Accepted Manuscript Effect of liquid surface tension on the filtration performance of coalescing filters Feng Chen, Zhongli Ji, Qiangqiang Qi PII: DOI: Reference:
S1383-5866(18)32271-8 https://doi.org/10.1016/j.seppur.2018.09.035 SEPPUR 14933
To appear in:
Separation and Purification Technology
Received Date: Revised Date: Accepted Date:
2 July 2018 26 August 2018 10 September 2018
Please cite this article as: F. Chen, Z. Ji, Q. Qi, Effect of liquid surface tension on the filtration performance of coalescing filters, Separation and Purification Technology (2018), doi: https://doi.org/10.1016/j.seppur. 2018.09.035
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Effect of liquid surface tension on the filtration performance of coalescing filters Feng Chen, Zhongli Ji*, Qiangqiang Qi Beijing Key Laboratory of Process Fluid Filtration and Separation, College of Mechanical and Transportation Engineering, China University of Petroleum, Beijing 102249, PR China *Corresponding author. Tel./fax: +86 10 89734336. Address: College of Mechanical and Transportation Engineering, China University of Petroleum, Beijing 102249, China. E-mail address:
[email protected]. Abstract Fibrous coalescing filters are used to collect droplets from air streams and are widely applied in the process industry but liquid surface tension affects the filter performance. In this work, the wet pressure drop, saturation, liquid distribution, and filtration efficiency of both oleophilic and oleophobic filters were investigated using four liquids with different surface tensions but similar densities and viscosities to produce aerosols. The results showed that an increase in surface tension accelerated the evolution of the wet pressure drop of filters. The jump pressure drop increased significantly while the excess channel pressure drop was constant with growing surface tension. The average saturation in the channel layers decreased as the surface tension increased and the decrease rate was much greater in oleophobic than oleophilic filters. The liquid distribution was also influenced by the surface tension. In oleophilic filters, higher surface tension led to a smaller number of channels and a larger size per channel. The channel boundaries were difficult to distinguish when the relative wettability between the liquid and filter media was very high or very low. Both oleophilic and oleophobic filters showed higher filtration efficiency at steady state in both sub-micrometer and micrometer droplet-size ranges when separating liquids with a larger surface tension. Additionally, superoleophilic and superoleophobic filters were prepared through a dip-coating method. When compared with untreated (oleophilic) and superoleophilic filters, the superoleophobic filters had the best overall performance for each liquid type due to the greatly reduced liquid hold-up. Keywords: Coalescence; Liquid aerosols; Filtration; Surface tension; Saturation 1. Introduction Coalescing filters are widely used to remove liquid aerosols consisting of droplets with 1
diameters of several nanometers to tens of micrometers from air streams. These aerosols are undesirable by-products of many industrial processes, including natural gas purification, engine crankcase ventilation, mechanical and metal cutting, and compressed air filtration [1,2]. In a natural gas long-distance transmission pipeline, for example, the presence of droplets may cause corrosion and erosion to the pipeline and instrument equipment, resulting in gas leakage and equipment shutdown [3]. Therefore, it is of great importance to study the removal of droplets from air streams. Glass fiber is one of the most effective and commonly used coalescence filter media. The majority of glass-fiber materials studied in existing works are oleophilic, although oleophobic filter media are also used in field applications. Superoleophobic glass-fibrous media, in particular, has received little attention in the industry. Gas-liquid coalescence filtration is a very complex process in which the filter performance is affected by a range of parameters, including properties of the liquid (viscosity and surface tension), operation conditions (gas velocity and liquid loading rate), and filter structure and properties (fiber diameter, pore size, thickness, and surface energy). Kampa et al. [4] conducted a study of the wet pressure drop and liquid transport through two typical coalescence filter media (one wettable and the other non-wettable). The results showed that a channel pressure drop and a jump pressure drop were generated by liquid transport through channels and across interfaces, which were confirmed with the help of a combination of polymerization techniques and scanning electron microscopy (SEM). A new phenomenological “jump-and-channel” model was then proposed qualitatively for the first time [4]. Further, the dependence of channel parameters on the layers, liquid delivery rate, and oil viscosity, as well as the effects of liquid loading rate and media properties on jump parameters, were analyzed to validate the model [5]. Subsequently, according to this model, the effect of a drainage layer on the filtration performance of coalescing filter cartridges at steady state and the saturation in the filtration process was evaluated experimentally by Chang et al. [6,7]. Kolb et al. [8] investigated the dependence of differential pressure, oil saturation level, and liquid distribution on the flow velocity of air and oil loading rate for oleophilic filters. Chang et al. [9] analyzed the pressure drop and saturation of non-wettable coalescing filter cartridges at different loading rates, and the influence of a drainage layer on the saturation and liquid distribution of oleophobic coalescing filter cartridges was also evaluated experimentally [10]. Chen et al. [11] systematically investigated the effect of pore size and layers on coalescence performance of both oleophilic and oleophobic filters. Although these 2
