- Email: [email protected]
Membrane treatment of biodiesel wash-water: A sustainable solution for water recycling in biodiesel production process
Journal of Water Process Engineering 19 (2017) 331–337
Contents lists available at ScienceDirect
Journal of Water Process Engineering journal homepage: www.elsevier.com/locate/jwpe
Membrane treatment of biodiesel wash-water: A sustainable solution for water recycling in biodiesel production process
MARK
⁎
Kosar Mozaffarikhaha, Ali Kargarib, , Meisam Tabatabaeic,d, Hossein Ghanavatic,d, Mohammad Mahdi A. Shirazie a
Department of Chemical Engineering, Mahshahr Branch, Islamic Azad University, Mahshahr, Iran Membrane Processes Research Laboratory (MPRL), Department of Chemical Engineering, Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran c Microbial Biotechnology Department, Agricultural Biotechnology Research Institute of Iran (ABRII), Agricultural Research, Education and Extension Organization (AREEO), Karaj, Iran d Biofuel Research Team (BRTeam), Karaj, Iran e Membrane Industry Development Institute, Tehran, Iran b
A R T I C L E I N F O
A B S T R A C T
Keywords: Biodiesel Wash-water Nanofiltration Wastewater treatment Environment Sustainable solution
Most biodiesel production processes lead to the generation of large volumes of highly polluted wastewater, namely wash-water effluent. The principal aim of this paper is to show the applicability and effectiveness of the nanofiltration process for treatment of biodiesel wash water effluent and possibility of recycling the treated wastewater. For this purpose, three types of commercial nanofiltration membranes (TW30, NE90 and NE70) have been examined. The effect of two important operating parameters, including operating pressure (6–14 bar) and feed flow-rate (100–250 L/h) on the permeate flux, flux decline, and also separation performances for biological oxygen demand (BOD), chemical oxygen demand (COD), total solids (TS) and total dissolved solids (TDS) were investigated for the TW30 membrane. The obtained results showed that the permeate stream obtained from a feed at a pressure of 12 bar and a flow rate of 250 L/h, through a single stage nanofiltration process have BOD, COD, TDS and TS of 85.6%, 85.8%, 97.3% and 98.8% lower than the initial wash water effluent, respectively. The results of testing the three membranes at 14 bar and 250 L/h showed that TW30 had the lowest fouling tendency and flux decline followed by NE90 and NE70, respectively. The highest rejection for BOD, COD, TDS and TS was also observed for the TW30 membrane which shows this membrane is a good candidate, among the studied membranes, for further consideration and commercial applications.
1. Introduction Fossil fuels are non-renewable sources of energy and their growing utilization has resulted in global warming and climate change. These in turn have endangered mankind’s long-term survival [1]. Biofuels as alternatives to fossil-based conventional fuels have been considered as a get-away strategy that could partially address some of the environmental concerns [2]. Among various biofuels, the Biodiesel could be produced from a variety of vegetable oils (edible and non-edible), animal fats, and waste cooking oil and has been used as an alternative to petroleum diesel [3]. Among the many advantages of biodiesel is that it can be used in diesel engines with no or minor modifications. Biodiesel is mostly produced through the transesterification reaction of triglycerides with a short chain alcohol (usually methanol) in the presence of a suitable catalyst such as sodium or potassium hydroxide [4]. Most biodiesel production processes involve a downstream process
⁎
namely the wet washing step in which hot water (50–60 °C) is used to strip the crude biodiesel of its impurities. This process leads to the production of a very polluting wastewater (i.e. the COD up to 35000 and the BOD up to 30000 mg O2/L) [5] also called biodiesel washwater. More specifically, based on our previous experiments, about 3–10 L of biodiesel wash-water is generated during this process for each liter of biodiesel [6]. It is also indicated in the literature that in 2011, worldwide generation of wash-water was approximately 28 million m3 per year [7]. This highly organic effluent consists of water, residual biodiesel, residual catalyst, soap, salts, methanol and traces of un-reacted oil and glycerol [8]. As a result, this wastewater should not be discharged of without a proper treatment step as it poses a serious threat to the environment. There are strict industrial effluent standard limits in this regard. For instance, the BOD, COD, and the oil and grease (O & G) content limitations in Malaysia are 20, 80 and 1.0 mg/L, respectively [9]. It is worth mentioning that since biodiesel wash-water is
Corresponding author. E-mail addresses: [email protected], [email protected] (A. Kargari).
http://dx.doi.org/10.1016/j.jwpe.2017.09.007 Received 8 April 2017; Received in revised form 23 August 2017; Accepted 6 September 2017 2214-7144/ © 2017 Elsevier Ltd. All rights reserved.
Journal of Water Process Engineering 19 (2017) 331–337
K. Mozaffarikhah et al.
Table 1 Pros. and Cons. of some biodiesel wash-water treatment processes [9]. Process Coagulation
Biological treatment
Adsorption
Microbial fuel cell
Pros.
Cons.
• Simple • Economical treatment technique • Proven • Economical arrangement for small areas • Versatile and suitable for small plants • Simple additional sludge is produced • Non of discharged wastewater is unaffected • pH COD removal capacity • High • High COD removal capacity
handling requirement • Chemical operation • Complicate low density sludge with low decomposition efficiency • Generates large amount of low density sludge • Generate consuming • Time to manage optimum condition first • Need further treatment • Need • Difficulties in off adsorbents discharging • Costly
reported and an economically-viable process which could be implemented on an industrial scale. Based on the above-mentioned studies, it could be concluded that the MF process is not very effective for treatment of biodiesel washwater because most of the pollutants present in wash-water are low molecular weight compounds that could not be rejected by MF membranes. Then, more tightened membranes such as ultrafiltration (UF), nanofiltration (NF) or reverse osmosis (RO) seem to be more effective. On such a basis, the present study was set to apply a commercially available NF membrane to treat biodiesel wash-water. Moreover, the impact of operating parameters such as feed pressure and flow rate on the chemical oxygen demand (COD), biological oxygen demand (BOD) and total dissolved solid (TDS) reduction of biodiesel wash-water by the developed system were thoroughly investigated.
