Study on pickling technology to control fouling of ceramic membrane treating secondary treated effluent Open Access

In recent years, membrane treatment technology has been widely used in water treatment. Compared with other membrane materials, flat ceramic membranes have the advantages of high mechanical strength, good chemical stability, easy cleaning and regeneration, and long service life. It has a development potential that cannot be ignored in the field of sewage treatment (Li et al. 2020). At present, the market share of inorganic membrane in water treatment technology industry is about one eighth. In the future, the market sales of inorganic membrane will double every three years, which has broad application prospects (Gong et al. 2019). The biggest obstacle restricting the development of membrane separation technology is membrane fouling (Sun 2019; Gao et al. 2021; Chen & Liu 2022). When the membrane fouling reaches the limit, it needs to be cleaned to restore the flux and prolong the service life (Asif & Zhang 2021).

Chemical cleaning is the most direct and efficient cleaning method for flux recovery. Commonly used chemical cleaning agents include acid, alkali, oxidant, surfactant, etc. The effect of chemical cleaning is closely related to the type of agent, method and conditions. However, in the process of chemical cleaning, there are many problems such as low mass transfer rate of active components, large amount of cleaning agents, and the destruction of membrane structure by cleaning agents, especially oxidizing agents, which obviously weaken the performance of the cleaning process (Porcelli & Judd 2010). Therefore, many new membrane cleaning methods have been born in recent years, such as H₂O₂-MnO₂ system, ultrasonic cavitation, nano-microbubble enhanced cleaning, etc. However, they have not been widely spread (He et al. 2019; Li 2021). On the contrary, in practice, the chemical immersion method is widely used due to the advantages of simple process, so it is particularly important to further optimize the chemical cleaning mode to improve the membrane cleaning efficiency. At present, there are few researches on acid cleaning in chemical cleaning, especially the research on the cleaning effect of ceramic membranes under different cleaning methods of strong acid and weak acid is still lacking compared with organic membranes, and the cleaning mechanism is not clear. The problem, to a large extent, restricts the popularization and application of ceramic membrane water treatment technology (Zhang et al. 2020). In the previous study, the research group found that the acid cleaning reagent (HCl) had a better cleaning effect on the ceramic membrane, and the flux recovery rate could reach more than 80%, indicating that ceramic membrane had good pickling potential, so it was necessary to carry out systematic research on ceramic membrane pickling technology (Zhang et al. 2017; Zhong 2019).

In this study, the cleaning effects of three strong acids and three weak acids in single-step, mixed, step-by-step and synergistic with surfactants on ceramic membranes were investigated and compared with polyvinylidene difluoride (PVDF) membranes. For ceramic membrane under the cleaning methods of different combinations of cleaning agents, different cleaning effects were analyzed and the best combination of acid cleaning agents was selected. Characterization and analysis were carried out by means of scanning electron microscopy and contact angle. In order to find out the cleaning mechanism, develop efficient flat ceramic membrane cleaning technology and provide reference for the practical application of cleaning agents.

The experimental water used in this study was the effluent from the municipal secondary biochemical tank of the reclaimed water station of a university in Jinan. The sewage in the reclaimed water station mainly came from the domestic drainage of the staff and students in the school. The water quality was shown in Table 1.

The ceramic membrane used in the experiment was from Shandong Industrial Ceramics Research and Design Institute (China), and the PVDF membrane was from Memster Company (China), both of which were flat membranes, and their physical and chemical properties and their own water treatment performance were shown in Table 2. Among them, the pure water flux was the corresponding flux when the operating pressure was 22 KPa; the critical flux was the secondary effluent of the municipal sewage shown in Table 2 as the raw water before the ceramic membrane, which was measured by the flux ladder method.

The main chemical cleaning agents involved in the experiment were shown in Table 3.