studies have contributed to the development of gas-liquid coalescence filtration mechanisms, a systematic study on the effect of liquid surface tension has not yet been carried out, despite the significance to the improvement of the jump-and-channel model. Besides, Kampa et al. [5] attempted to vary the surface tension of oils they used during the validation of the model, but no suitable test liquid was found. Hence, this work focuses on the study of the effect of liquid surface tension. In the coalescence filtration process, the interaction between fibers and droplets directly influences the morphology of droplets on the fiber surface, thus also affecting the droplet coalescence and migration process, as well as the drop re-entrainment. Hence, the saturation and liquid distribution in the filters may be changed by this interaction. Finally, the pressure drop and filtration efficiency—the two most important filtration performance parameters—are also likely to be greatly affected. Such fiber-droplet interaction is closely related to both the liquid surface tension and the fiber surface energy. For the fiber filter media with a fixed surface energy, droplets with the same density and viscosity but various surface tensions will present different morphological characteristics on the media surface, which is generally characterized and distinguished by the static contact angle. In a few cases, the filter media will present a transition of wettability between wettable and non-wettable for the liquid with different surface tensions. In addition, a porous fibrous filter can be regarded as a system of capillaries [12]. The capillarity, which can affect the features of the liquid film on the inlet or outlet surface, is highly related to the surface tension. In other words, the surface tension theoretically has an important effect on both the channel and jump parameters and hence, there is a growing need to analyze the effect of surface tension on the filtration performance of coalescing filters. Besides, the resultant findings may help to provide technical support for field applications due to varying types of liquid in practical conditions. There are a few studies on the effect of surface tension. The work carried out by Briscoe et al. [13] indicated that the distribution patterns of droplets on a single fiber were mainly affected by the surface tension, contact angle, droplet volume, and fiber diameter. Letts et al. [14] conducted a series of experimental tests using bis(2-ethylhexyl) sebacate (BEHS) and a straight mineral oil. The results showed that the differences in the pressure drop increase and the magnitude of efficiency reductions were not significantly dependent on the type of liquid used. Contal et al. [15] tested the influence of physicochemical characteristics of three different liquids and found that a larger surface tension offered a higher specific surface area, resulting in faster clogging of the filter. Their results also 3
corroborated the effect of surface tension on the arrangement of drops on a single fiber, which was proposed by Briscoe et al. [13]. Hsiao et al. [16] conducted an experimental study on the transition pressure drop characteristics of fibrous filters that were loaded with four different oil particles. The results showed that the surface tension was the dominant physical property that affected the pressure drop of the glass-fiber filter, while the pressure drop profiles of cellulose filter media were influenced by both oil surface tension and viscosity due to the absorption effect of cellulose. The effect of surface tension may be different in oleophilic and oleophobic filters due to the difference in fiber surface energy and so this effect should be analyzed separately for these two kinds of filters. Meanwhile, some studies have determined the effect of fiber surface energy on the filter performance. Patel et al. [17] conducted a study of the effect of surface energy of woven drainage channels. The results showed that the filter embedded with Teflon fiber drainage channels at 45° downward angles had the best overall performance. Chang et al. [6,10] investigated the effect of non-woven drainage layers with various wettabilities on the filtration performance of oleophilic and oleophobic glass-fiber filter cartridges. Letts et al. [14] selected three different materials of glass, polyester, and polyaramid fibers to improve mist filters, and found that utilizing higher surface energy fibers allowed lower levels of liquid retention and thus lower pressure drop at steady state. Furthermore, it should be mentioned that all the above research was carried out using traditional oleophilic and oleophobic filter materials. The materials with super wettability (superoleophilicity and superoleophobicity) have been successfully applied and broadly reported in a range of other fields, such as corrosion resistance, self-cleaning, anti-fogging, electronic protection, drag reduction, and oil/water separation [18-20]. However, in the gas-liquid coalescence filtration field, research on the materials with super wettability remains undeveloped despite the preparation and application of such materials being recommended as a key area for further work [2]. To the best of our knowledge, only one study, which we previously published, has reported a filter with super wettability [21] and where a test aerosol of di-ethyl-hexyl-sebacate (DEHS) was selected according to EN779: 2012 [22]. This aerosol is widely used in performance testing of high-efficiency particulate air (HEPA) and ultra-low-penetration air (ULPA) filters. The results showed that the superoleophobic filter had the best overall performance for both small and large oil mists. However, it is necessary to verify whether such a superoleophobic filter is still beneficial in separating aerosols with other properties and, if so, whether it has a similar relationship with surface tension as that of a conventional 4