a highly stable emulsion containing grease, oil and soap, its treatment by using conventional techniques seems difficult. In other words, water washing has been proven to result in the biodiesel meeting the international standard specifications laid out for biodiesel [10]. However, it gives rise to some disadvantages, among which high water consumption, despite the current water shortage and considerable increase in fresh water demand worldwide [11], is of major concern. As a result, purification of this environmentally hazardous wastewater via hightech strategies should be taken into serious consideration. To date, several treatment processes have been developed for treating the biodiesel wash-water such as physic-chemical treatment, electro-chemical treatments, coupled chemical and electrochemical treatments, advanced oxidation technologies, biological treatments and, integrated treatment processes [7,9]. The pros and cons for some of these processes are listed in Table 1. Membrane separation processes are clean and energy effective processes [12], in which fluid separation takes place without the contribution of a third phase and is usually without phase change. Among those, the most commonly known process is water desalination, which has been traditionally performed by highly energy intensive distillation processes like multi-effect distillation (MED) and multi-stage flash distillation (MSF) [13]. The use of membranes as the separation medium has revolutionarily converted this process as well as many other processes into cleaner, more environmentally-friendly, less expensive, less energy-intense, and more compact processes with less foot-prints and without any by products or residual wastes generated [14]. Accordingly, it is believed that the application of membrane separation processes in biodiesel processing [15], in particular its wastewater treatment, could also be very effective. However, only few works have been performed on biodiesel and its wastewaters by membrane separation processes [7,9]. Shirazi et al. [6] used a microfiltration (MF) process by a superhydrophobic electrospun nano-fibrous polystyrene membrane to treat the biodiesel's water-washing effluent and reportedly managed to decrease COD and BOD by 75% and 55%, respectively. They used a contact heating method to modify the membranes’ surface, and then characterized them using atomic force microscopy (AFM) for their surface features. In another work, Jaber and his co-workers [16] used a simple microfiltration-based procedure to treat and subsequently re-use biodiesel wash-water. They applied commercial polypropylene (PP) MF membranes with 1 and 5 μm pore sizes and also commercial activate carbon and sand filter cartridges to perform the treatment in a hybrid system. The highest COD and BOD reductions obtained in their work stood at around 18.5% and 12.4%, respectively, which are lower compared to the work of Shirazi et al. [6]. Moreover, they argued that their innovative PP-based microfiltration followed by sand and activated carbon separations and in combination with 70% dilution rate with fresh water enabled the re-use of biodiesel wash-water which still led to a standard-quality biodiesel product and resulted in up to 15% less water consumption. Despite all the efforts put into treating this stringent wastewater, there is still a wide gap between the findings
2. Experimental 2.1. Materials Biodiesel wash-water was provided by the Biofuel Research Team (BRT, Karaj, Iran). The biodiesel itself was produced from waste cooking oil and methanol (1:6 molar ratio) through the alkali-catalyzed transesterification reaction (potassium hydroxide 1 wt.%). The specification of the crude wash-water is shown in Table 2. A commercially available NF membrane, (TW30-1812-50) from Dow-Filmtec Co. [17] was used as the separation media. The membrane had a thin-film composite structure composed of a very thin and selective polyamide layer at the top, a polysulfone UF structure as the mid-layer and a non-woven polyester fabric as the support layer. The specification of the membrane module, which has been reported by the manufacturer, is tabulated in Table 3. The used membrane was cut from the membrane module. 2.2. Methods 2.2.1. Experimental setup Fig. 1 represents the schematic of the experimental setup. A 25 L Table 2 The average specification of the crude wash-water. Property
Value
Unit
EC pH COD BOD TDS TSS TS
1086 6.98 19880 18335 1238 1370 2608
μS/cm – mg/L mg/L mg/L mg/L mg/L
EC: Electrical conductivity. TSS: Total suspended solids. TS: Total solids.
332
Journal of Water Process Engineering 19 (2017) 331–337
K. Mozaffarikhah et al.
controlled at 30 °C. For each experiment, the membrane was soaked in distilled water for at least 48 h. The feed was pumped into the membrane module at predetermined pressure and flow rate. More specifically, the effect of different feed flow rates (i.e. 100, 150, 200, and 250 L/h) at constant pressure and the effect of inlet pressure (i.e. 6, 8, 10, 12 and 14 bar) at constant feed flow rate were investigated. After measuring the permeate stream flow rate, it was recycled (together with the retentate stream) back into the feed tank. Sampling from the permeate stream was done when the steady-state condition was established. The time required for attaining the steady-state was found different according to the circumstances of each experiment, but an average of 7 h of operation (i.e. recirculation) was sufficient for all the experiments to reach the steady-state condition. Permeate flux and solute rejections were calculated using Eq. (1) and (2):
Table 3 The specification of the NF membrane module used in this study [17]. Specification
Value
Unit
Area Diameter Length Max. working pressure Max flow rate Max. feed water temperature pH range
4.8 1.75 10 300 7.6 45 2–11
ft2 in in psi L/min °C –
stainless steel tank was used as the feed reserve. A low-pressure pump (So-Pure, Korea) and a 5 μm pore size propylene depth pre-filter were used as the pre-filtration system before the feed enters the main highpressure pump. A vertical multistage centrifugal stainless steel pump (LVB 1–36, Leo Co., China) with maximum pressure and flow of 25 bar and 1.8 m3/h was used as the main pump in the feed stream, respectively. The adjustment of the pump outlet pressure was performed by an AC frequency control system (N700E, LG Co., Korea). Then, the feed stream entered the membrane module, which was a handmade stainless steel, cross-flow filtration system containing a flat sheet membrane. The active area of the flat membrane was 78 cm2 (W = 5.2 and L = 15 cm). Finally, the retentate stream passed through a rotameter (25–250 L/h, Fischer & Porter 3400, Gottingen, Germany)) and returned into the feed tank. The final adjustment of the feed stream pressure and flowrate was performed by two globe valves before and after the membrane module. As shown in the schematic diagram, two pressure gauges (Bourdon Sedeme, 0–16 bar and Sangan Sanat, 0–40 bar) were placed before and after the membrane module to monitor the operating pressure. Temperature adjustment of the feed was performed by a cold water stainless steel coil, which was immersed into the feed reserve. Electrical conductivity (EC) was measured by a Metrohm conductivity meter (Metrohm 720, Switzerland) and used for determination of TDS. A digital pH meter (Metrohm-780 Herisau, Switzerland) equipped with a Pt-1000 thermometer for temperature correction was used for measuring the pH of the aqueous solutions.
Jp =
Qp A
Cp ⎞ R (%) = ⎜⎛1 − ⎟ × 100 Cf ⎠ ⎝
(1)
(2)
Where Qp is the permeate stream flowrate (L/h),A is the effective area of the used membrane (m2), R (%) is the percentage of rejection for a specific solute, Cp and Cfare the permeate stream and feed concentrations, respectively [18,19].
2.2.3. Analytical methods The Open-reflux and Azide methods were used for COD and BOD5 measurements, respectively, according to the standard method for examination of water and wastewater. The pH and electrical conductivity (EC) was measured in the samples with an electrometric pH meter and EC meter, respectively; and TDS (total dissolved solids), TSS (total suspended solids), and TS (total solids) were detected by the method described in the previous work [6]. The applied membrane, before and after NF experiments, was characterized using SEM (scanning electron microscopy) for its morphological features and fouling characteristics, as shown in Fig. 2.
2.2.2. Procedure In all the experiments conducted, the feed temperature was
Fig. 1. A general scheme of the experimental apparatus.
333
Journal of Water Process Engineering 19 (2017) 331–337
K. Mozaffarikhah et al.
Fig. 2. SEM images of applied commercial NF membrane in this work, before (left) and after (right) wash-water treatment experiment.
thinner polarized layer at higher feed flows are at times significantly less than the thicker one for lower feed flow rates [23]. As a consequence, the time for attaining steady-state flux would be shorter as the feed flow rate increases. Having considered the above-mentioned results, it could also be stated that the feed flow rate is effective on wash-water characteristics, e.g. TDS, TS, BOD and COD. The effect of variations in feed flow (100, 150, 200 and 250 L/h) on the rejection of these items at 14 bar is shown in Fig. 4. The results showed that the rejection percentage (%) for TS and TDS are almost identical by using the proposed NF membrane at 14 bar operating pressure. High rejection values (i.e. > 98%) for both TDS and TS were obtained, which is a considerable result compared with conventional treatment processes (see Table 2 in Ref. [7]). However, the effect of feed flow rate on the variation of TDS and TS rejections was negligible. Fig. 4 also shows the effect of feed flow rate on BOD and COD rejection at the mentioned inlet pressure (i.e. 14 bar). As could be observed, by increasing the flow rate, the rejection rate was increased as well, but similar BOD and COD rejection behavior is observed. The BOD and COD rejections under the proposed condition were at the range of 67–81%. It could be claimed that this is an appropriate result for a single-stage membrane process. The obtained results mentioned above could be concluded by this fact that an increase in the concentration polarization layer has two drawbacks on the membrane performance. On the one hand, by increasing the polarization layer, the hydraulic resistance of the
3. Results and discussion 3.1. Effect of feed flow rate Feed flow rate is one of the factors that can affect membrane performance. To test the effect of this operating parameter, experiments were conducted with biodiesel wash-water under different feed flow rates ranging from 100 to 250 L/h at 14 bar constant inlet pressure. As could be observed in Fig. 3, the permeate fluxes decreased during the first 25 min of the experiment for all inlet flow rates. This could be due to the gradual compaction of the membrane [20] and development of the concentration polarization layer [21] at different feed flow rates. This declining trend of the flux proceeded until it reached a slight steady state and remained constant over time for the upper range of flow rates, i.e. 200 and 250 L/h. While for the lower bond of flow rate, i.e. 100 L/h, the permeate flux became constant after about 375 min. Moreover, the experimental results showed that by increasing the feed flow rate from 100 to 250 L/h (under constant operating pressure), the membrane permeate flux was increased. For instance, the initial flux for the lowest and highest flow rates are about 47 L/h and 54 L/h, and then reduced to 33 L/h and 48 L/h for the same issues, respectively. It is worth quoting that the time to reach the steady-state decreased by increasing the flow rate value. This fact could be ascribed to the concentration polarization effect. More specifically, by increasing the feed flow rate, the thickness of the concentration polarization layer is decreased resulting in less hydraulic resistance and higher flux through the membrane [22]. It is claimed that the time for formation of a
Fig. 3. Effect of feed flow rate on the membrane permeate flux at 14 bar and 30 °C.