Taking the municipal secondary effluent as the raw water before the membrane, the initial membrane flux was 50 L·(m²·h)⁻¹, the filtration time was 10 min, the intermittent running time was 1 min, the hydraulic recoil time was 20 s, the aeration intensity was 5 L·min⁻¹ (non-intermittent), and the constant temperature was 25 °C until the membrane fouling reached the limit. Then, the membrane whose fouling reached the limit was washed for 1 min under the water flow of 11 L/min to remove most of the reversible pollution on the membrane surface, and then backwashed for 20 s under the water flow of 1 L/min, followed by chemical cleaning. The strong acid cleaning agents used in this test were H₂SO₄, HCl and HNO₃, and the weak acid cleaning agents were oxalic acid, acetic acid and citric acid, and their pH was 3. And then the cleaning effect of ceramic membrane and PVDF membrane after being soaked in strong acid and weak acid for 2 h was investigated. The strong acid mixed cleaning combinations were HCl + H₂SO₄, HCl + HNO₃, H₂SO₄ + HNO₃; the weak acid mixed cleaning combinations were oxalic acid + acetic acid, oxalic acid + citric acid, acetic acid + citric acid; the respective cleaning concentrations were 1:1 and the soaking time was 2 h. Then, according to the two cleaning agents with better effect in the mixed cleaning of strong acid and weak acid of ceramic membrane and PVDF membrane, the different cleaning effects produced by different cleaning sequences during step-by-step cleaning were explored. The ceramic membrane and PVDF membrane were soaked in the first acid cleaning agent for 1 h. After removing the water in the membrane, the flux was measured, and then the membrane was soaked in another acid for 1 h and the flux was measured again. In the cleaning of ‘acid + surfactant’, the acid with better effect in single acid cleaning was used, and three representative types of surfactants were selected: cationic surfactant – dodecyl trimethyl ammonium chloride (DTAC), anionic surfactant – sodium dodecyl benzene sulfonate (SDBS) and non-ionic surfactant- nonylphenol polyoxyethylene ether (OP-10). The three surfactants were used four times of their respective critical micelle concentration (CMC).

Scanning electron microscope (SEM) was a technique that used secondary electron signal imaging to observe the surface morphology of samples. In this study, QUANTA FEG250 scanning electron microscope from FEI company in the United States was used to measure the surface topography of the membrane at a magnification of 2,000 times. The membrane samples were dried naturally and sprayed with gold, and then placed under SEM to observe the surface morphology.

Contact angle (CA) referred to the tangent of the gas-liquid interface made at the intersection of the gas, liquid and solid three phases. The angle θ between the tangent on the liquid side and the solid-liquid boundary was a measure of moisture level. The contact angle could reflect the hydrophilicity of the membrane surface. The smaller the contact angle, the better the hydrophilicity of the membrane. On the contrary, the larger the contact angle, the worse the hydrophilicity of the membrane.

In this paper, DSA30 contact angle tester from KRUSS company in Germany was used to measure the CA of film and pollutants by the lying drop method. During the measurement, 4 μL of deionized water was used and gently dropped on the membrane surface. The contact angle measurement was performed by taking a stable image capturing a water droplet on the film surface. Each sample was measured 10 times, the maximum and minimum values were discarded and the average value was calculated as the study.

It could be seen from Figure 2(a) that for flat ceramic membranes, H₂SO₄ had the highest flux increment (14.69%) relative to physical cleaning, followed by HCl (13.17%). For PVDF membranes, HCl soaked the membrane with the highest flux increment (13.83%) relative to physical cleaning, followed by H₂SO₄ (11.75%). The test showed that both HCl and H₂SO₄ had good cleaning effect. This might be because HCl and H₂SO₄ could dissolve insoluble matter or composite pollutants in the membrane fouling layer, which destroyed the bond between the filter cake structure and the membrane surface (Tian et al. 2013). However, after the flat ceramic membrane and PVDF membrane were soaked in HNO₃, the value-added was very small relative to physical cleaning, and there was almost no cleaning effect. This might be because some pollutants were decomposed by it, forming more fine dirt during the cleaning process, resulting in less effective cleaning (Ochando-Pulido et al. 2015).

It could be seen from Figure 2(b) that for flat ceramic membranes, oxalic acid immersion had the highest increase in membrane flux (15.75%) relative to physical cleaning, and for PVDF membranes, citric acid immersion had the highest increase in membrane flux relative to physical cleaning (25.56%), which was about 10% higher than oxalic acid (15.75%). Experiments had shown that oxalic acid had a significant cleaning effect on both flat ceramic membranes and PVDF membranes. This might be because the membrane fouling layer was mainly composed of inorganic salts and sludge flocs, and oxalate ions could complex insoluble salt metal ions, reducing the degree of binding between the fouled membrane surface and pollutants. After the flat ceramic membrane and PVDF membrane were soaked and cleaned in acetic acid, the membrane flux hardly changed. Furthermore, citric acid had little effect on flux of flat ceramic membranes, but had a very significant cleaning effect on PVDF membranes. This might be because citric acid could hydrolyze some organics and was very effective for cleaning carbonate deposits on contaminated PVDF membrane surfaces, as it not only dissolved and removed them, but also formed complexes that were easier to remove from the membrane surface (Goon et al. 2021).