oleophobic filter. In particular, a novel concept of the “bounce-collide-drain” mechanism was proposed in [21] to interpret the improvement of the filtration performance after modification, but there is still a gap in the correlation between the improved performance and the change in the liquid hold-up in the filters. The aim of the present work was to systematically analyze the effect of liquid surface tension on the filtration performance of coalescence filters using two typical glass-fiber filter media and four different liquids in terms of wet pressure drop, saturation, liquid distribution features, and filtration efficiency. By combining with the saturation results, the advantage of superoleophobic filter media to collect droplets with different surface tensions was also validated using the as-prepared superoleophilic and superoleophobic filter materials through a wet chemical coating technique. 2. Experimental details 2.1. Filter materials The filter materials used in this study were glass microfiber media, one of which was oleophilic (W) and the other oleophobic (NW) and both are widely used in industrial gas-liquid coalescing filters. The properties of the filter materials are listed in Table 1. Wettability was differentiated by measuring contact angles with DEHS (dynamic viscosity 23 mPa s, density 912 kg/m3 at 25 ºC), which is a standard test oil for the performance evaluation of filters [22]. The filter was defined as oleophilic if the contact angle was 150°, which is defined as the lower limit of contact angle for superphobic materials. In other words, when the wettability between the liquid and filter media decreased to a certain degree (one extreme case was superphobic), the liquid distribution did not show the sharply defined boundaries between the wetted and non-wetted areas. This was surprising and is most likely related to the method of transporting the liquid between the filter layers.
Fig. 5. Liquid distribution patterns of layer 6 within eight-layer filters at steady state. The top and bottom four images represent the distribution patterns for oleophilic and oleophobic filters, respectively, loaded with different liquid aerosols. From left to right: liquid G, liquid D, liquid B, and liquid P.
Generally, the liquid channels begin to form from the second layer in multilayer filters. It was also the case here when oleophobic filters collected droplets with low surface tension. A different result appeared, however, if the wettability between the liquid and the filter was very poor. Take the 15
case of oleophobic filters collecting aerosols of liquid P in which the individual layers of filter materials were obtained by carefully disassembling the filters after reaching steady state. It was found that a quite obvious layer of liquid film existed on the internal surface of every two neighboring filter layers, as shown in Fig. 6. We speculate that this film is the primary reason for the dramatically decreased channel transport. Due to the effect of light on the photography, Movie 1 (shown below) better shows the film from different directions. Apparently, the film illustrated here differed greatly from the one previously mentioned (Fig. 3) and also from that presented in [4] and [6]. This film was more suitable to be termed an “apparent film” because it would not be sucked into the filter media after the airflow was shut off. Also, no SEM or complicated polymerization technique was required to observe it and the film could be directly observed by the naked eye. Although the existence of the film would introduce an error into the measurement of saturation, repeated experimental results showed good repeatability and reproducibility. It should be noted that the appearance of this apparent film was not the result of an insufficient clamping force on the filter layers because the absolute leak-tightness was confirmed and all experimental runs applied the same clamping force and operation procedure. Therefore, the channel boundaries in the sixth layer were not clearly defined, as shown in Fig. 5h. Apart from the first layer, the liquid distribution patterns in other layers were similar to those of the sixth layer. The first layer had a comparatively better collection effect because its entire upstream surface was in contact with the aerosols, which led to a significantly higher liquid hold-up (Fig. 3b). Overall, when the relative wettability between the liquid and fiber materials varied from philicity to phobicity and then to superphobicity, the liquid channels changed from a number of small ones (difficult to distinguish and count) to a few large ones (easy to identify and count), and finally back to a distribution pattern that was difficult to distinguish.
