| dc.description.abstract | Dehydration is an ancient method of fish preservation, and comprehensive experience has been gathered from various products and treatments. Several attempts have been made to improve the understanding of the dehydration process in the fish muscle, and there is extensive data on the diffusion coefficients of water in fish muscle as well as of salt in fish muscle, mainly generated in Scotland and Canada. Improvement of product quality has been an objective for centuries, and in the recent decades, some concepts have emerged. One is to start the osmotic dehydration (i.e. salting) by a dilute brine, and preserve a flexible, light surface. This approach is also known as Gaspécure, after the research laboratory at Gaspé, Canada. This concept has been developed further, and submitted to research funded by the Norwegian Research Council. Regarding drying, on the other hand, beneficial effects of low temperature initial drying have been documented by the Dewatering Laboratory at NTNU / SINTEF, Norway.
This work has attempted to uncover the mechanisms of moisture binding, its release and transport under different points of view. Several methods of investigation have been applied, and some combinations are novel. Particularly the microscopy of the ultrastructure of the muscle tissue combined with capillary models and sorption isotherms is a new angle of investigation. Transport coefficients have been extracted from literature data, and used in diffusion models, in the form of analytical solutions to Fick’s 2nd law.
One part of this work has been to examine the ultrastructure of fish muscle in a porosity perspective, in search of a viable approach for model construction (Paper I). A capillary approach has been attempted, by examining the moisture binding at different centrifugal pressures. The utltrastructure of the filaments may be characterised by cylindrical chords, respectively thick filaments of myosin and thin filaments of actin, arranged in a regular, hexagonal pattern. The influence of salt is strong on the ultrastructure of the muscle, as well as the sorption isotherm. Osmotic dehydration is a strong treatment, in spite of its apparent simplicity. Convective dehydration, however, imposed no irreversible changes to the ultrastructure. At this level of accuracy, one was unable to detect differences between tissue that had been dried at low temperature or medium temperature.
A salting experiment in semi-industrial scale has been carried out, with different brine strengths in the first 48 hrs (Paper II). Light curing as initial salting has proven feasible for improving product quality and yield, which has been explained by initial swelling of the proteins of the muscular tissue.
A laboratory scale salting experiment was carried out to verify diffusion coefficients and mathematical models (Paper III). Attempts have been made to estimate mass transfer resistance through skin and through stagnant boundary layer at surface. These resistances have been converted to equivalent product thickness or "length" in order to apply analytical solutions to the diffusion equations (i.e. Fick's second law). A similar approach has been used to estimate the extra mass transfer resistance due to the stagnant brine, as most experiments in literature have been made under agitated conditions, in contrary to industrial practice, which is in stagnant brine.
Further, a convective experiment was undertaken, with unsalted fillet of different species (Paper IV). This experiment was repeated in two different temperature programs, low temperature (LT) program and medium temperature (MT) program. The drying process showed different characteristics during the two temperature programs, while negligible differences were found between the species. The dried products from the two temperature programs were different. Among other parameters, the rehydration index was significantly higher in the low temperature dried products than in the medium temperature dried products. Several species of low fat muscle have been examined in this work. Generally, the transport coefficients and models obtained with cod and swordfish could be applied directly. Substitution of cod may be obtained with some related species, like grenadier and mora, whereas dogfish show other properties less favourable.
Finally, a convective drying of salted fish was carried out, also at the same two temperature programs as described above (Paper V). The process characteristic showed similar behavior as the previous experiment, however the differences in the product groups were less pronounced in the salted fish after drying.
The models give good correlation to experimental data, as long as the boundary conditions have been appropriately formulated. Also, convective analysis has been undertaken to predict the heat and mass transfer coefficients between fish and drying air. A factor of great importance is the drying temperature, or more precisely, the drying air temperature in the case of adiabatic drying chambers.
This work has documented the kinetics of a drying process at two temperature programs, including the effect of raising the temperature during dehydration. The medium temperature program rapidly developed a layer of high transport resistance, whereas the low temperature program (frozen tissue) did not. When increasing the temperature level, the drying rate increased significantly for some 20 hours in this case. Studies of the dehydration of fish muscle, both osmotic and convective, have revealed that an excessive initial liquid flow to the surface has adverse effects on the drying rate as well as the quality and yield of the product. During osmotic dehydration, the use of diluted brine allows the tissue to swell and retain soluble compounds before an outward liquid flow is established. Likewise, during convective dehydration, a freezing of the tissue will immobilise the liquid phase and allow sublimation of water without a solute transport to the surface. Thus, a vapour transport through the surface may prevent a solute build-up at the surface forming a barrier to moisture transport. Later, when the temperature is increased, some liquid flow will be established from the interior to the drying zone. The drying zone is not a plane surface however, but a 3-dimensional zone with considerably increased surface area, which seems to be open to transport across the surface. The significant increase in drying rate following the temperature rise confirms this assumption.
The beneficial effects of low temperature drying have been documented in previous research works, particularly at the NTNU and SINTEF, and these effects have been confirmed in this work. In an economical view, there is a trade-off between low temperature and decreased drying rate at low temperatures. This is correct regarding the drying machinery capacity, where the SMER ratio necessarily decreases with decreasing temperature in an adiabatic dryer. Regarding the internal transport resistance however, the drying rate disadvantage may be overcome by optimising the temperature program of dehydration, for instance by a stepwise increase of temperature. A temperature rise from low temperature resulted in a significant rise of drying rate, although no attempt was made to optimise the drying rate of the low temperature drying, and still, the dried amount of moisture tended to coincide after some 70 hours.
A possible moisture-binding model could be multilayer adsorption to and around these chords, applying the BET isotherm for multilayer adsorption. The BET isotherm describes multilayer adsorption on a plane, not a cylindrical surface. Thus, the amount of adsorbed moisture per layer increases more than the BET isotherm describes, which has to be accounted for in some manner. One option is to use an equivalent number of layers to compensate for the increased moisture volume per additional adsorbed layer due to the cylindrical geometry.
This work may modify the conception of fish muscle as a porous system, as the basic system is not characterised by liquid-filled pores surrounded by a solid, but a filament of solid chords surrounded by liquid more or less strongly adsorbed. In consequence of this perspective, the validity of the BET isotherm may be extended into the high-moisture range of the sorption isotherm. In conclusion, fish muscle is a very complex material with different drying properties from many other dried materials, like ceramics and sand. In spite of this, dehydration may obey relatively simple equations, provided the correct boundary conditions are given. | nb_NO |
