## 3.1 Percentage Extraction

In this section, the percentage extraction of all four solvents, hexane, heptane, toluene and cyclohexanol, are compared as a function of flow rate and flow ratio. Extraction with hexane and heptane was conducted in the PMMA microchannel. The feed flow rate was in the range of 0.1–1.2 mL/min, and the solvent flow rate was maintained in the range of 0.02–3.6 mL/min. The flow ratio varied from 0.25-3.0. The flow ratio is defined as the ratio of solvent to feed flow rate. The results of percentage extraction are shown in Fig. 3.1 and Fig. 3.2, respectively. At each flow ratio, the percentage extraction is observed to decrease with the increase in flow rate, as known. This may be because of the insufficient contact time between the solvent and the feed at higher flow rates. Next, the percentage extraction variation with respect to the flow ratio is seen from the same figures wherein the separation percentage increases with the flow ratio, which may be due to the proportionate increase in the solvent rate at the given conditions. A significant difference is observed at flow ratios 0.25 and 3. Although this seems to be advantageous, the flow ratio increase is not beneficial from the operations aspect as it would increase solvent consumption. Furthermore, the separation obtained at higher flow ratios is not very different. (vast difference). Overall, the percentage extraction of n-hexane is in the range of 12 to 22%.

The extraction performance of heptane is nearly similar to that of hexane; however, not more than a 2% increase. The maximum percentage is obtained at the lowest feed rate, where the maximum contact time between the solvent and the feed is obtained. At flow ratio 1, the percentage extraction lies in the range of 21.2–16.2%. In the same way, the Effect of the flow ratio for heptane is seen to follow the same trend as n-hexane. At a flow ratio of 0.25, the percentage extraction is in the range of 19.1–16.1%. The same flow ratio 3 varies from 24 − 17%, which is the maximum at the given conditions. Such lower percentage separation suggests that a single-stage extraction is not adequate to completely remove the solute from the feed and needs many more stages. A good solvent requires to have more distribution coefficient and high selectivity, where it can dissolve more solute selectivity and thereby reduce the required number of stages.

Extraction with toluene and cyclohexanol was conducted in a stainless steel microchannel as these solvents would damage the PMMA material. Hence for comparisons, the flow rates were adjusted such that their performance could be compared with hexane and heptane at equivalent residence times. In this case, the flow rate of the feed solutions was maintained from 0.04 to 0.45 mL/min, and the solvent was 0.01–1.35 mL/min. Similarly, the flow ratio varied from 0.25 to 3. The results of the percentage extraction of PA as a function of flow rate and flow ratio are shown in Fig. 3.3 and Fig. 3.4, respectively. The results indicate that the percentage extraction of toluene and cyclohexanol is twofold higher than the first two alkane solvents. At each flow ratio, the percentage extraction decreases with the increase in total flow rate, as expected. It should also be highlighted that at each flow ratio, there is about an 8–10% difference exists in the percentage separation. The overall percentage extraction of toluene changes from 26.1 to 41.2%. In cyclohexanol’s case, a relatively high separation is achieved then all the first three solvents. The overall percentage separation of cyclohexanol ranges from 56 to 60% with cyclohexanol. At each flow ratio, for the given flow rates, nearly 4 to 5% separation is achieved. Comparing all the solvents, hexane and heptane are observed to possess poor separation characteristics, while toluene and cyclohexanol exhibit relatively better performance. The extent of separation in terms of percentage separation implies that the solute is not completely removed from the feed in a single stage. This creates the need for several stages for the maximum removal of the solute. In the previous work, Singh et al., (2020) studied the extraction of PA with hexane & toluene and reported the highest percentage extraction of 36% and 19.4% for toluene and hexane, respectively at 0.1 mL/min total flow rate & flow ratio 1.

## 3.4 Number of Stages

It was intended to estimate the number of stages required for the maximum removal of the solute PA from the aqueous solution (feed) using all four solvents. Accordingly, the feed and the solvent were fed into the first stage wherein the solvent with the solute called extract was collected separately and the leftover feed called the raffinate was collected separately while the raffinate was fed into the subsequent microchannels to recover the PA. In each stage, the pure solvent was used for extraction. The flow rates and the flow ratio were adjusted such that the percentage extraction & extraction efficiency was maximum. Hence, the flow rate of the feed and the solvent was maintained equal in the range of 0.1 to 1.2 mL/min. Figure 3.13 to 3.16 shows the Effect of flow rate on concentration change, percentage extraction, extraction efficiency, volumetric mass transfer coefficient and the number of stages.