334
Journal of Water Process Engineering 19 (2017) 331–337
K. Mozaffarikhah et al.
Fig. 4. Effect of feed flow rate on the rejection of TDS, TS, BOD and COD at 14 bar and 30 °C.
Fig. 6. Effect of feed pressure on the steady state flux at feed flow rate of 250 L/h and 30°C.
membrane increases which results in lower permeation rates. On the other hand, the real concentration of the rejected species on the membrane surface increases which results in higher concentration of the permeated stream (i.e. lower rejection rates). Therefore, membrane separations at higher feed flow rates and higher operating pressures would be more favorable, specifically for samples like biodiesel washwater. The obtained results are also in agreement with the known Kedem–Katchalsky theory in which higher permeation rates would result in higher rejections [24]. Hence, the 250 L/h feed flow rate was investigated for the following experiments.
As all the experiments were carried out at the highest feed flow rate, based on the conclusion in the previous section, the initial flux decline for each case was not considerable especially at low pressures (see Fig. 5). In other words, in addition to the membrane compaction, the rate of the polarized layer formation and cake compaction are lower and only a small flux decline is observed. By increasing the feed pressure over 8 bar, further membrane compaction, cake formation and polarization, and as a consequence considerable flux decline, are expected [25]. It is worth quoting that under such feed pressures, the rate of permeation fluxes were high which led to reaching pseudo-steady conditions within a few seconds that could not be measured manually, meaning that the initial sharp flux decline could not be observed at higher operating pressures ( > 8 bar). By increasing the feed pressure, the time for establishing the real steady-state condition is also increased. This time at 6 bar is about 100 min while this time for 14 bar is measured at about 360 min. The influence of feed pressure on the steady state permeation flux is shown in Fig. 6. As could be observed, there is a linear relationship between the steady permeate flux and applied pressure. A little deviation is observed at pressures of 6 and 8 bar. This is due to the
3.2. Effect of feed pressure The effect of operating pressure (i.e., 6, 8, 10, 12 and 14 bar) on the membrane permeate flux at constant 250 L/h feed flow rate and 30 °C is shown in Fig. 5. Under different operating pressures, membrane permeates decreased over the operation time for individual inlet pressures. This is due to the gradual formation of concentration polarization and the cake layer on the membrane surface during almost the first 150 min of operation. This is while steady state permeate fluxes were reached after this time.
Fig. 5. Effect of operating pressure on the membrane permeate flux at 250 L/h feed flow rate and 30°C.
335
Journal of Water Process Engineering 19 (2017) 331–337
K. Mozaffarikhah et al.
Table 4 Separation performance of the three studied membranes for treatment of biodiesel wash water (test condition: feed pressure: 14 bar, feed flow rate: 250 L/h). Membrane Type
Rejection [%]
TW30 NE90 NE70
COD
BOD
TDS
TS
81.31 77.94 80.88
80.56 74.13 79.17
98.72 86.56 45.55
98.60 86.19 29.04
available membranes, i.e. NE70 and NE90 both from CSM, Korea. These membranes are used widely for brackish water desalination, especially where high fluxes are expected. The results of testing these membranes are shown in Fig. 8. The test condition was selected as 14 bar feed pressure and 250 L/h feed flow rate. The results reveal that the order of initial flux for these three membranes is NE70 > NE90 > TW30. In the case of NE70, a sharp flux decline was observed that shows severe fouling and a certain extent of compaction of the membrane in the first 200 min of operation. NE90 showed lower permeate flux than NE70 with lower flux decline. After about 600 min of operation, the flux of NE90 and TW30 membranes were nearly the same and both higher than NE70. The interesting point in these experiments was that TW30 showed a nearly flat flux with time. In other words, the permeate flux of the TW30 membrane declined very slowly compared to the other two membranes. This reveals that the TW30 membrane is an anti-fouling membrane and is more suitable than the other two membranes for treatment of biodiesel wash water. The overall flux decline of the three membranes within 600 min was calculated as well. The lowest flux decline belongs to the TW30
Fig. 7. Effect of operating pressure on the rejection of TDS, TS, BOD and COD (feed flow rate of 250 L/h feed flow rate and temperature of 30°C).
polarization effect and initial flux decline. However, as the feed pressure is increased, the steady state flux becomes more linear. In other words, according to the well-known Darcy's law of filtration, the effect of inlet pressure on the hydraulic resistance against membrane permeation is negligible so that the permeation flux becomes a linear function of applied pressure [26,27]. The effect of various inlet pressures on the rejection of TDS, TS, BOD and COD at constant 250 L/h feed flow rate is shown in Fig. 7. Approximately the same rejection behavior for TDS and TS, and also for COD and BOD could be observed, respectively. By increasing the feed pressure, the rejection value (%) for all species is increased. Considering the effect of feed pressure on the permeation flux (see Fig. 6), the permeation flux increases as the feed pressure is increased, which is in agreement with the Kedem–Katchalsky theory for rejection, as mentioned earlier. The maximum rejection values for BOD, COD, TDS and TS at 14 bar are measured as 80.7, 81.3, 98.7 and 98.6 percent, respectively.
L
L
⎛ ⎞⎞ ⎛ ⎞⎞ ⎛ ⎛ 2 2 membrane ⎜0.11 ⎝ m h. h ⎠ ⎟, then to NE90 ⎜0.26 ⎝ m h. h ⎠ ⎟ and finally to NE70 ⎝ ⎠ ⎝ ⎠ ⎛ L ⎞⎞ ⎛ 2 ⎝m .h⎠ ⎜0.61 h ⎟. These results show that the best membrane for this ap-
⎝ ⎠ plication is TW30. The separation performance of the three types of NF membranes for treatment of the wastewater is compared in Table 4. The results revealed that the TW30 membrane exhibited the best performance among the studied membranes. High rejection values for BOD, COD, TS and TDS in the TW30 membrane make it a good candidate for commercial applications. Although higher initial fluxes were obtained from NE70 and NE90, the separation performance of these two membranes
3.3. Effect of membrane type In order to evaluate the performance of the membrane used in this study, experiments were carried out using two other commercially
Fig. 8. Effect of NF membrane type on the permeate flux changes with time (test condition: feed flow rate of 250 L/h, feed pressure 14 bar and temperature 30 °C).
336
Journal of Water Process Engineering 19 (2017) 331–337
K. Mozaffarikhah et al.
especially for reduction of BOD and COD is somewhat lower than TW30. In the case of NE70, none of the separation performances are acceptable. The rejection performance of TDS and TS for this membrane was obtained as low as 45.55 and 29.04 percent, respectively. These low separation performances and also high fouling tendency for NE70 could be resulted and this membrane should be omitted from the list of candidates. NE90 had better performance for rejection of TDS and TS with respect to NE70, but lower rejections for BOD and COD even lower than those for NE70, makes it unsuitable for the current application.
