Keywords

1 Introduction

Bilgewater is a chronic source of oil pollution that is being increasingly discharged to the marine ecosystem. Bilgewater is complex shipboard wastewater comprising mainly seawater in addition to various ship pollutants (e.g., fuels, lubrication oils, solid particulates, etc.) [1]. The environmental and ecological impacts of ship fluids discharge into the ocean are very far-reaching. Coastal plants are very vulnerable to the hydrocarbon residues in the bilgewater. The heavy residues can accumulate in the soil and adversely affect its permeability. Marine mammals are as well threatened by the bilgewater pollutants, they are susceptible to accidental consumption of tar residuals and physical contact with oils all of which cause severe health conditions. Therefore, the treatment of bilgewater is mandatory for sustaining the ecosystem and for potential reutilization for the irrigation of coastal plants [2]. Minimization or elimination of ship fluids discharge to the ecosystem adheres to the Sustainable Development Goals (SDGs) by sustaining the marine life, the coastal plants, and the water ways (Goals No. 14, 15, and 6 respectively).

Various treatment technologies have been proposed for the treatment of oil-contaminated seawater including electrodialysis, adsorption, coagulation and floatation, photodegradation, and chemical oxidation [3, 4]. Among the proposed processes, membrane technologies exhibited outstanding performance in terms of effluent quality and process cost [5,6,7]. Pressure-driven membrane technologies such as ultrafiltration and reverse osmosis have been reported widely in the literature for the treatment of oil-polluted wastewater [8,9,10,11]. However, the presence of oil in water causes severe fouling and subsequent deterioration of the membrane performance [12,13,14,15]. Previous studies tested ultrafiltration units for the treatment of bilgewater. Crossflow pilot ultrafiltration experiments were conducted for the treatment of oily bilgewater containing 6100 Total dissolved solids (TDS) and 27.2 ppm oil. The oil content in the obtained permeates was reduced by 80% however, no considerable removal of salts was observed [16,17,18].

Membrane distillation processes are generally utilized for the elimination of salts from oily-saline wastewater. Membrane distillation is a thermally driven membrane process that can utilize waste energy or renewable energy to desalinate water [19, 20]. Capillary polypropylene commercial membrane (Accurel PP V8/2 HF) was tested in direct contact membrane distillation (DCMD) for the treatment of bilgewater from a port in the Baltic Sea (TDS ~ 6500 ppm, average oil = 40 ppm). Excellent separation efficiency was achieved, and the obtained distillate was practically freshwater [21, 22]. Treatment of high-oil content bilgewater can foul MD membranes and cause severe wetting. A recent study investigated the treatment of bilgewater containing 360 ppm oil and TDS of 3700 ppm using a capillary polypropylene membrane in DCMD configuration. Oil rejection was virtually 100% however, salt rejection was only ~ 10% indicating severe membrane wetting. The authors included an ultrafiltration pretreatment step to minimize the oil load on the MD module from 360 to 10 ppm. The permeate quality was enhanced greatly and freshwater was obtained [23, 24].

In this work, hydrophobic polytetrafluoroethylene membrane is—tested for the treatment of bilgewater from a ship berth in the Suez Canal using air gap membrane distillation (AGMD). Membrane characterization is carried out to assess the fouling behavior and separation efficiency of the membrane. To our knowledge, no previous study tested the treatment of bilgewater in AGMD configuration using polytetrafluoroethylene and investigated the fouling behavior on the membrane surface.

2 Materials and Methods

Polytetrafluoroethylene (PTFE) membrane was obtained from STERLITECH USA, the membrane characteristics are shown in Table 1. Bilgewater was collected from a ship berth in the city of Port Said, Egypt.

Table 1 Characteristics of the membrane utilized in this study
Full size table

Membrane characterization was carried out to assess the fouling behavior of the membrane-bilgewater system. Surface morphology was analyzed by scanning electron microscopy (TESCAN MIRA). The samples were fixed on pieces of conductive carbon tape and coated with 10 nm of silver using argon plasma sputtering. Surface composition was detected before and after experimental bilgewater treatment by using energy dispersive X-ray (OXFORD Xplore) using a landing energy of 15 keV and a beam current of 100 picoamps. The active functional groups and organic fouling bonding to the membrane were detected by Fourier transform infrared spectroscopy (FTIR) (BRUKER ALPHA II equipped with ATR crystal). The surface roughness of the membrane was determined by atomic force microscopy (Nanosurf FlexAFM).