For the solvent hexane, Fig. 3.13(a) displays the Effect of multistage the change extraction in PA the raffinate and the extract. This plot displays the maximum separation obtained at equal flow rates of hexane and the feed at 0.1 mL/min. As expected, the concentration of PA in the raffinate decreases with each stage and reaches a constant value at the fifth stage and vice-versa in the extract. In each stage, nearly 20% extraction is achieved. Figure 3.13(b) shows the Effect of total flow rate on percentage extraction for each stage. As anticipated, the percentage extraction decreases with the flow rate. While it increases with each stage due to the increase in PA concentration difference in the successive stage, respectively. It should be noted that overall 74.3% solute is separated by hexane from the first to the fifth stage. In each stage, about 23% separation is achieved by hexane. Similarly, the extraction efficiency is plotted in Fig. 3.13(c). Similarly, the volumetric mass transfer coefficient is seen in Fig. 3.13(d). It is seen in the figures that efficiency decreases with each stage and the overall values lie between 97.96 and 90.66 at 0.2 mL/min. The maximum efficiency is obtained in stage 1, later the extraction efficiency decreases in each stage due to the decrease in concentration ratio (According to Eq. 2.2). In the same manner, the volumetric mass transfer coefficient is found to decrease with each stage because of the decrease in concentration difference between the saturated and the inlet PA concentration. The overall volumetric mass transfer coefficient is in the range of 0.45 to 0.88 s− 1.

Heptane exhibits an identical extraction performance like that of hexane. The extraction results of n-heptane are shown in Fig. 3.14(a-d). Here too, the PA concentration in the raffinate becomes stable at the fifth stage. The PA molarity in the extract is increased from range to 0.825 g/L. The overall PA separation by heptane is about 86.06% in five stages. The percentage extraction increases in each stage and the corresponding values are about 21% in the first two stages and from the third to the fifth stage, its value is in the following order 27.3%, 36.4%, 51.3% respectively. The maximum extraction efficiency obtained with heptane is about 98.9%. The extraction efficiency is found to decrease as well as with the stage, and the latter decreases with the flow rate as expected and also with each stage. The volumetric mass transfer coefficient also exhibits an opposite behaviour where the KLa increases with the increase in flow rate but decreases with each stage. The maximum range of KLa is obtained in stage 1 which is 0.16–0.91 s− 1. The next solvent used was toluene. The extraction characteristics of toluene are shown in Fig. 3.15(a-d). The concentration profile for PA in toluene for both the raffinate and extract is shown in Fig. 3.15(a). It is seen that nearly 3 stages are required for the solute concentration in the raffinate to reach the minimum value of 0.05 M. The percentage extraction plot further increases in stage results, that the maximum percentage extraction is reached at stage 3 which is about 50% at 0.2 mL/min total flow rate. The respective percentage separation in each stage is about 35.5%, 46.1% and 62.2%. The overall separation percentage through the 3 stages is approximately 87.1% which is higher than the first two solvents. The corresponding extraction efficiency is nearly 100% in stage 1 and decreases further with the flow rate and stages. The maximum KLa value is obtained in stage 1, which is nearly in the range of 0.25-1.0 s− 1. Finally, cyclohexanol was tested for determining the number of microchannel stages required to bring down the PA concentration in the raffinate to the lowest level. The extraction results of cyclohexanol are shown in Fig. 3.16(a-d). In this case, merely 2 stages are required by the solvent to achieve a PA concentration of 0.04 M in the raffinate where other solvents needed 3.5 stages for reaching the same molarity. The percentage of extraction of PA by cyclohexanol in the first stage is 57.2% while it is 89% in the second stage. The overall separation of the solute by cyclohexanol is about 95.3% which is very high than the other solvents. Likewise, the volumetric mass transfer coefficient for cyclohexanol is in the range of 0.2–1.2 s− 1 in the first stage which is the maximum.

## 3.5 Total annual Cost Analysis (TAC)

The TAC (Total Annual Cost) comparison was performed for the propionic acid extraction for the solvent used. TAC was calculated using the below Eq. (30,31).