Figure 2(c) showed the flux change when the ceramic membrane cleaned with H₂SO₄ and the PVDF membrane cleaned with HCl were used to treat municipal secondary effluent. It could be seen that the flux of ceramic membrane and PVDF membrane after acid washing were 51.2 LMH and 49.8 LMH, respectively, which were basically consistent with the initial membrane flux. The reactor ran for 1 h under the above working conditions, and the flux of ceramic membrane and PVDF membrane decreased rapidly. This was because the particles in the raw water whose particle size was much smaller than the pore size of the membrane entered the membrane through the membrane surface, and the particles in the raw water with a particle size larger than or slightly equal to the membrane pore size were adsorbed on the membrane surface. When the reactor ran for 30 h the flux of ceramic membrane and PVDF membrane still decreased with the increase of time, but the flux reduction rate of ceramic membrane was lower than PVDF membrane. Therefore, the flux of ceramic membrane was greater than that of PVDF membrane, and this phenomenon still remained with the increase of time. After the reactor was operated for 100 h, the flux remained almost unchanged with the increase of time. This was because the membrane pores were almost filled with small-sized pollutant particles, and the pollutants in the raw water were adsorbed and deposited on the membrane surface. Under the action of pressure, the pollution layer was continuously squeezed to form a dense filter cake layer. At this time, the flux tended to stabilize.

It could be clearly seen from Figure 2(c) that at 100 h, the flux of ceramic membrane and PVDF membrane after pickling basically tended to be stable. At this time, the flux of ceramic membrane was 13 LMH and that of PVDF membrane was 10 LMH. The test results showed that when the flux of ceramic membrane after H₂SO₄ cleaning and PVDF membrane after HCl cleaning decreased to the same level, the running time of ceramic membrane was longer than that of PVDF membrane.

When the same concentration of HCl and H₂SO₄ were mixed to a pH of 3, the concentration of Cl⁻ in the cleaning solution was more than ⁠, and Cl⁻ had a significant impact on the oxidation of organic pollutants, and reacted with to generate free radicals ⁠, which pushed the reaction (1) to the right and further depletes Cl⁻, and the effective content of Cl⁻ in the solution decreased. Subsequently, and Cl⁻ can be further transformed into through Equation (3), and then into free chlorine (Cl₂ and HOCl). The efficiency of cleaning also decreased.

The increase of flux of ceramic membrane after mixed cleaning with HCl and HNO₃ was 7.72% lower than that of single-step cleaning with HCl, and 4.42% higher than that of single-step cleaning with HNO₃. The increase of flux of PVDF membrane was 13.03% lower than that of single-step cleaning with HCl, and 3.35% lower than that of single-step cleaning with HNO₃. The increase of the flux of ceramic membrane after mixed cleaning with H₂SO₄ and HNO₃ was 5.75% lower than that of the single-step cleaning with H₂SO₄, and 4.52% higher than that of the single-step cleaning with HNO₃. The increase of flux of PVDF membrane was 0.65% lower than that of single-step cleaning with H₂SO₄, and 9.03% higher than that of single-step cleaning with HNO₃. This might be due to the presence of ⁠, ⁠, plasma in the mixed cleaning solution, and would undergo redox reaction with ⁠, thereby weakening the concentration of the cleaning agent (Katsoufidou et al. 2010).

The mixed cleaning results of different weak acid combinations shown in Figure 3(b) showed that the mixed cleaning effect of weak acid combination was not necessarily higher than that of single-step cleaning, which was consistent with the conclusion reached when exploring the mixed cleaning effect of strong acid combination. After the mixed cleaning of ceramic membrane and PVDF membrane with oxalic acid + acetic acid, the increase of flux was lower than that of single-step cleaning with oxalic acid and higher than that of single-step cleaning with acetic acid. The flux increment of oxalic acid + citric acid, acetic acid + citric acid mixed cleaning of ceramic membrane was improved compared with that of single-step cleaning, which might be related to the complex reaction of citric acid. The inorganic salt and organic-inorganic complex in the polluted layer would compete with citric acid reaction (Hilal et al. 2005; Beattie et al. 2014; Ferrer et al. 2016). When only citric acid was used, due to the competition of pollutants, it would react with citric acid, leading to the decrease of citric acid concentration, and then the cleaning effect decreased. When mixed with acetic acid or oxalic acid, these two weak acids could react with inorganic salt to remove inorganic pollutants. Therefore, the effect of these two mixed cleaning methods was more obvious than that of a single agent. For PVDF membrane, the flux increment of mixed and single-step reagent cleaning decreased slightly, which might be related to the inorganic content of the two membrane pollutants.