16
Fig. 6. An apparent film on the surface of layer 6 within an eight-layer oleophobic filter at steady state. The filter material used was NW and the liquid was P. Movie 1. The apparent film on the surface of layer 6 within an eight-layer oleophobic filter at steady state. The filter material used was NW and the liquid was P.
3.1.2. Filtration efficiency of filters Because of the different principles employed by the two particle-size spectrometers SMPS and APS to measure the aerosol concentrations, the results were studied separately. SMPS is based on the scanning electrical mobility diameter while APS is based on the aerodynamic diameter. The droplet-size ranges measured by the SMPS and APS were 0.05–0.8 μm and 0.8–20 μm, respectively, and thus the droplets were named sub-micrometer-sized and micrometer-sized droplets. Fig. 7 shows the filtration efficiency of different filters for sub-micrometer-sized droplets at steady state. A larger surface tension produced a higher filtration efficiency as expected, corresponding to the higher pressure drop in Fig. 2. This finding was applicable to both single-layer and eight-layer filter media. The results of single-layer filters were used to make a comparison with those of modified filters (which will be discussed in Section 3.2). In general, the sub-micrometer-sized droplets are mainly captured by the mechanisms of diffusion and interception, while the inertia effect is poor. This indicates that the increase in the interstitial velocity is disadvantageous to the collection of droplets in this range. According to Fig. 5, at higher surface tensions, less area on the surface of the oleophilic filter layer was wetted, leading to a lower interstitial velocity. Therefore, for sub-micrometer-sized droplets, the filtration efficiency of the oleophilic filters increased with increasing surface tension. Similarly, for oleophobic filters, as mentioned earlier, the rate of decrease of saturation with the increasing surface tension was remarkably higher than that for the oleophilic filters, thereby contributing to the effective separation of sub-micrometer-sized droplets. The one-layer oleophobic filter had a significantly lower filtration efficiency for liquid G than for the other three liquids. This is most likely due to the obviously smaller contact angle of liquid G on the surface of the filter material NW (Table 2). Note, however, that due to the differences in the structural parameters, there was not necessarily an association between the filtration efficiency of oleophilic and oleophobic filters.
17
90
100.0
(a) Oleophilic
1-layer 8-layer
100 Steady-state SMPS efficiency (%)
Steady-state SMPS efficiency (%)
100
99.6
80 99.2 70 98.8 60 98.4
50 40
98.0 G
D
B
P
90
100.0
(b) Oleophobic
1-layer 8-layer
99.8
80 99.6 70 99.4 60 99.2
50 40
99.0 G
D
B
P
Liquid
Liquid
Fig. 7. Steady-state filtration efficiency of one- and eight-layer (a) oleophilic and (b) oleophobic filters measured by the SMPS and loaded with different liquid aerosols.