[7]
[8] [9]
[10]
4. Conclusion
[11]
An effective method for BOD and COD removal from biodiesel’s wash-water by a commercially available NF membrane was introduced in this study. The effect of feed flow rate and inlet pressure on the rejection and permeate flux was investigated. Based on the obtained results, it can be concluded that by increasing the feed flow rate, the rejection value was enhanced for all investigated specifications. This was due to lowering the adverse effect of concentration polarization. Moreover, at a constant feed flow rate, the rejection value was enhanced by increasing the inlet pressure which shows that the system follows the Kedem-Katchalsky model for permeate flux. The maximum values for BOD and COD removals were measured at 80.7% and 81.3%, respectively, while the TDS and TS rejections were measured at 98.7% and 98.6%, under 12 bar feed pressure and 250 L/h feed flow rate. Generally speaking, it could be concluded that the NF system investigated in this work could effectively be used for BOD and COD reduction from biodiesel wash-water. As a result, the obtained effluent is soft enough to be discharged of.
[12]
[13]
[14]
[15]
[16]
[17] [18]
[19] [20]
Acknowledgement
[21]
The authors acknowledge Islamic Azad University, Mahshahr Branch, for partially supporting this research.
[22]
References [23] [1] M.R. Anur, A.Z. Abdullah, Challenges in biodiesel industry with regards to feedstock, environmental, social and sustainability issues: a critical review, Renew. Sustain. Energy Rev. 58 (2016) 208–223. [2] A. Datta, B.K. Mandal, A comprehensive review of biodiesel as an alternative fuel for compression ignition engine, Renew. Sustain. Energy Rev. 57 (2016) 799–821. [3] M.M.A. Shirazi, A. Kargari, M. Tabatabaei, B. Mostafaei, M. Akia, M. Barkhi, M.J.A. Shirazi, Acceleration of biodiesel-glycerol decantation through NaCl-assisted gravitational settling: a strategy to economize biodiesel production, Bioresour. Technol. 134 (2013) 401–406. [4] P. Verma, M.P. Sharma, G. Dwivedi, Impact of alcohol on biodiesel production and properties, Renew. Sustain. Energy Rev. 56 (2016) 319–333. [5] I.M. Atadashi, M.K. Aroua, A.A. Aziz, Biodiesel separation and purification: a review, Renew. Energy 36 (2011) 437–443. [6] M.M.A. Shirazi, A. Kargari, S. Bazgir, M. Tabatabaei, M.J.A. Shirazi, M.S. Abdullah,
[24]
[25] [26]
[27]
337
T. Matsuura, A.F. Ismail, Characterization of electrospun polystyrene membrane for treatment of biodiesel's water-washing effluent using atomic force microscopy, Desalination 329 (2013) 1–8. V.B. Veljkovic, O.S. Stamenkovic, M.B. Tasic, The wastewater treatment in the biodiesel production with alkali-catalyzed transestrification, Renew. Sustain. Energy Rev. 32 (2014) 40–60. I.M. Atadashi, M.K. Aroua, A.R.A. Aziz, N.M.N. Sulaiman, Refining technologies for the purification of crude biodiesel, Appl. Energy 88 (2011) 4239–4251. N.M. Daud, S.R.S. Abdullah, H.A. Hassan, Z. Yaakob, Production of biodiesel and its wastewater treatment technologies: a review, Process Safety Environ. Protection 94 (2015) 487–508. I.M. Atadashi, M.K. Aroua, A.A. Aziz, High quality biodiesel and its diesel engine application: a review, Renew. Sustain. Energy Rev. 14 (2010) 1999–2008. M.M.A. Shirazi, A. Kargari, A.F. Ismail, T. Matsuura, Computational fluid dynamic (CFD) opportunities applied to the membrane distillation process: state-of-the-art and perspectives, Desalination 377 (2016) 73–90. J. Kim, M. Park, H.K. Shon, J.H. Kim, Performance analysis of reverse osmosis, membrane distillation, and pressure-retarded osmosis hybrid processes, Desalination 380 (2016) 85–92. M.M.A. Shirazi, A. Kargari, M. Tabatabaei, Evaluation of commercial PTFE membranes in desalination by direct contact membrane distillation, Chem. Eng. Proc.: Process Intensif. 76 (2014) 16–25. S. Mohammadi, A. Kargari, H. Sanaeepour, K. Abbasian, A. Najafi, E. Mofarrah, Phenol removal from industrial wastewaters: a short review, Desal. Water Treat. 53 (2015) 2215–2234. S.H. Shuit, Y.T. Ong, K.T. Lee, B. Subhash, S.H. Tan, Membrane technology as a promising alternative in biodiesel production: a review, Biotechnol. Adv. 30 (2012) 1364–1380. R. Jaber, M.M.A. Shirazi, J. Toufaily, A.T. Hamieh, A. Noureddin, H. Ghanavati, A. Ghaffari, A. Zenouzi, A. Karout, A.F. Ismail, M. Tabatabaei, Biodiesel wash-water reuse using microfiltration: toward zero-discharge strategy for cleaner and economized biodiesel production, Biofuel Res. J. 2 (2015) 148–151. FILMTEC Membranes, Product Information Catalog Form No. 609-09010-0406, (2017). A.M. Hidalgo, G. Leon, M. Gomez, M.D. Murcia, E. Gomez, J.L. Gomez, Application of the Spiegler–Kedem–Kachalsky model to the removal of 4-chlorophenol by different nanofiltration membranes, Desalination 315 (2013) 70–75. L.D. Nghiem, A.I. Schafer, Critical risk points of nanofiltration and reverse osmosis processes in water recycling applications, Desalination 187 (2006) 303–312. T.O. Mahlangu, T.A.M. Msagati, E.M.V. Hoek, A.R.D. Verliefde, B.B. Mamba, Rejection of pharmaceuticals by nanofiltration (NF) membranes: effect of fouling on rejection behaviour, Physics Chem. Earth Part A/B/C 76-78 (2014) 28–34. M.M.A. Shirazi, A. Kargari, M. Tabatabaei, A.F. Ismail, T. Matsuura, Assessment of atomic force microscopy for characterization of PTFE membranes for membrane distillation (MD) process, Desal. Water Treat. 54 (2015) 295–304. C.V. Gherasim, P. Mikulasek, Influence of operating variables on the removal of heavy metal ions from aqueous solutions by nanofiltration, Desalination 343 (2014) 67–74. N. Hilal, O.O. Ogunbiyi, N.J. Miles, R. Nigmatullin, Methods employed for control of fouling in MF and UF membranes: a comprehensive review, Sep. Sci. Technol. 40 (2005) 1957–2005. T. Kaghazchi, M. Mehri, M. Takht Ravanchi, A. Kargari, A mathematical modeling of two industrial seawater desalination plants in the Persian Gulf region, Desalination 252 (2010) 135–142. W. Zhang, J. Luo, L. Ding, M.Y. Jaffrin, A review on flux decline control strategies in pressure-driven membrane processes, Ind. Eng. Chem. Res. 54 (2015) 2843–2861. A. Fouladitajar, F. ZokaeeAshtiani, H. Rezaei, A. Haghmoradi, A. Kargari, Gas sparging to enhance permeate flux and reduce fouling resistances in cross flow microfiltration, Ind. Eng. Chem. 20 (2014) 624–632. P.T.P. Aryanti, S. Subagjo, D. Ariono, I.G. Wenten, Fouling and rejection characteristics of humic substances in polysulfone ultrafiltration membrane, J. Membr. Sci. Res. 1 (2015) 41–45.