Membrane performance was tested using a pilot-scale AGMD system. A graphical illustration of the system is shown in Fig. 1 and a photograph of the actual system is shown in Fig. 2. The system comprises four main sections:

Fig. 1
An illustration of the pilot M D module, showing the active connections of a pressure transmitter, temperature transmitter, temperature control, and rotameter to both the hot and cold fluid cycles, along with their electrical connections.

Illustration of the pilot MD module utilized in this work. The red lines represent the hot fluid cycle, the blue lines represent the cold fluid cycle, and the dashed lines represent the electrical connections. The abbreviations “PT”, “TT”, “TC”, and “FI” are “pressure transmitter”, “temperature transmitter”, “temperature control”, and “rotameter”, respectively

Full size image
Fig. 2
A photo of an experimental system includes synthetic P W tank, feed pump, rotameter, thermocouples, pressure transmitters, membrane module, cooling water pump, cooling water tank, refrigeration cycle, data logger, electricity panel.

The actual experimental system utilized in this study; 1—synthetic PW tank, 2—feed pump, 3—rotameter, 4—thermocouples, 5—pressure transmitters, 6—membrane module, 7—cooling water pump, 8—cooling water tank, 9—refrigeration cycle, 10—data logger, 11—electricity panel

Full size image
  1. 1.

    Hot feed cycle.

  2. 2.

    Cooling fluid cycle.

  3. 3.

    The membrane unit

  4. 4.

    Monitoring and control.

The separation efficiency of the PTFE membrane was tested in the system described. The membrane used was 10 cm × 10 cm. The feed temperature is 70 °C and the cooling temperature was fixed at 20 °C. The feed flow rate was 2 l/min.

The purification efficiency was determined by measuring the TDS, TSS, Turbidity, and chemical oxygen demand (COD) of the bilgewater before and after treatment. TDS and TSS were measured by conductivity measurement and gravimetric method respectively. The COD was measured by the dichromate method using spectrophotometry. The error in COD measurements due to the presence of a high concentration of chloride ion in bilgewater was compensated by diluting the sample and using a COD kit that contains mercuric (II) sulfate. The flux, J (l/m2 h) was calculated using the following equation (Eq. 1):

$$J= \frac{W}{\rho A\Delta t}$$
(1)

where (W) is the permeate mass, (\(\rho )\) is the density (A) is the area of the membrane, and (\(\Delta t)\) is the experiment time.

3 Results and Discussion

3.1 Surface Morphology and Surface Composition

Scanning electron microscope images of the PTFE membrane before and after the treatment of bilgewater are shown in Fig. 3. In MD operation, the surface hydrophobicity of the membrane is crucial to prevent the deposition of salts and membrane wetting. In addition to the presence of the hydrophobic C-F function group, the presence of a rough surface at the Cassie-Baxter state may capture micro air bubbles and minimize wetting (Fig. 3). However, higher roughness may induce the deposition of inorganic scales. The morphology of the PTFE membrane (Fig. 4, left) indicated the presence of surface roughness. Deposition of salts is obvious (Fig. 4 right) after the treatment of bilgewater. The size of the deposits is variable, coarse salt particles are observed at magnifications of 1 kx and 5 kx. At a magnification of 45 kx, the attachment of particles to the membrane surface is evident.

Fig. 3
An illustration of a circular chamber labeled, Drop or Cassie Baxter over a drop with air pockets between them.

Illustration of the underwater Cassie-Baxter state

Full size image
Fig. 4
4 microscopic images of P T F E membrane with the cracks and dark spots over the surface.figure 4

Field-emission scanning electron microscope images of the pristine PTFE membrane (left) and images of the PTFE membrane after treatment of the bilgewater (right)

Full size image

Energy dispersive X-ray analyses were performed to assess the nature of the deposited particles. EDS analyses of the PTFE membrane before and after treatment are given in Table 2. Elemental mapping of the membrane after treatment is shown in Fig. 5.