Fig. 8 shows the filtration efficiency of different filters for micrometer-sized droplets at steady state. The droplets in this size range were mainly captured by the interception and inertial effect. As stated above, a larger surface tension produced a lower interstitial velocity, which in turn caused a decrease in the filtration efficiency for micrometer-sized droplets. This seems inconsistent with the results shown in Fig. 8, i.e., that the filtration efficiency increased with the surface tension. It should be noted, however, that due to the drop re-entrainment, an increase in the downstream aerosol concentration of the test filter was prone to occurring for the droplets within this size range. Thus, we suggest that the drop re-entrainment was more likely to appear during the separation of liquids with a lower surface tension. For oleophilic filters, the thickness of the liquid film on the outlet surface increased with the surface tension. Generally, a thicker liquid film has greater stiffness. Therefore, due to the low face velocity, the film was more likely to be affected by the airflow for the liquid with a lower surface tension. This might increase the rate of bubble formation and then bursting [26] from the liquid film, leading to the higher possibility of drop re-entrainment. For oleophobic filters, the most likely reason for the drop re-entrainment was the weaker droplet fiber adhesion forces, which are related to the “clamshell” droplets formed on the fiber materials [27]. The saturation and interstitial velocity were higher at lower surface tension, thereby increasing the rate of occurrence for the drop re-entrainment.
18
99.8
100.00
(a) Oleophilic
1-layer 8-layer
100.0 Steady-state APS efficiency (%)
Steady-state APS efficiency (%)
100.0
99.96
99.6
99.92
99.4
99.88
99.2
99.84
99.0
99.80 G
D
B
99.6
1-layer 8-layer
99.999
99.2
99.998
98.8
99.997
98.4
99.996 G
P
100.000
(b) Oleophobic
Liquid
D
B
P
Liquid
Fig. 8. Steady-state filtration efficiency of one- and eight-layer (a) oleophilic and (b) oleophobic filters measured by the APS and loaded with different liquid aerosols.
3.2. Improvement of filtration performance by superoleophobic filters In the above section, the oleophilic and oleophobic filter media that are typically used in the process industry were studied. Due to the rapid development of bio-inspired and modified technology, the media with super wettability have received a wide attention. The wet chemical coating method is an effective method to modify the wettability of glass-fiber media. In our previous work [21], glycerol propoxylate triglycidyl ether (GPTE) and perfluoroalkyl acrylic copolymer (PFAP) water solution were applied as superoleophilic and superoleophobic treatments, respectively, onto the non-woven glass fiber through a dip-coating method. Herein, three types of single-layer filters (superoleophilic, untreated (oleophilic), and superoleophobic) were obtained based on the same control material and treating methods as in [21], which describes the detailed preparation procedures. The contact angles of different liquids on the treated materials were also tested previously and the as-prepared superoleophobic filter materials had a contact angle >150° to the liquid with a surface tension >31 mN/m. Hence, only liquid G did not present a superoleophobic characteristic, but an obviously oleophobic property was shown with θ > 110°. It may be necessary to promote the treatment method to make the filter media superoleophobic to the liquid with such a low surface tension, which will be studied in future work. For convenience, the filter treated by the PFAP water solution is still termed the superoleophobic filter. The filter materials treated by the GPTE solution showed an obviously superphilic character to all four test liquids and the droplets were immediately sucked into the filter media within 1 s. 19
Fig. 9 shows the wet pressure drop profiles of single-layer superoleophobic filters. For clarity, the results of the superoleophilic and untreated filters, which were similar to that of the oleophilic filters in Fig. 2a, are not included in the figure. As shown, in the jump stage, the increase rate of the wet pressure drop with loading time became greater as the surface tension increased. Meanwhile, a larger surface tension led to a greater pressure drop of filters at steady state due to the poorer wettability and attendant higher flow resistance to liquid. Hence, the surface tension dependence of the wet pressure drop is similar between oleophobic and superoleophobic filters. 12 Liquid P
Wet P (kPa)
9
Liquid B
6
Liquid D
3
Liquid G
0 0
40
80
120
160
200
240
Time (min)
Fig. 9. Wet pressure drop profiles of one-layer superoleophobic filters in the filtration process.
Fig. 10 shows the filtration efficiency of single-layer filters (untreated and treated with wet chemical coating methods) at steady state for sub-micrometer-sized and micrometer-sized droplets as measured by the SMPS and APS, respectively. In both size ranges of droplets, the superoleophobic filters always showed the highest filtration efficiency for the collection of each liquid type. Moreover, similar to the oleophobic filters, their collection efficiency increased with the increase in surface tension. For the superoleophilic and untreated filters, the filtration efficiencies were almost the same at low surface tension. As the surface tension increased, the difference in the filtration efficiency of these two kinds of filters increased slightly but both types of filters were substantially less efficient than the superoleophobic filters.