Contents lists available at ScienceDirect
Journal of Water Process Engineering journal homepage: www.elsevier.com/locate/jwpe
Membrane treatment of biodiesel wash-water: A sustainable solution for water recycling in biodiesel production process
MARK
⁎
Kosar Mozaffarikhaha, Ali Kargarib, , Meisam Tabatabaeic,d, Hossein Ghanavatic,d, Mohammad Mahdi A. Shirazie a
Department of Chemical Engineering, Mahshahr Branch, Islamic Azad University, Mahshahr, Iran Membrane Processes Research Laboratory (MPRL), Department of Chemical Engineering, Amirkabir University of Technology (Tehran Polytechnic), Tehran, Iran c Microbial Biotechnology Department, Agricultural Biotechnology Research Institute of Iran (ABRII), Agricultural Research, Education and Extension Organization (AREEO), Karaj, Iran d Biofuel Research Team (BRTeam), Karaj, Iran e Membrane Industry Development Institute, Tehran, Iran b
A R T I C L E I N F O
A B S T R A C T
Keywords: Biodiesel Wash-water Nanofiltration Wastewater treatment Environment Sustainable solution
Most biodiesel production processes lead to the generation of large volumes of highly polluted wastewater, namely wash-water effluent. The principal aim of this paper is to show the applicability and effectiveness of the nanofiltration process for treatment of biodiesel wash water effluent and possibility of recycling the treated wastewater. For this purpose, three types of commercial nanofiltration membranes (TW30, NE90 and NE70) have been examined. The effect of two important operating parameters, including operating pressure (6–14 bar) and feed flow-rate (100–250 L/h) on the permeate flux, flux decline, and also separation performances for biological oxygen demand (BOD), chemical oxygen demand (COD), total solids (TS) and total dissolved solids (TDS) were investigated for the TW30 membrane. The obtained results showed that the permeate stream obtained from a feed at a pressure of 12 bar and a flow rate of 250 L/h, through a single stage nanofiltration process have BOD, COD, TDS and TS of 85.6%, 85.8%, 97.3% and 98.8% lower than the initial wash water effluent, respectively. The results of testing the three membranes at 14 bar and 250 L/h showed that TW30 had the lowest fouling tendency and flux decline followed by NE90 and NE70, respectively. The highest rejection for BOD, COD, TDS and TS was also observed for the TW30 membrane which shows this membrane is a good candidate, among the studied membranes, for further consideration and commercial applications.
1. Introduction Fossil fuels are non-renewable sources of energy and their growing utilization has resulted in global warming and climate change. These in turn have endangered mankind’s long-term survival [1]. Biofuels as alternatives to fossil-based conventional fuels have been considered as a get-away strategy that could partially address some of the environmental concerns [2]. Among various biofuels, the Biodiesel could be produced from a variety of vegetable oils (edible and non-edible), animal fats, and waste cooking oil and has been used as an alternative to petroleum diesel [3]. Among the many advantages of biodiesel is that it can be used in diesel engines with no or minor modifications. Biodiesel is mostly produced through the transesterification reaction of triglycerides with a short chain alcohol (usually methanol) in the presence of a suitable catalyst such as sodium or potassium hydroxide [4]. Most biodiesel production processes involve a downstream process
⁎
namely the wet washing step in which hot water (50–60 °C) is used to strip the crude biodiesel of its impurities. This process leads to the production of a very polluting wastewater (i.e. the COD up to 35000 and the BOD up to 30000 mg O2/L) [5] also called biodiesel washwater. More specifically, based on our previous experiments, about 3–10 L of biodiesel wash-water is generated during this process for each liter of biodiesel [6]. It is also indicated in the literature that in 2011, worldwide generation of wash-water was approximately 28 million m3 per year [7]. This highly organic effluent consists of water, residual biodiesel, residual catalyst, soap, salts, methanol and traces of un-reacted oil and glycerol [8]. As a result, this wastewater should not be discharged of without a proper treatment step as it poses a serious threat to the environment. There are strict industrial effluent standard limits in this regard. For instance, the BOD, COD, and the oil and grease (O & G) content limitations in Malaysia are 20, 80 and 1.0 mg/L, respectively [9]. It is worth mentioning that since biodiesel wash-water is
Corresponding author. E-mail addresses: [email protected], [email protected] (A. Kargari).
http://dx.doi.org/10.1016/j.jwpe.2017.09.007 Received 8 April 2017; Received in revised form 23 August 2017; Accepted 6 September 2017 2214-7144/ © 2017 Elsevier Ltd. All rights reserved.
Journal of Water Process Engineering 19 (2017) 331–337
K. Mozaffarikhah et al.
Table 1 Pros. and Cons. of some biodiesel wash-water treatment processes [9]. Process Coagulation
Biological treatment
Adsorption
Microbial fuel cell
Pros.
Cons.
• Simple • Economical treatment technique • Proven • Economical arrangement for small areas • Versatile and suitable for small plants • Simple additional sludge is produced • Non of discharged wastewater is unaffected • pH COD removal capacity • High • High COD removal capacity
handling requirement • Chemical operation • Complicate low density sludge with low decomposition efficiency • Generates large amount of low density sludge • Generate consuming • Time to manage optimum condition first • Need further treatment • Need • Difficulties in off adsorbents discharging • Costly
reported and an economically-viable process which could be implemented on an industrial scale. Based on the above-mentioned studies, it could be concluded that the MF process is not very effective for treatment of biodiesel washwater because most of the pollutants present in wash-water are low molecular weight compounds that could not be rejected by MF membranes. Then, more tightened membranes such as ultrafiltration (UF), nanofiltration (NF) or reverse osmosis (RO) seem to be more effective. On such a basis, the present study was set to apply a commercially available NF membrane to treat biodiesel wash-water. Moreover, the impact of operating parameters such as feed pressure and flow rate on the chemical oxygen demand (COD), biological oxygen demand (BOD) and total dissolved solid (TDS) reduction of biodiesel wash-water by the developed system were thoroughly investigated.
