**4. Discussion**

Although there is a visible difference between the olive oil and water present in the sponge, it is extremely challenging to quantitatively retrieve the interface between both liquids (Figure 3d). As shown in Table 1, the attenuation coefficient values of these fluids were in close proximity to each other.

The contrast was therefore improved by using two different contrast agents namely magnetite powder and brominated vegetable oil. The dispersion capabilities of the contrast agents in the olive oil and the influence of the contrast agen<sup>t</sup> on the properties of the olive oil are of high importance and limit the practically achievable concentration. Due to the inert chemical behaviour of magnetite powder with olive oil and their difference in bulk densities, sedimentation of magnetite powder occurred at concentrations higher than 50% (wt/volume). For brominated vegetable oil, such sedimentation was not observed as the two liquids were miscible. However, with a thorough premix of the solution, the olive oil acts as a suitable carrier for magnetite powder [43] and the solution remains stable for sufficiently long time before being applied on to the sponge.

With the help of 2D cross sectional images a thin film may be seen on the walls of the sponge pores due to the adhesion of the contrasted olive oil, making it partially a closed cell structure. Also due to the uneven distribution of contrasted olive oil, the sponge tends to lose its stability and collapses towards the heavier side. This was one of the challenges that had to be faced while conducting the experimental protocol. After each step, the sponge was rearranged before the scan to have the full view of the structure.

The traces of magnetite powder dispersed olive oil were present in the sponge even after two stages of cleaning. A possible cause of this phenomenon is the magnetite powder without olive oil which tends to stick to the sponge material due to the heterogeneous solution. This separation of magnetite powder from its solution will not be helpful as the main purpose of cleaning the sponge becomes questionable. The influence of the contrast agen<sup>t</sup> on the olive oil properties, particularly with respect to the interaction with the sponge material, is a very complex research question. Solving this is out of the scope of this manuscript but part of parallel research in the same research consortium. However, the use of brominated vegetable oil as contrast agen<sup>t</sup> for olive oil can be justified as bromine results in a higher attenuation coefficient and the two liquids form a stable solution.

The contrasted olive oil added onto the sponge could impregnate purely by gravity and capillary forces. According to laws of capillarity, the small pores cause higher capillary pressure for the wetting phase (contrasted olive oil) to move towards non-wetting phase (air filled pores) hence making it difficult for imbibition of contrasted olive oil through the sponge. This resulted in the concentration of contrasted olive oil at the periphery (Figures 4a and 5a).

To assess the entire volume of contrasted olive oil present in the sponge, the different phases were segmented in 3D analysis software. The limitation to find the optimal threshold value together with partial volume effects were some of the reasons for the difference between the measured amount of olive oil in the uncleaned sponge and the added amount of olive oil (Table 2, 3.6 mL of magnetite powder dispersed olive oil and 1.2 mL of brominated vegetable oil mixed olive oil compared to 5 mL inserted in the system). One of the other important reasons may be the loss of contrasted olive oil through the outlet after addition, for which solutions will be sought. The amount of solution of brominated vegetable oil in olive oil left behind in the sponge was relatively less and unlike magnetite powder the dissolved brominated vegetable oil did not adhere to the sponge and therefore provides a better approach to contrasting.

As the system was not operated in vacuum, the displacement of the wetting phase by a non-wetting phase resulted in trapping of contrasted olive oil droplets inside the sponge [44]. Dynamic processes like squeezing and rinsing were necessary to mobilize and emulsify this trapped soil droplets for easy removal. Process 2 and Process 3 illustrated in Section 2.7 helped in mobilizing water-soluble particles present in the sponge and in emulsification of fatty soils respectively [45]. These processes facilitated for removal of contrasted olive oil dispersed in the sponge.

The flow cell with its cyclic action and with the help of water channels was successful in removing the contrasted olive oil present inside the sponge. Comparing the percentage of cleaning and the enhancement in attenuation coefficient of olive oil, brominated vegetable oil becomes a favorable choice as contrast agent. The 3D rendered images in Figures 4 and 5 depict the displacement and relocation of fluid inside the porous structure in relation to time and hence can be used as an input for developing a model for fatty oil absorption and removal with a detergent solution. This method can be employed to optimize a detergent formulation that can quickly wet the kitchen sponge and the oil trapped in between the pores, emulsify the fat into oil droplets and completely remove them via squeezing action.

X-ray 4D-μCT (time dependent imaging) is a suitable choice for visualizing in situ experiments as it needs no sample preparation and gives both qualitative and quantitative data. Application of suitable contrast agents have further enabled the technique to image low attenuating samples in a multiphase fluid system non-destructively.