20
100.0
90
99.6
80
99.2
98.8
70
60
superoleophilic-SMPS untreated-SMPS superoleophobic-SMPS
superoleophilic-APS untreated-APS superoleophobic-APS
G
B
D
APS filtration efficiency (%)
SMPS filtration efficiency (%)
100
98.4
P
Liquid
Fig. 10. Steady-state filtration efficiency of one-layer filters (untreated and treated with wet chemical coating methods) as measured by the SMPS and APS.
To compare the overall performance of single-layer filters that were untreated and treated with wet chemical coating methods, the filtration efficiency as measured by the SMPS and APS (Fig. 10) and the steady-state pressure drop data were substituted into Eq. (2). Then, the quality factors of filters for sub-micrometer-sized and micrometer-sized droplets were obtained, as shown in Fig. 11. In both droplet-size ranges, the superoleophobic filters showed the best overall filtration performance for each kind of liquid aerosol. In contrast, the superoleophilic treatment had little influence on the overall filter performance. In fact, the surface tension range of the liquids used in this work included that of common liquid impurities in the industrial oil-mist filtration process. Therefore, these results confirm the significant advantages of superoleophobic filters in the field of liquid aerosol coalescence filtration.
Quality factor (1/kPa)
2.5 2.0
superoleophilic-SMPS untreated-SMPS superoleophobic-SMPS
superoleophilic-APS untreated-APS superoleophobic-APS
1.5 1.0 0.5 0.0 G
D
B
P
Liquid
Fig. 11. Steady-state quality factors of one-layer filters that were untreated and treated with wet chemical coating methods. 21
Fig. 12 shows the steady-state saturation of single-layer filters that were untreated and treated with wet chemical coating methods. For each liquid type, the saturation of the superoleophobic filters was significantly lower (about half that of the other two filters or even less). As stated in [6] and [27], in the coalescence filtration process, the liquid hold-up in the filters continuously increased while the corresponding filtration efficiency gradually decreased until a steady state was reached. This implies that the only reason for the reduction in the filtration efficiency from the initial to steady state was the increase in the liquid hold-up. Accordingly, it is reasonable to conclude that the fundamental cause of the high filtration efficiency of superoleophobic filters was the efficient suppression of the increase in the liquid hold-up. A subtler feature was found in Fig. 12, i.e., that by comparison with the superoleophilic and untreated filters, the steady-state saturation of superoleophobic filters showed a decline with the increase in surface tension. Based on the positive relationship between the filtration efficiency and the surface tension in Fig. 10, it confirms that the decrease in the liquid hold-up in the filters contributed to the improvement in the filtration performance. In addition, for the superoleophobic filters, the variation in the saturation with the surface tension was similar to that for the oleophobic filters (Fig. 3b), which suggests the similar effects of surface tension on these two kinds of filter media. 1.0
Saturation (-)
0.8 0.6 0.4 0.2 0.0
superoleophilic untreated superoleophobic
G
D
B
P
Liquid
Fig. 12. Steady-state saturation of one-layer filters that were untreated and treated with wet chemical coating methods.
