**3. Results**

#### *3.1. Characterization of Materials*

Figure 3 illustrates the characterization of the sponge microstructure in absence of load, with and without liquids inside. In Figure 3a the pore structure of the dry sponge can be seen and Figure 3c illustrates the distribution of water and olive oil inside the sponge. With the help of 2D cross sectional images the structure and distribution of the fluids at any given layer inside the sponge can be visualized in 2D and 3D. Although the location of olive oil and water inside the sponge could be visually assessed however, in a single system it was impossible to accurately discriminate between olive oil, water and sponge material (Figure 3d).

**Figure 3.** (**a**) Cross sectional image indicating the microstructure of the dry sponge. (**c**) Cross sectional image of sponge with water and olive oil without application of contrast agent. (**b**,**d**) Zoomed images of the result shown in (**<sup>a</sup>**,**<sup>c</sup>**), depicting the microstructure of the sponge and water and non-contrasted olive oil in the sponge, respectively. The colormap illustrates the reconstructed attenuation coefficient of different materials.

Table 1 illustrates the experimental attenuation coefficient values of each of the test fluids retrieved separately in multiple scans with same scanner settings. The images were loaded in Octopus Analysis and the average grey value was determined over a volume of interest of approximately 3140 mm3. This grey value was later converted to a linear attenuation coefficient μ using the calibration from the reconstruction. Although these values are separated by 2σ (standard deviation), it should be noted that these are obtained in a container of pure material. In a real system such as the sponge, the features to be recognized are small and partial volume effects have a significant contribution [6]. Furthermore, these measurements are obtained from a high-quality scan, and in-situ experiments will yield higher noise levels.


**Table 1.** Solvents with their experimental linear X-ray attenuation coefficient (μ) and the achieved standard deviation σ (μ) indicating the close proximity in attenuation coefficient values of test fluids and attenuation coefficient enhancement of olive oil by addition of contrast agents.

#### *3.2. Assessment of Contrast Agent Specificity*

In the specificity test, both contrast agents were dispersed evenly in the solution by severe mixing. The magnetite powder solution however proved to be unstable over time due to sedimentation of the powder. The stability of the solution depends on the concentration of magnetite powder dispersed in olive oil. With lower concentrations the onset of sedimentation only happens if the solution remains stationary in its liquid state without being applied onto sponge.

The values of contrasted olive oil in Table 1 indicate that there is a considerable improvement in the attenuation coefficient of olive oil with dispersion of magnetite powder and brominated vegetable oil respectively. 10% wt/volume concentration of contrast agen<sup>t</sup> dispersed in olive oil was chosen to be optimal for all the experiments considering the sedimentation property of magnetite powder in olive oil.

#### *3.3. Experiments of Contrasted Olive Oil on Sponges: Cleaning Assessment of the Custom-Built Device and Quantification of Contrasted Olive Oil in the Sponge*

Octopus Analysis software was used to calculate the volume of contrasted olive oil present in the sponge at different stages of the cleaning process and it was compared with the actual volume of contrasted olive oil added to the sponge. Although 5 mL of contrasted olive oil was added on to the sponge only 3.6 mL of magnetite powder dispersed olive oil and 1.2 mL of brominated vegetable oil with olive oil could be recorded through image segmentation (Table 2). The factors influencing this difference in value will be explained in the Discussion section.

**Table 2.** Volume of contrasted olive oil at different stages of cleaning indicating the removal of contrasted olive oil from the sponge. Stage 1 of cleaning process: soiled sponge thoroughly rinsed with 5 mL of detergent is squeezed for 10 min at 10 cycles/min with continuous water feed. Stage 2 of cleaning process: 5 mL of detergent is added to the same sponge and squeezed for 10 min at 10 cycles/min with continuous water feed.


Using Equation (1) the percentages of cleaning for the sponge were determined. In Table 3 the percentage of cleaning for the given constant volume at two cleaning stages are given. Sponge applied with brominated vegetable oil mixed olive oil showed a better cleaning percentage compared to sponge with magnetite powder dispersed olive oil and the change in the cleaning percentage between the two stages were less for the latter case.


**Table 3.** Percentage of cleaning for two sponges, sponge 1 (magnetite powder dispersed olive oil) and sponge 2 (brominated vegetable oil mixed olive oil).

#### *3.4. Experiments: Dynamics of Soil Removal from Sponges under Loading*

From the stack of reconstructed slices, a 3D volume of both the uncleaned and the cleaned sponges were rendered using VGStudio MAX 3.2 software. Figures 4 and 5 give the visual representation of the contrasted olive oil with respectively magnetite powder dispersed olive oil and brominated vegetable oil with olive oil in the sponge before and after two stages of cleaning along with the sponge after Process 3 (Section 2.7). The removal of contrasted olive oil in the sponge is made possible by the squeezing action of the plunger, the external compressive force overcomes the capillary forces and thereby the trapped fluid gets displaced by a non-wetting phase. Dynamic action of the plunger not only increases the pore size of the sponge but also helps in mobilizing the water droplets and detergent throughout the porous structure.

**Figure 4.** 3D rendering of sponge with magnetite powder dispersed olive oil (**a**) before cleaning, (**b**) intermediate Process 3 (Section 2.7) and (**c**) after two stages of cleaning. Pseudo coloration is performed based on the segmentation: blue represents magnetite powder dispersed olive oil, pale blue represents water, red represents detergent and yellow colour represents the sponge. The residue was present even after 2 stages of cleaning for the magnetite powder dispersed olive oil sponge.

**Figure 5.** 3D rendering of sponge with brominated vegetable oil dispersed olive oil (**a**) before cleaning, (**b**) intermediate Process 3 (Section 2.7) and (**c**) after two stages of cleaning. Pseudo coloration is performed based on the segmentation: blue represents brominated vegetable oil with olive oil, pale blue represents water, red represents detergent and yellow colour represents the sponge. There were no traces of contrasted olive oil in the cleaned sponge.