a highly stable emulsion containing grease, oil and soap, its treatment by using conventional techniques seems difficult. In other words, water washing has been proven to result in the biodiesel meeting the international standard specifications laid out for biodiesel [10]. However, it gives rise to some disadvantages, among which high water consumption, despite the current water shortage and considerable increase in fresh water demand worldwide [11], is of major concern. As a result, purification of this environmentally hazardous wastewater via hightech strategies should be taken into serious consideration. To date, several treatment processes have been developed for treating the biodiesel wash-water such as physic-chemical treatment, electro-chemical treatments, coupled chemical and electrochemical treatments, advanced oxidation technologies, biological treatments and, integrated treatment processes [7,9]. The pros and cons for some of these processes are listed in Table 1. Membrane separation processes are clean and energy effective processes [12], in which fluid separation takes place without the contribution of a third phase and is usually without phase change. Among those, the most commonly known process is water desalination, which has been traditionally performed by highly energy intensive distillation processes like multi-effect distillation (MED) and multi-stage flash distillation (MSF) [13]. The use of membranes as the separation medium has revolutionarily converted this process as well as many other processes into cleaner, more environmentally-friendly, less expensive, less energy-intense, and more compact processes with less foot-prints and without any by products or residual wastes generated [14]. Accordingly, it is believed that the application of membrane separation processes in biodiesel processing [15], in particular its wastewater treatment, could also be very effective. However, only few works have been performed on biodiesel and its wastewaters by membrane separation processes [7,9]. Shirazi et al. [6] used a microfiltration (MF) process by a superhydrophobic electrospun nano-fibrous polystyrene membrane to treat the biodiesel's water-washing effluent and reportedly managed to decrease COD and BOD by 75% and 55%, respectively. They used a contact heating method to modify the membranes’ surface, and then characterized them using atomic force microscopy (AFM) for their surface features. In another work, Jaber and his co-workers [16] used a simple microfiltration-based procedure to treat and subsequently re-use biodiesel wash-water. They applied commercial polypropylene (PP) MF membranes with 1 and 5 μm pore sizes and also commercial activate carbon and sand filter cartridges to perform the treatment in a hybrid system. The highest COD and BOD reductions obtained in their work stood at around 18.5% and 12.4%, respectively, which are lower compared to the work of Shirazi et al. [6]. Moreover, they argued that their innovative PP-based microfiltration followed by sand and activated carbon separations and in combination with 70% dilution rate with fresh water enabled the re-use of biodiesel wash-water which still led to a standard-quality biodiesel product and resulted in up to 15% less water consumption. Despite all the efforts put into treating this stringent wastewater, there is still a wide gap between the findings
2. Experimental 2.1. Materials Biodiesel wash-water was provided by the Biofuel Research Team (BRT, Karaj, Iran). The biodiesel itself was produced from waste cooking oil and methanol (1:6 molar ratio) through the alkali-catalyzed transesterification reaction (potassium hydroxide 1 wt.%). The specification of the crude wash-water is shown in Table 2. A commercially available NF membrane, (TW30-1812-50) from Dow-Filmtec Co. [17] was used as the separation media. The membrane had a thin-film composite structure composed of a very thin and selective polyamide layer at the top, a polysulfone UF structure as the mid-layer and a non-woven polyester fabric as the support layer. The specification of the membrane module, which has been reported by the manufacturer, is tabulated in Table 3. The used membrane was cut from the membrane module. 2.2. Methods 2.2.1. Experimental setup Fig. 1 represents the schematic of the experimental setup. A 25 L Table 2 The average specification of the crude wash-water. Property
Value
Unit
EC pH COD BOD TDS TSS TS
1086 6.98 19880 18335 1238 1370 2608
μS/cm – mg/L mg/L mg/L mg/L mg/L
EC: Electrical conductivity. TSS: Total suspended solids. TS: Total solids.
332
Journal of Water Process Engineering 19 (2017) 331–337
K. Mozaffarikhah et al.
controlled at 30 °C. For each experiment, the membrane was soaked in distilled water for at least 48 h. The feed was pumped into the membrane module at predetermined pressure and flow rate. More specifically, the effect of different feed flow rates (i.e. 100, 150, 200, and 250 L/h) at constant pressure and the effect of inlet pressure (i.e. 6, 8, 10, 12 and 14 bar) at constant feed flow rate were investigated. After measuring the permeate stream flow rate, it was recycled (together with the retentate stream) back into the feed tank. Sampling from the permeate stream was done when the steady-state condition was established. The time required for attaining the steady-state was found different according to the circumstances of each experiment, but an average of 7 h of operation (i.e. recirculation) was sufficient for all the experiments to reach the steady-state condition. Permeate flux and solute rejections were calculated using Eq. (1) and (2):
Table 3 The specification of the NF membrane module used in this study [17]. Specification
Value
Unit
Area Diameter Length Max. working pressure Max flow rate Max. feed water temperature pH range
4.8 1.75 10 300 7.6 45 2–11
ft2 in in psi L/min °C –
stainless steel tank was used as the feed reserve. A low-pressure pump (So-Pure, Korea) and a 5 μm pore size propylene depth pre-filter were used as the pre-filtration system before the feed enters the main highpressure pump. A vertical multistage centrifugal stainless steel pump (LVB 1–36, Leo Co., China) with maximum pressure and flow of 25 bar and 1.8 m3/h was used as the main pump in the feed stream, respectively. The adjustment of the pump outlet pressure was performed by an AC frequency control system (N700E, LG Co., Korea). Then, the feed stream entered the membrane module, which was a handmade stainless steel, cross-flow filtration system containing a flat sheet membrane. The active area of the flat membrane was 78 cm2 (W = 5.2 and L = 15 cm). Finally, the retentate stream passed through a rotameter (25–250 L/h, Fischer & Porter 3400, Gottingen, Germany)) and returned into the feed tank. The final adjustment of the feed stream pressure and flowrate was performed by two globe valves before and after the membrane module. As shown in the schematic diagram, two pressure gauges (Bourdon Sedeme, 0–16 bar and Sangan Sanat, 0–40 bar) were placed before and after the membrane module to monitor the operating pressure. Temperature adjustment of the feed was performed by a cold water stainless steel coil, which was immersed into the feed reserve. Electrical conductivity (EC) was measured by a Metrohm conductivity meter (Metrohm 720, Switzerland) and used for determination of TDS. A digital pH meter (Metrohm-780 Herisau, Switzerland) equipped with a Pt-1000 thermometer for temperature correction was used for measuring the pH of the aqueous solutions.
Jp =
Qp A
Cp ⎞ R (%) = ⎜⎛1 − ⎟ × 100 Cf ⎠ ⎝
(1)
(2)
Where Qp is the permeate stream flowrate (L/h),A is the effective area of the used membrane (m2), R (%) is the percentage of rejection for a specific solute, Cp and Cfare the permeate stream and feed concentrations, respectively [18,19].
2.2.3. Analytical methods The Open-reflux and Azide methods were used for COD and BOD5 measurements, respectively, according to the standard method for examination of water and wastewater. The pH and electrical conductivity (EC) was measured in the samples with an electrometric pH meter and EC meter, respectively; and TDS (total dissolved solids), TSS (total suspended solids), and TS (total solids) were detected by the method described in the previous work [6]. The applied membrane, before and after NF experiments, was characterized using SEM (scanning electron microscopy) for its morphological features and fouling characteristics, as shown in Fig. 2.
2.2.2. Procedure In all the experiments conducted, the feed temperature was
Fig. 1. A general scheme of the experimental apparatus.
333
Journal of Water Process Engineering 19 (2017) 331–337
K. Mozaffarikhah et al.
Fig. 2. SEM images of applied commercial NF membrane in this work, before (left) and after (right) wash-water treatment experiment.
thinner polarized layer at higher feed flows are at times significantly less than the thicker one for lower feed flow rates [23]. As a consequence, the time for attaining steady-state flux would be shorter as the feed flow rate increases. Having considered the above-mentioned results, it could also be stated that the feed flow rate is effective on wash-water characteristics, e.g. TDS, TS, BOD and COD. The effect of variations in feed flow (100, 150, 200 and 250 L/h) on the rejection of these items at 14 bar is shown in Fig. 4. The results showed that the rejection percentage (%) for TS and TDS are almost identical by using the proposed NF membrane at 14 bar operating pressure. High rejection values (i.e. > 98%) for both TDS and TS were obtained, which is a considerable result compared with conventional treatment processes (see Table 2 in Ref. [7]). However, the effect of feed flow rate on the variation of TDS and TS rejections was negligible. Fig. 4 also shows the effect of feed flow rate on BOD and COD rejection at the mentioned inlet pressure (i.e. 14 bar). As could be observed, by increasing the flow rate, the rejection rate was increased as well, but similar BOD and COD rejection behavior is observed. The BOD and COD rejections under the proposed condition were at the range of 67–81%. It could be claimed that this is an appropriate result for a single-stage membrane process. The obtained results mentioned above could be concluded by this fact that an increase in the concentration polarization layer has two drawbacks on the membrane performance. On the one hand, by increasing the polarization layer, the hydraulic resistance of the
3. Results and discussion 3.1. Effect of feed flow rate Feed flow rate is one of the factors that can affect membrane performance. To test the effect of this operating parameter, experiments were conducted with biodiesel wash-water under different feed flow rates ranging from 100 to 250 L/h at 14 bar constant inlet pressure. As could be observed in Fig. 3, the permeate fluxes decreased during the first 25 min of the experiment for all inlet flow rates. This could be due to the gradual compaction of the membrane [20] and development of the concentration polarization layer [21] at different feed flow rates. This declining trend of the flux proceeded until it reached a slight steady state and remained constant over time for the upper range of flow rates, i.e. 200 and 250 L/h. While for the lower bond of flow rate, i.e. 100 L/h, the permeate flux became constant after about 375 min. Moreover, the experimental results showed that by increasing the feed flow rate from 100 to 250 L/h (under constant operating pressure), the membrane permeate flux was increased. For instance, the initial flux for the lowest and highest flow rates are about 47 L/h and 54 L/h, and then reduced to 33 L/h and 48 L/h for the same issues, respectively. It is worth quoting that the time to reach the steady-state decreased by increasing the flow rate value. This fact could be ascribed to the concentration polarization effect. More specifically, by increasing the feed flow rate, the thickness of the concentration polarization layer is decreased resulting in less hydraulic resistance and higher flux through the membrane [22]. It is claimed that the time for formation of a
Fig. 3. Effect of feed flow rate on the membrane permeate flux at 14 bar and 30 °C.
