. The images at crucial stages were digitally recorded. In the TG/DTA, the following quantities have been measured as a function of time and temperature in controlled atmospheres (N2, Ar, Ar+1%O2, Ar+5%O2, Air and O2): (i) the mass loss, (ii) change in mass loss, and (iii) enthalpy change. Nine different heating rates, ranging from 0.1 to 50ºC/min, have been employed. The results were found to depend upon the atmosphere and the heating rates, and were summarized below. The mass loss curves show three (or two) peaks in oxidizing (or inert) atmospheres. With increasing heating rate the peaks move to higher temperatures. Extensive data for the shapes of the curves are collected and used for developing kinetic models of degradation of the coating. Plots of enthalpy changes show two positive peaks (implying exothermicity) in oxidizing atmospheres (corresponding to the second and third peaks in the mass loss curves) and none in inert atmospheres. The higher the oxygen content of the atmosphere, the greater is the enthalpy change. The microscopic examinations are consistent with the view that each peak is associated material (as received) is uneven, with almost hemispherical cavities or “weak points”. During the first stage, these cavities turn into channels reaching the aluminium surface. In the second step more such channels are formed, and appear against a darker background, mostly carbon. Combustion, the third stage, starts on particles, found to be silica, and spreads out until the whole surface becomes carbon-free. Just before the first mass loss peak there is a release of methane detected by the MS. In the hot stage microscope at the same temperature this effect manifests itself as more than a threefold increase in the number of cavities. For the first peak no special emissions are observed. For the second peak, in both kinds of atmospheres, concentration peaks for acetaldehyde and ester linkeage are observed with the MS. In addition combustion gases, CO2 and H2O, are released. In argon, CO is produced. For the third peak only combustion gases CO2 and H2O are emitted. The reaction (degradation) rate dα/dt =f(α)k(T) is described in the literature typically as a product of a mass function f(α) and temperature dependent function k(T), where k(T) is given by the Arrhenius equation. A goal of this work is to deduce the form of f(α). In the past, the reaction had frequently been assumed to be of first order. For instance, Várhegyi [Chapter 4: Várhegyi, 2002], employing the experimental data from experiments conducted in the present work, assuming that the rate is proportional to the fraction converted. He describes parallel first order reactions in his model. The classical Kissinger solution emplyed initially in this work makes use of the temperature when the mass loss peaks occure. By plotting the reciprocal of the absolute temperature versus the logarithm of the heating rate (an Arrhenius plot), the activation energy and frequency factor are found. The lifetime of the coat in different atmospheres can then be estimated. The temperature integral derived in the literature give a hypergeometric function. Many approximations to the temperature integral are presented in the literature. In the work presented here a continued fraction solution is derived. On truncating this continued fraction many of the published approximations are obtained. Two models for the mass dependence are proposed, that give a function f(α) different from that used in first order kinetics. “Weak points concentration models” are deduced by assuming that the change of the number of “weak points” is proportional to the mass change. Also this change is proportional to the number or concentration of weak points in the remaining lacquer. When the fraction coating converted is small, the two models give the same result that f(α) is a decreasing exponential function of the fraction coat converted, α. This model break down, as may be expected, when less than 10% of the coating is left. Each degradation step is associated with an activation energy that is not dependent upon the heating rate. The activation energy is determined by minimizing the standard deviation over the average of the temperature integral for all heating rates. The activation energies for the different models derivate from the mean by less than 8%. The activation energies show the same trend for the various models. Comparing the predictions with experimental results it is observed that the rate constant for the second peak is proportional to the square root of the partial pressure of oxygen. The rate constant for the first peak is nearly independent of the oxygen pressure. It is believed that the degradation model presented here for decoating aluminium scrap can be employed to describe the behaviour of other polymer coats as well. The model should be useful for choosing suitable atmosphere and temperature profiles in industrial decoating units.