TAC = Fixed cost + Operating cost (3.1)

The fixed costs in Eq. (3.1) represent the cost of purchase and installation of microchannel stack extractor. The operating cost includes the cost of raw materials, solvents, labour and electricity. To calculate the TAC, a few assumptions were made for processing 1000Kg per day, which are as follows.350 days of the operation in the year, the dollar price 1$ = 73.83 INR, the cost of the labour per day $3.88 and the cost of the electricity is about $0.092/unit. Although the maximum extraction efficiency of the solvents varies from 97.5–99.8%. An average efficiency value was adopted for the TAC calculation, which is 98.3%. Accordingly, the required number of microchannel units for a single stage is 4197, 3,723, 11,993, and 10,119 for the solvents hexane, heptane, toluene and cyclohexanol respectively. The capital cost per unit of the microchannel is found to be about $536 for hexane/heptane and $731 for toluene &cyclohexanol. Thus, for the above-mentioned repeating units and for the respective number of total stages the capital cost for hexane, heptane, toluene and cyclohexanol are in the following order $1,12,58,242, $1,10,19,504, $2,33,36,474, and $1,47,98,632 respectively. Thus, the capital cost calculation is about 86.2–94%, while the solvent cost contributes close to 2-12.4% of the TAC. These values suggest that repeating individual microchannels is not a feasible option.

The second case of the TAC estimation was based on the microchannel stack by adopting Arora 2010, microchannel stack in place of single microchannels unit is expected to minimize the capital cost significantly, Arora (2010) had used such a method i.e., stacked microchannel distillation unit for separating methanol-water mixture from a bio-diesel plant by distillation. The actual stack had 204 repeating units each having 426 microchannels for processing 1409 kg/h mixture. Based on this information, the cost of capital cost for processing 1000 Kg/day capacity is $73,627 for one stage. Hence, applying the same design for extraction the total number of the repeating unit for a single stage for the following systems PA-hexane, Pa-heptane, PA-toluene, PA-cyclohexanol are 9.85, 8.7, 28.1 and 24 units respectively. For the overall stages, the total number of stacks required is, nearly 50 units are required for the solvent hexane, 43 units for heptane,75 units for toluene and 48 units for cyclohexanol. The results of the TAC analysis is shown in Fig. 3.17 and 3.18. According to the calculations, the overall capital cost for the total number of stages in the order of solvents is $36,26,129, $32,02,774and $62,06,756, $35,34,096 for hexane, heptane, toluene and cyclohexanol respectively. This corresponds to 69–93% of the TAC. In this design, the solvent cost contributes the major portion of the operation cost which is nearly 6 to 30% of the same. Accordingly, the total annual cost is $41,83,379.75, $45,96,971.50, $66,47,213.10 and $49,10,453.00 for the above solvents in the same order.

In the third case, of TAC estimation, the cost of fabrication was based on the fabrication of the stacked microchannel units locally (in India). The cost of a stacked microchannel (consisting of 426 channels) for the hexane & heptane is $625, and for toluene & cyclohexanol, it is $876 respectively. Then the capital cost for the total number of stacked units (for overall stages) are $30,781.25, $27,187.5, $73,846.8 and $42,048forhexane, heptane, toluene & cyclohexanol respectively. Now the contribution of the capital cost is mere 1.9–14%of the TAC. On the other hand, now the TAC is dominated by the solvent cost only which corresponds to 81–97% of the TAC. In this new estimation, the total annual cost is $5,88,031, $14,21,384.5, $5,14,303.80 and $14,18,405.0 for the above order of solvents. It is seen that the overall annual cost is found to be less for toluene than all other solvents. Although cyclohexanol needs merely two stages the total annual cost is increased by its solvent cost. Thus, among the solvents, toluene seems to be a better option due to its relatively lowest TAC. If a suitable co-solvent can be employed with cyclohexanol the overall operation cost may significantly be reduced to a desirable value. Moreover, a solvent even more effective than these is expected to reduce the overall annual cost. For microchannel stack was calculated based on the local charges i.e. charges incurred in our region for the stack design.

**Conclusion**

Multistage extraction for the separation of propionic acid (7.04 wt%) from its aqueous solution by solvent extraction using n-hexane, heptane, toluene and cyclohexanol was conducted in microchannels. These solvents were chosen based on their ease of recovery by distillation except the last. The Effect of flow rate, flow ratio, stages on PA concentration in raffinate and extract, percentage extraction, extraction efficiency and volumetric mass transfer coefficient were studied. Besides, a total annual cost for the multistage extraction process was also made.