According to the cleaning agents with better effect of strong acid and weak acid in mixed cleaning for ceramic membrane and PVDF membrane, the effects produced by different cleaning sequences during step-by-step cleaning were explored.

As could be seen from Figure 4(a), the cleaning effects of the two cleaning methods of HCl first and then H₂SO₄ and H₂SO₄ first and then HCl were not significantly different, and both were lower than the effect produced by the H₂SO₄ + HCl. After cleaning ceramic membrane with HCl first and then H₂SO₄, the flux recovery rate was 61.01%, which was 10.52% higher than that of physical cleaning. After cleaning ceramic membrane with H₂SO₄ first and then HCl, the flux recovery rate was 60.98%, which was 8.63% higher than that of physical cleaning. After ceramic membrane was cleaned with HCl + H₂SO₄, the flux recovery rate was 69.05%, which was 11.79% higher than that of physical cleaning. It could be seen from the results that the cleaning effect of each agent soaked for 1 h was not much different than that of soaking for 2 h. Therefore, it was inferred that the insignificant cleaning effect was related to the reaction time between each agent and pollutants. In addition, the mixing of Cl⁻, and was not the best cleaning method.

The effect of the two cleaning sequences for PVDF membrane was more than 90%. However, comparing the flux increment of two cleaning sequences relative to physical cleaning, it was obvious that HNO₃ first and then H₂SO₄ (27.24%) was higher than H₂SO₄ first and then HNO₃ (17.31%). Compared with the physical cleaning, the flux increment of PVDF membrane with H₂SO₄ + HNO₃ was 13.18%, which was not much different from the cleaning effect of H₂SO₄ first and then HNO₃.

It could be seen from Figure 4(b) that the cleaning sequence of oxalic acid and citric acid had no obvious effect on ceramic membrane. The flux recovery rate of ceramic membrane after cleaning was below 50%. The two cleaning methods increased the flux by about 5% relative to the physical cleaning. However, the flux increment of oxalic acid + citric acid mixed cleaning ceramic membrane relative to physical cleaning was 18.2%. Citric acid as a complexing agent and oxalic acid as a cleaning agent for inorganic scale, the mixing of the two would have obvious synergistic effect, and the effect of mixed cleaning was better than step-by-step cleaning.

The effects of the two cleaning sequences for PVDF membranes were both above 90%. Comparing the flux increment of the two cleaning methods relative to physical cleaning, oxalic acid first and then citric acid (8.57%) was almost the same as citric acid first and then oxalic acid (8.59%), and the cleaning sequence of oxalic acid and citric acid had no effect on the cleaning effect of PVDF membrane. The effect of oxalic acid + citric acid mixed cleaning for PVDF membrane was 88.46%, and the flux increment relative to physical cleaning was 12.78%. The phenomenon that the step-by-step cleaning effect was lower than the mixed cleaning might be related to the synergistic effect of citric acid and oxalic acid.

According to the single-step cleaning results, the most significant cleaning effects of strong acid single-step cleaning on ceramic membrane and PVDF membrane were H₂SO₄ and HCl, respectively. Therefore, only the effects of mixed cleaning of H₂SO₄ and surfactant on ceramic membrane and mixed cleaning of HCl and surfactant on PVDF membrane were considered. The cleaning effects were shown in Figure 5(a).

Figure 5(a) showed that after H₂SO₄ + DTAC, H₂SO₄ + OP-10 and H₂SO₄ + SDBS mixed cleaning, the flux increment relative to physical cleaning was 34.28%, 18.67% and 3.42% respectively. H₂SO₄ + DTAC mixed cleaning was the most effective for ceramic membrane. The reason might be that the nitrogen atom in DTAC molecule contained lone pair electrons, which could combine with hydrogen in acid molecule by hydrogen bond to make the amino group positively charged, so H₂SO₄ + DTAC had good cleaning effect (Geng 2012; Zhu et al. 2015; Zhong 2019).

After the PVDF membrane was mixed cleaning with HCl + DTAC, HCl + OP-10, and HCl + SDBS, the flux increment relative to physical cleaning was 23.34%, 13.53%, and 5.5%, respectively. HCl + DTAC mixed cleaning was the most effective for PVDF membrane.

According to the results of the above research, H₂SO₄ + DTAC and HCl + DTAC had significant cleaning effects on ceramic membrane and PVDF membrane. Next, it was necessary to explore the effects of step-by-step cleaning of these two cleaning methods, as shown in Figure 5(b).