4. Conclusions In this work, the effect of liquid surface tension on the coalescence filtration performance of glass-fiber filter media was studied experimentally. Four kinds of liquid were selected with similar 22
densities and viscosities but different surface tensions ranging from 20.6 to 46.3 mN/m. The filters were fabricated into sandwiched structures using two types of filter materials (one oleophilic and one oleophobic) separately, both of which are typically applied in the process industries. The results showed that the jump pressure drop was highly dependent on the surface tension because the increased surface tension results in an increase in both the capillary retention force and the thickness of the liquid film. Conversely, the excess channel pressure drop did not change with the variation in the surface tension. Particularly for oleophobic filters, as the surface tension increased, the excess channel pressure drop was balanced by the increase in flow resistance to liquid and the decrease in saturation. However, during the filtration process, both the rate of increase of the jump pressure drop and the excess channel pressure drop with time increased as the surface tension increased, leading to earlier cut-off points between the filter stages. The average saturation of the channel region, which was inversely correlated to the surface tension, was more affected by the surface tension in oleophobic filters than in oleophilic filters. Moreover, the evolution of the liquid distribution patterns with the surface tension showed the opposite characteristics in oleophilic and oleophobic filters. In the case of oleophilic filters collecting low-surface-tension liquids and oleophobic filters collecting high-surface-tension liquids, the channel patterns were unclear and difficult to distinguish. The former case was due to the formation of a number of small channels and liquid bridges, while the latter case was the result of a layer of liquid film that existed at the interface of the filter layers. We termed this film “apparent film” because it was visible to the naked eye and was not sucked into the filter media by capillary effect. Nevertheless, in the case of oleophilic filters collecting high-surface-tension liquids and oleophobic filters collecting low-surface-tension liquids, bright liquid channels of small number and large size were present. In addition, for both sub-micrometer-sized and micrometer-sized droplets, larger surface tensions led to higher filtration efficiency at steady state, which applied to both oleophilic and oleophobic filters. The filtration performance of filters with super wettability (superphilicity and superphobicity) for different liquid aerosols was also evaluated. The effect of liquid surface tension on the superoleophobic filters was similar to the oleophobic filters case. Most importantly, when compared with the untreated (oleophilic) and superoleophilic filters, the superoleophobic filters showed the best overall filtration performance for each liquid type studied. The significant decrease in the liquid hold-up by the surface modification greatly contributes to the filtration performance. 23
Acknowledgment Authors acknowledge the financial support from National Natural Science Foundation of China under grant No. 51376196. References [1] G. Gonfa, M.A. Bustam, A.M. Sharif, N. Mohamad, S. Ullah, Tuning ionic liquids for natural gas dehydration using COSMO-RS methodology, J. Nat. Gas Sci. Eng. 27 (2015) 1141-1148. [2] R. Mead-Hunter, A.J.C. King, B.J. Mullins, Aerosol-mist coalescing filters – a review, Sep. Purif. Technol. 133 (2014) 484-506. [3] J.S. Stahley, Dry Gas Seals Handbook, PennWell Corporation, Tulsa, 2005. [4] D. Kampa, S. Wurster, J. Buzengeiger, J. Meyer, G. Kasper, Pressure drop and liquid transport through coalescence filter media used for oil mist filtration, Int. J. Multiph. Flow 58 (2014) 313-324. [5] D. Kampa, S. Wurster, J. Meyer, G. Kasper, Validation of a new phenomenological “jump-and-channel ” model for the wet pressure drop of oil mist filters, Chem. Eng. Sci. 122 (2015) 150-160. [6] C. Chang, Z.L. Ji, F.Y. Zeng, The effect of a drainage layer on filtration performance of coalescing filters, Sep. Purif. Technol. 170 (2016) 370-376. [7] C. Chang, Z.L. Ji, J.L. Liu, The effect of a drainage layer on the saturation of coalescing filters in the filtration process, Chem. Eng. Sci. 160 (2017) 354-361. [8] H.E. Kolb, J. Meyer, G. Kasper, Flow velocity dependence of the pressure drop of oil mist filters, Chem. Eng. Sci. 166 (2017) 107-114. [9] C. Chang, Z.L. Ji, J.L. Liu, Pressure drop and saturation of non-wettable coalescing filters at different loading rates, AIChE J. 64 (2018) 180-185. [10]
C. Chang, Z.L. Ji, J.L. Liu, The effect of a drainage layer on saturation and liquid distribution of oleophobic coalescence filters, Sep. Purif. Technol. 194 (2018) 355-361.
[11]
F. Chen, Z.L. Ji, Q.Q. Qi, Effect of pore size and layers on filtration performance of coalescing filters with different wettabilities, Sep. Purif. Technol. 201 (2018) 71-78.
