Date of Award

Fall 2013

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Agricultural and Biological Engineering

First Advisor

Martin Okos

Committee Chair

Martin Okos

Committee Member 1

Carol J. Boushey

Committee Member 2

Osvaldo Campanella

Committee Member 3

Ganesan Narsimhan

Abstract

Density depends on the experimental technique and structural properties of food. True, apparent and bulk are different types of densities. For porous foods such as baked food products and complex food systems, accurate measurement of density is challenging. Empirical correlations dependent on composition and temperature exist to calculate true density. Thus if composition and apparent density are known, porosity can be calculated, porosity = [1 - (apparent density/ true density)]. Various non-destructive diagnostic tools have been employed to study food microstructure. Among these, computed tomography (CT) and digital radiography (DR) utilizing x-ray imaging showed greater capability for evaluation of food density. The x-ray attenuation through materials is dependent on the chemical composition and is proportional to density as described by the Beer-Lambert's law. This relationship forms the hypothesis of this study. The proposed experiments were conducted to verify and develop this hypothesis further for food systems. The objectives of this study were 1) To determine the strengths and limitations of traditional and conventional density measurement techniques, 2) To determine the apparent density of food materials using x-ray linear and mass attenuation coefficients, and 3) To determine porosity of food materials using x-ray linear attenuation coefficients and image processing & analysis. Traditional density measurement techniques were operator-dependent and appropriate selection was necessary according to the nature and type of food. X-ray DR was effectively used to determine x-ray linear and mass attenuation coefficients of fundamental pure components of foods. Further, x-ray mass attenuation coefficients were used to directly determine the apparent density of food materials with known sample thickness. On comparison with density obtained from external dimensions, density determined from mass attenuation coefficient resulted in SD = 0.125 g/cm3 between the two methods for the same foods. In a different approach, apparent density was determined directly from linear attenuation coefficients by scanning at two energy levels using apparent density equation without any external measurement. Several liquid foods were also analyzed to establish the method. On comparison with density obtained from external dimensions, density determined from linear attenuation coefficients resulted in SD = 0.052 g/cm3. Porosity was determined directly from linear attenuation coefficients and food composition. However on comparison to porosity determined by calculation, porosity values resulted in SD = 0.137 for DR whereas SD = 0.167 for CT measurements. A mathematical technique based on Otsu's algorithm was used to determine porosity by analyzing x-ray images resulting in SD = 0.083 on similar comparison. Alternatively, porosity from apparent density obtained from x-ray linear attenuation coefficient resulted in SD = 0.06 on comparison. Thus, the potential of x-ray imaging via DR and micro CT to determine apparent density and porosity of foods was effectively demonstrated for foods with varying porosities.

Download

Share

COinS