334
Journal of Water Process Engineering 19 (2017) 331–337
K. Mozaffarikhah et al.
Fig. 4. Effect of feed flow rate on the rejection of TDS, TS, BOD and COD at 14 bar and 30 °C.
Fig. 6. Effect of feed pressure on the steady state flux at feed flow rate of 250 L/h and 30°C.
membrane increases which results in lower permeation rates. On the other hand, the real concentration of the rejected species on the membrane surface increases which results in higher concentration of the permeated stream (i.e. lower rejection rates). Therefore, membrane separations at higher feed flow rates and higher operating pressures would be more favorable, specifically for samples like biodiesel washwater. The obtained results are also in agreement with the known Kedem–Katchalsky theory in which higher permeation rates would result in higher rejections [24]. Hence, the 250 L/h feed flow rate was investigated for the following experiments.
As all the experiments were carried out at the highest feed flow rate, based on the conclusion in the previous section, the initial flux decline for each case was not considerable especially at low pressures (see Fig. 5). In other words, in addition to the membrane compaction, the rate of the polarized layer formation and cake compaction are lower and only a small flux decline is observed. By increasing the feed pressure over 8 bar, further membrane compaction, cake formation and polarization, and as a consequence considerable flux decline, are expected [25]. It is worth quoting that under such feed pressures, the rate of permeation fluxes were high which led to reaching pseudo-steady conditions within a few seconds that could not be measured manually, meaning that the initial sharp flux decline could not be observed at higher operating pressures ( > 8 bar). By increasing the feed pressure, the time for establishing the real steady-state condition is also increased. This time at 6 bar is about 100 min while this time for 14 bar is measured at about 360 min. The influence of feed pressure on the steady state permeation flux is shown in Fig. 6. As could be observed, there is a linear relationship between the steady permeate flux and applied pressure. A little deviation is observed at pressures of 6 and 8 bar. This is due to the
3.2. Effect of feed pressure The effect of operating pressure (i.e., 6, 8, 10, 12 and 14 bar) on the membrane permeate flux at constant 250 L/h feed flow rate and 30 °C is shown in Fig. 5. Under different operating pressures, membrane permeates decreased over the operation time for individual inlet pressures. This is due to the gradual formation of concentration polarization and the cake layer on the membrane surface during almost the first 150 min of operation. This is while steady state permeate fluxes were reached after this time.
Fig. 5. Effect of operating pressure on the membrane permeate flux at 250 L/h feed flow rate and 30°C.
335
Journal of Water Process Engineering 19 (2017) 331–337
K. Mozaffarikhah et al.
Table 4 Separation performance of the three studied membranes for treatment of biodiesel wash water (test condition: feed pressure: 14 bar, feed flow rate: 250 L/h). Membrane Type
Rejection [%]
TW30 NE90 NE70
COD
BOD
TDS
TS
81.31 77.94 80.88
80.56 74.13 79.17
98.72 86.56 45.55
98.60 86.19 29.04
available membranes, i.e. NE70 and NE90 both from CSM, Korea. These membranes are used widely for brackish water desalination, especially where high fluxes are expected. The results of testing these membranes are shown in Fig. 8. The test condition was selected as 14 bar feed pressure and 250 L/h feed flow rate. The results reveal that the order of initial flux for these three membranes is NE70 > NE90 > TW30. In the case of NE70, a sharp flux decline was observed that shows severe fouling and a certain extent of compaction of the membrane in the first 200 min of operation. NE90 showed lower permeate flux than NE70 with lower flux decline. After about 600 min of operation, the flux of NE90 and TW30 membranes were nearly the same and both higher than NE70. The interesting point in these experiments was that TW30 showed a nearly flat flux with time. In other words, the permeate flux of the TW30 membrane declined very slowly compared to the other two membranes. This reveals that the TW30 membrane is an anti-fouling membrane and is more suitable than the other two membranes for treatment of biodiesel wash water. The overall flux decline of the three membranes within 600 min was calculated as well. The lowest flux decline belongs to the TW30
Fig. 7. Effect of operating pressure on the rejection of TDS, TS, BOD and COD (feed flow rate of 250 L/h feed flow rate and temperature of 30°C).
polarization effect and initial flux decline. However, as the feed pressure is increased, the steady state flux becomes more linear. In other words, according to the well-known Darcy's law of filtration, the effect of inlet pressure on the hydraulic resistance against membrane permeation is negligible so that the permeation flux becomes a linear function of applied pressure [26,27]. The effect of various inlet pressures on the rejection of TDS, TS, BOD and COD at constant 250 L/h feed flow rate is shown in Fig. 7. Approximately the same rejection behavior for TDS and TS, and also for COD and BOD could be observed, respectively. By increasing the feed pressure, the rejection value (%) for all species is increased. Considering the effect of feed pressure on the permeation flux (see Fig. 6), the permeation flux increases as the feed pressure is increased, which is in agreement with the Kedem–Katchalsky theory for rejection, as mentioned earlier. The maximum rejection values for BOD, COD, TDS and TS at 14 bar are measured as 80.7, 81.3, 98.7 and 98.6 percent, respectively.
L
L
⎛ ⎞⎞ ⎛ ⎞⎞ ⎛ ⎛ 2 2 membrane ⎜0.11 ⎝ m h. h ⎠ ⎟, then to NE90 ⎜0.26 ⎝ m h. h ⎠ ⎟ and finally to NE70 ⎝ ⎠ ⎝ ⎠ ⎛ L ⎞⎞ ⎛ 2 ⎝m .h⎠ ⎜0.61 h ⎟. These results show that the best membrane for this ap-
⎝ ⎠ plication is TW30. The separation performance of the three types of NF membranes for treatment of the wastewater is compared in Table 4. The results revealed that the TW30 membrane exhibited the best performance among the studied membranes. High rejection values for BOD, COD, TS and TDS in the TW30 membrane make it a good candidate for commercial applications. Although higher initial fluxes were obtained from NE70 and NE90, the separation performance of these two membranes
3.3. Effect of membrane type In order to evaluate the performance of the membrane used in this study, experiments were carried out using two other commercially
Fig. 8. Effect of NF membrane type on the permeate flux changes with time (test condition: feed flow rate of 250 L/h, feed pressure 14 bar and temperature 30 °C).
336
Journal of Water Process Engineering 19 (2017) 331–337
K. Mozaffarikhah et al.
especially for reduction of BOD and COD is somewhat lower than TW30. In the case of NE70, none of the separation performances are acceptable. The rejection performance of TDS and TS for this membrane was obtained as low as 45.55 and 29.04 percent, respectively. These low separation performances and also high fouling tendency for NE70 could be resulted and this membrane should be omitted from the list of candidates. NE90 had better performance for rejection of TDS and TS with respect to NE70, but lower rejections for BOD and COD even lower than those for NE70, makes it unsuitable for the current application.