[12]
B.J. Mullins, R.D. Braddock, Capillary rise in porous, fibrous media during liquid immersion, Int. J. Heat Mass Transf. 55 (2012) 6222-6230. 24
[13]
B.J. Briscoe, K.P. Galvin, P.F. Luckham, A.M. Saeid, Droplet coalescence on fibres, Colloids Surf. 56 (1991) 301-312.
[14]
G.M. Letts, P.C. Raynor, R.L. Schumann, Selecting fiber materials to improve mist filters, J. Aerosol Sci. 34 (2003) 1481-1492.
[15]
P. Contal, J. Simao, D. Thomas, T. Frising, S. Callé, J.C. Appert-Collin, D. Bémer, Clogging of fibre filters by submicron droplets. Phenomena and influence of operating conditions, J. Aerosol Sci. 35 (2004) 263-278.
[16]
T.C. Hsiao, D.R. Chen, Experimental observations of the transition pressure drop characteristics of fibrous filters loaded with oil-coated particles, Sep. Purif. Technol. 149 (2015) 47-54.
[17]
S.U. Patel, P.S. Kulkarni, S.U. Patel, G.G. Chase, The effect of surface energy of woven drainage channels in coalescing filters, Sep. Purif. Technol. 87 (2012) 54-61.
[18]
H. Wang, H. Zhou, A. Gestos, J. Fang, T. Lin, Robust, superamphiphobic fabric with multiple self-healing ability against both physical and chemical damages, ACS Appl. Mater. Interfaces 5 (2013) 10221-10226.
[19]
K. Liu, Y. Tian, L. Jiang, Bio-inspired superoleophobic and smart materials: Design, fabrication, and application, Prog. Mater. Sci. 58 (2013) 503-564.
[20]
T.A. Otitoju, A.L. Ahmad, B.S. Ooi, Superhydrophilic (superwetting) surfaces: A review on fabrication and application, J. Ind. Eng. Chem. 47 (2017) 19-40.
[21]
X. Wei, F. Chen, H.X. Wang, H. Zhou, Z.L. Ji, T. Lin, Efficient removal of aerosol oil-mists using superoleophobic filters, J. Mater. Chem. A 6 (2018) 871-877.
[22]
EN 779: 2012 Particulate Air Filters for General Ventilation – Determination of the Filtration Performance, British Standards Institution, 2012.
[23]
C.N. Davies, Air Filtration, Academic Press, London, 1973.
[24]
ISO 14419-2010 Textiles - Oil Repellency - Hydrocarbon Resistance Test, ISO, 2010.
[25]
R.C. Brown, Air Filtration: An Integrated Approach to the Theory and Applications of Fibrous Filters, Pergamon Press, New York, 1993.
[26]
S. Wurster, J. Meyer, H.E. Kolb, G. Kasper, Bubbling vs. blow-off – On the relevant mechanism(s) of drop entrainment from oil mist filter media, Sep. Purif. Technol. 152 (2015) 70-79. 25
[27]
B.J. Mullins, R. Mead-Hunter, R.N. Pitta, G. Kasper, W. Heikamp, Comparative performance of philic and phobic oil-mist filters, AIChE J. 60 (2014) 2976-2984.
26
0.6
0.8
0.2 0.0
G
D B Liquid
0.4
P
Liquid G Liquid D Liquid B
0.2
0.6
0.0
3
0.4
2
0.2
1
0.0
G
D
B
P
0
Liquid
0.4 Liquid P
0.2
Liquid P
0.6
3
Oleophobic
0.4
V (cm )
Saturation (-)
0.8
1.0
0.6
Average S of channel region
Oleophilic
Saturation (-)
1.0
Average S of channel region
Graphical abstract
Liquid G
Liquid D
Liquid B
0.0 1
2
3
4
5
6
7
8
1
Layer
2
3
4
5
Layer
27
6
7
8
Highlights
Increase of surface tension accelerated the evolution of wet pressure drop of filters.
Jump pressure drop increased linearly while excess channel pressure drop remained constant with increasing surface tension.
Channel boundaries were difficult to distinguish when relative wettability between liquid and filter media was very high and very low.
Larger surface tension led to higher filtration efficiency at steady state.
Superoleophobic filters showed the best overall performance for each liquid type studied.
28