It could be clearly seen from Figure 5(b) that the cleaning method of H₂SO₄ first and then DTAC was the most effective, followed by H₂SO₄ first and then OP-10, H₂SO₄ first and then SDBS. The order of the cleaning effect was consistent with the conclusion drawn during mixed cleaning. In addition, after ceramic membrane was cleaned by H₂SO₄ first and then DTAC, the flux recovery rate reached 97%, which was 45% higher than that of physical cleaning and higher than the flux increment of H₂SO₄ + DTAC cleaning relative to physical cleaning (34.28%). The results showed that in the first step of H₂SO₄ cleaning, the cleaning effect of H₂SO₄ on the three membranes was almost the same. In the second step of three kinds of surfactant cleaning, the cleaning effect of DTAC was the best, which was consistent with the above-mentioned sulfuric acid + DTAC as the best cleaning effect.

After the PVDF membrane was cleaned by HCl first and then SDBS, HCl first and then DTAC, and HCl first and then OP-10; the flux increment relative to physical cleaning was 15.11%, 13.2%, and 7.04% respectively. The two step-by-step cleaning methods of HCl first and then SDBS and HCl first and then DTAC had similar effects on PVDF membrane.

From the above test results, it was clear that the effect of DTAC mixed with strong acid and step-by-step cleaning was better than the other two surfactants. Next, the effect of DTAC mixed with weak acid with remarkable single-step cleaning effect and step-by-step cleaning on ceramic membrane and PVDF membrane was explored, as shown in Figure 5(c).

It could be seen from Figure 5(c) that the flux recovery rate of ceramic membrane after oxalic acid first and then DTAC and oxalic acid + DTAC cleaning was 41.56% and 39.47% respectively, and the flux increment relative to physical cleaning was 9.95% and 6.54%, respectively. It showed that in the step-by-step cleaning, the recovery rate of the first-step oxalic acid cleaning and the second-step DTAC cleaning was lower than two agents single-step cleaning. The reason for this phenomenon might be that the soaking time of each agent in step-by-step cleaning was lower than that in single-step cleaning.

The flux recovery rate of PVDF membrane cleaned by citric acid first and then DTAC and citric acid + DTAC were 48.57% and 51.82%, respectively, and the flux increment relative to physical cleaning was 10.13% and 6.91%, respectively. The results showed that the effect of PVDF membrane cleaned by citric acid first and then DTAC was higher than that of citric acid + DTAC. The results showed that in the step-by-step cleaning, the recovery rate of the first step citric acid cleaning and the second step DTAC cleaning was lower than that of the single-step cleaning of the two cleaning agents. The reason for this phenomenon might be that the soaking time of each agent in the step-by-step cleaning was lower than that of the single-step cleaning.

In order to observe the hydrophilicity and hydrophobicity of the ceramic membrane, its contact angle was characterized, as shown in Table 4.

It could be seen from Table 4 that the contact angle of the new membrane was 48.49°, which had good hydrophilicity. The contact angle of the membrane whose pollution reached the limit was 88.24°. The contact angle of the polluted membrane after physical cleaning decreased from 88.24° to 68.90°, and the hydrophilic was worse than that of the new membrane. After cleaning with H₂SO₄ first and then DTAC, the surface contact angle of the fouling membrane was smaller than that of the new membrane, and the hydrophilic was better than that of the new membrane. The reason for this phenomenon was the same as the corresponding explanation for the SEM image above, which also explained the phenomenon that the flux of the ceramic membrane cleaned by this cleaning method exceeded that of the new membrane. The efficiency of the optimal cleaning scheme recommended in this study was further verified.

This study focused on the cleaning effects of three strong acids, three weak acids and surfactants on ceramic membrane, and compared it with PVDF membrane. In single-step cleaning, H₂SO₄ and HCl had the best cleaning effect on ceramic membrane and PVDF membrane, respectively. Moreover, the municipal secondary effluent was treated with the ceramic membrane cleaned by H₂SO₄ and the PVDF membrane cleaned by HCl. When the fluxes of the two membranes dropped to the same level, the operating time of the ceramic membrane was longer than that of the PVDF membrane. In agent combination cleaning, there may be obvious inhibition between the cleaning agents, so that the cleaning effect may not exceed that of single-step cleaning, and the cleaning sequence of different acids had little effect on the cleaning effect. According to the cleaning methods in this experiment, the most obvious cleaning effect on ceramic membrane was H₂SO₄ first and then DTAC. The flux recovery rate after cleaning reached 96.94%, basically reaching the level of new membrane.

This work was financially supported by the Key R&D Program of Shandong Province (2016CYJS07A03-2).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.