The percentage extraction is found to decrease with the total flow rate for all the solvent. On the other hand, the same increased with the increase in flow ratio. The overall percentage extraction for hexane is in the range of 12 to 22%, for heptane it is in the range of 16–24%. The percentage separation of PA with toluene and cyclohexanol is in the range of 41.2 to 26% and 60.5 to 52%. Thus toluene and cyclohexanol are found to be better than the first two solvents considered. Percentage extraction increased with flow ratio in all the cases however the increment was not more than 1–3% in the case of hexane and heptane but with toluene and cyclohexanol, it was in the range of nearly 4–5%.

The Effect of flow rate and flow ratio on extraction efficiency was studied. Extraction efficiency decreased with the total flow rate and reached the maximum at the lowest flow rate and flow ratio 1. The Effect of flow ratio indicated increased extraction efficiency with the flow ratio increase from 0.25 to 1.0 and the same decreased with further increase in flow ratio from 1.5 to 3.0. Extraction efficiency was also evaluated for the individual solvents. The maximum value obtained for hexane is in the range of 97.5–87.1%. Similarly, heptane is in the range of 96.3 to 86.3%. The highest extraction efficiency was obtained for toluene and cyclohexanol that is 97.7–87.5% and 99.8–93.8% respectively. The volumetric mass transfer coefficient was evaluated with respect to flow rate and flow ratio. It was found that the KLa increased with the increase in flow rate. The maximum KLa was obtained at flow ratio 0.25 which is for hexane in the range of 0.15–0.87 s− 1, for heptane 0.08–0.91 s− 1, for toluene it is 0.17–1.12 s− 1 and finally, for cyclohexanol, it is the range of 0.16–1.23 s− 1

The number of stages required for the maximum removal of PA from the feed was estimated. Also, the Effect of stages on percentage extraction, extraction efficiency and the volumetric mass transfer coefficient was studied. Consecutively, the required number of stages for hexane and heptane was 5 and for toluene, it is 3 and finally, for cyclohexanol, it is 2. The percentage extraction increased with each stage due to the increase in PA concentration difference. In the successive stages. Hexane recovered 74.3% of PA from the feed-in five stages. In each stage, about 23.7% of separation was achieved. Similarly, heptane removed nearly 86% of PA overall and in the first two stages about 22% separation was obtained and from the third to the fifth stage, it increased from 27 to 51.3%. Toluene required three extraction stages for separating 87.1% PA. The separation obtained in the three stages is around 35.5%, 45.1% and 62.2% respectively. In the same way, cyclohexanol required only 2 stages to separate nearly 95.3% PA overall. The separation obtained in the two stages is 57% and 89% respectively.

The overall volumetric mass transfer coefficient decreased with each multiple extraction stage for all the solvents. The overall KLa values are in the range of 0.15-0.87s− 1, 0.08-0.91s− 1, 0.17–1.12 s− 1 and 0.16-1.23s− 1 for hexane, heptane, toluene and cyclohexanol respectively. The effect of stages on extraction efficiency was evaluated. Maximum efficiency was obtained in stage 1 for all the solvents. The extraction efficiency values for the above solvents (in the same order) are 97.9–86.3%, 98.9–86.2%, 99.7–89.5% and 99.8–93.8%, respectively. The extraction efficiency decreased with the increase in stage numbers; however, the change is not so significant, which is only 1–3%.

The total cost analysis was made for the multistage extraction of propionic acid involving all the solvents. For processing 1000 Kg/day, 4197, 3723, 11,993 and 10,119 individual microchannels were required. In this design, the capital cost of the microchannels contributed about 86–96% of the TAC, while the solvent cost contributed 2–12% of the overall costs.

In the second estimation technique, a microchannel stack design containing 426 microchannels was adopted from the literature to estimate TAC comparison. Accordingly, about 50, 84 and 48 total units were required for the solvents one to four listed at the beginning. In this design, the capital cost of the microchannel stack accounts for 8-33.9% and the solvent costs were increased to 45-81.5% of the overall TAC. The third case involved the TAC estimation based on the fabrication cost in India. The results indicated that the TAC is greatly reduced and the capital cost for the stacks accounts for only 1.2–14.3% of the TAC. The solvent cost, on the other hand, contributes 81–96% of the TAC. Overall, solvent toluene contributes the minimum TAC than the other solvents. Thus, it can be concluded that effective solvents can greatly reduce the number of stages required for the maximum recovery of PA from the feed. At the same time, the TAC of the process can be greatly reduced by employing a suitable, efficient solvent.