[7]
[8] [9]
[10]
4. Conclusion
[11]
An effective method for BOD and COD removal from biodiesel’s wash-water by a commercially available NF membrane was introduced in this study. The effect of feed flow rate and inlet pressure on the rejection and permeate flux was investigated. Based on the obtained results, it can be concluded that by increasing the feed flow rate, the rejection value was enhanced for all investigated specifications. This was due to lowering the adverse effect of concentration polarization. Moreover, at a constant feed flow rate, the rejection value was enhanced by increasing the inlet pressure which shows that the system follows the Kedem-Katchalsky model for permeate flux. The maximum values for BOD and COD removals were measured at 80.7% and 81.3%, respectively, while the TDS and TS rejections were measured at 98.7% and 98.6%, under 12 bar feed pressure and 250 L/h feed flow rate. Generally speaking, it could be concluded that the NF system investigated in this work could effectively be used for BOD and COD reduction from biodiesel wash-water. As a result, the obtained effluent is soft enough to be discharged of.
[12]
[13]
[14]
[15]
[16]
[17] [18]
[19] [20]
Acknowledgement
[21]
The authors acknowledge Islamic Azad University, Mahshahr Branch, for partially supporting this research.
[22]
References [23] [1] M.R. Anur, A.Z. Abdullah, Challenges in biodiesel industry with regards to feedstock, environmental, social and sustainability issues: a critical review, Renew. Sustain. Energy Rev. 58 (2016) 208–223. [2] A. Datta, B.K. Mandal, A comprehensive review of biodiesel as an alternative fuel for compression ignition engine, Renew. Sustain. Energy Rev. 57 (2016) 799–821. [3] M.M.A. Shirazi, A. Kargari, M. Tabatabaei, B. Mostafaei, M. Akia, M. Barkhi, M.J.A. Shirazi, Acceleration of biodiesel-glycerol decantation through NaCl-assisted gravitational settling: a strategy to economize biodiesel production, Bioresour. Technol. 134 (2013) 401–406. [4] P. Verma, M.P. Sharma, G. Dwivedi, Impact of alcohol on biodiesel production and properties, Renew. Sustain. Energy Rev. 56 (2016) 319–333. [5] I.M. Atadashi, M.K. Aroua, A.A. Aziz, Biodiesel separation and purification: a review, Renew. Energy 36 (2011) 437–443. [6] M.M.A. Shirazi, A. Kargari, S. Bazgir, M. Tabatabaei, M.J.A. Shirazi, M.S. Abdullah,
[24]
[25] [26]
[27]
337
T. Matsuura, A.F. Ismail, Characterization of electrospun polystyrene membrane for treatment of biodiesel's water-washing effluent using atomic force microscopy, Desalination 329 (2013) 1–8. V.B. Veljkovic, O.S. Stamenkovic, M.B. Tasic, The wastewater treatment in the biodiesel production with alkali-catalyzed transestrification, Renew. Sustain. Energy Rev. 32 (2014) 40–60. I.M. Atadashi, M.K. Aroua, A.R.A. Aziz, N.M.N. Sulaiman, Refining technologies for the purification of crude biodiesel, Appl. Energy 88 (2011) 4239–4251. N.M. Daud, S.R.S. Abdullah, H.A. Hassan, Z. Yaakob, Production of biodiesel and its wastewater treatment technologies: a review, Process Safety Environ. Protection 94 (2015) 487–508. I.M. Atadashi, M.K. Aroua, A.A. Aziz, High quality biodiesel and its diesel engine application: a review, Renew. Sustain. Energy Rev. 14 (2010) 1999–2008. M.M.A. Shirazi, A. Kargari, A.F. Ismail, T. Matsuura, Computational fluid dynamic (CFD) opportunities applied to the membrane distillation process: state-of-the-art and perspectives, Desalination 377 (2016) 73–90. J. Kim, M. Park, H.K. Shon, J.H. Kim, Performance analysis of reverse osmosis, membrane distillation, and pressure-retarded osmosis hybrid processes, Desalination 380 (2016) 85–92. M.M.A. Shirazi, A. Kargari, M. Tabatabaei, Evaluation of commercial PTFE membranes in desalination by direct contact membrane distillation, Chem. Eng. Proc.: Process Intensif. 76 (2014) 16–25. S. Mohammadi, A. Kargari, H. Sanaeepour, K. Abbasian, A. Najafi, E. Mofarrah, Phenol removal from industrial wastewaters: a short review, Desal. Water Treat. 53 (2015) 2215–2234. S.H. Shuit, Y.T. Ong, K.T. Lee, B. Subhash, S.H. Tan, Membrane technology as a promising alternative in biodiesel production: a review, Biotechnol. Adv. 30 (2012) 1364–1380. R. Jaber, M.M.A. Shirazi, J. Toufaily, A.T. Hamieh, A. Noureddin, H. Ghanavati, A. Ghaffari, A. Zenouzi, A. Karout, A.F. Ismail, M. Tabatabaei, Biodiesel wash-water reuse using microfiltration: toward zero-discharge strategy for cleaner and economized biodiesel production, Biofuel Res. J. 2 (2015) 148–151. FILMTEC Membranes, Product Information Catalog Form No. 609-09010-0406, (2017). A.M. Hidalgo, G. Leon, M. Gomez, M.D. Murcia, E. Gomez, J.L. Gomez, Application of the Spiegler–Kedem–Kachalsky model to the removal of 4-chlorophenol by different nanofiltration membranes, Desalination 315 (2013) 70–75. L.D. Nghiem, A.I. Schafer, Critical risk points of nanofiltration and reverse osmosis processes in water recycling applications, Desalination 187 (2006) 303–312. T.O. Mahlangu, T.A.M. Msagati, E.M.V. Hoek, A.R.D. Verliefde, B.B. Mamba, Rejection of pharmaceuticals by nanofiltration (NF) membranes: effect of fouling on rejection behaviour, Physics Chem. Earth Part A/B/C 76-78 (2014) 28–34. M.M.A. Shirazi, A. Kargari, M. Tabatabaei, A.F. Ismail, T. Matsuura, Assessment of atomic force microscopy for characterization of PTFE membranes for membrane distillation (MD) process, Desal. Water Treat. 54 (2015) 295–304. C.V. Gherasim, P. Mikulasek, Influence of operating variables on the removal of heavy metal ions from aqueous solutions by nanofiltration, Desalination 343 (2014) 67–74. N. Hilal, O.O. Ogunbiyi, N.J. Miles, R. Nigmatullin, Methods employed for control of fouling in MF and UF membranes: a comprehensive review, Sep. Sci. Technol. 40 (2005) 1957–2005. T. Kaghazchi, M. Mehri, M. Takht Ravanchi, A. Kargari, A mathematical modeling of two industrial seawater desalination plants in the Persian Gulf region, Desalination 252 (2010) 135–142. W. Zhang, J. Luo, L. Ding, M.Y. Jaffrin, A review on flux decline control strategies in pressure-driven membrane processes, Ind. Eng. Chem. Res. 54 (2015) 2843–2861. A. Fouladitajar, F. ZokaeeAshtiani, H. Rezaei, A. Haghmoradi, A. Kargari, Gas sparging to enhance permeate flux and reduce fouling resistances in cross flow microfiltration, Ind. Eng. Chem. 20 (2014) 624–632. P.T.P. Aryanti, S. Subagjo, D. Ariono, I.G. Wenten, Fouling and rejection characteristics of humic substances in polysulfone ultrafiltration membrane, J. Membr. Sci. Res. 1 (2015) 41–45.