Table 2 Elemental analysis of the PTFE membrane before and after treatment of bilgewater
Full size table
Fig. 5
2 illustrations. a. E D S intensity chart with 12 different elements mapped. b. Elemental mapping of a surface with small dots densely packed.

Elemental distribution data of the PTFE membrane after bilgewater treatment. a The EDS intensity chart and b elemental mapping of the membrane surface

Full size image

The Elemental analysis revealed the presence of a wide variety of salts and minerals in the bilgewater. Sodium and calcium are the most abundant cations while chloride is the most abundant anion. The presence of oxygen may be attributed to the presence of Silicates, iron oxides, or carbonate ions in the bilgewater. The elemental mapping illustrates that sodium ions (which represent the formation of NaCl scales) are distributed uniformly over the membrane surface. However, Calcium ions tend to be present in large deposits or particle aggregates, these observations suggest two different scaling mechanisms. Sodium chloride precipitates due to the effect of concentration polarization, while calcium salts normally precipitate from water at higher temperatures and form salt clusters.

3.2 Chemical Structure

The FTIR spectrum was captured within the wave number region of 400–4000 cm−1 with a resolution of 4 cm−1. The IR absorption spectrum for the PTFE membrane is shown in Fig. 6. The CF2 characteristic peaks of the symmetric and asymmetric stretching of C–F are observed at 1204 cm−1 and 1148 cm−1 respectively. Lower wave number CF2 characteristic peaks of the deformation and rocking modes are observed at 550 cm−1 and 640 cm−1 respectively. The C–H bending peak at 755 cm−1 and the strong C–H stretching peaks at 2850 and 2930 cm−1 represent the polypropylene support. The presence of noise in the FTIR chart is due to the trapping of air molecules between the very thin membrane coupon and the ATR crystal.

Fig. 6
A graph of absorbance units versus wavenumber has a fluctuating curve with its highest peak at (2800. 0.012). Values are approximated.

FTIR absorbance spectrum of the polytetrafluoroethylene membrane

Full size image

3.3 Membrane Surface Roughness

Numerous studies have established the relationship between surface roughness and the liquid-repel ability of the membrane surface. Surface roughness was determined by obtaining height images from atomic force microscopy. The membrane was imaged in dynamic force mode using a gold-coated silicone tip with an average resonance frequency of 191 kHz and height of 13 μm. The height color map and 3D height map for a sample area of 225 μm2 are shown in Fig. 7.

Fig. 7
2 microscopic images of a membrane uneven surface with peaks and patches of different lengths.

Atomic force microscopy height images featuring the topography of the membrane surface

Full size image

The average line roughness parameter Ra and area roughness parameter Sa were 0.7 μm and 0.73 μm respectively. The height fluctuation ranged from − 0.9 to 2.8 μm.

3.4 Membrane Testing Experiment

The membrane flux was approximately stable during the experiment with little fluctuations. The membrane flux and salt rejection percentage are plotted as a function of time and the results are presented in Fig. 8. The initial water flux was 4 l/m2 h. The flux remained stable at 3.4 l/m2 h during the first hour then deteriorated to 2.6 l/m2 h and finally increased again to 3.3 l/m2 h and remained stable. The initial deterioration in flux is attributed to the initial deposition of foulants on the membrane surface. However, the membrane flux increased again and stabilized due to the widening of pores which is obvious in the SEM images after treatment (Fig. 4, right). The increase in pore size is due to the simultaneous effect of the elevating membrane temperature and the passage of hot water vapor.

Fig. 8
Scatter plot for flux and salt rejection percentage versus time in minutes has data points plotted for Flux and Rejection.

AGMD performance of the PTFE membrane against the oily-saline bilgewater. The feed and cold temperatures were 70 ± 0.5 °C and 20 ± 0.5 °C respectively. The cross flow velocity of the feed was 0.167 m/s

Full size image

Salt rejection was always above 99%. The COD of the permeate was 70% less than that of the feed. The feed and permeate water qualities are given in Table 3.

Table 3 Water quality of the feed and the permeate
Full size table

4 Conclusion

An experimental study was carried out to assess the effectiveness of using AGMD for treating saline oil-polluted seawater. Various membrane characteristics were measured and their influence on the membrane performance and fouling tendency were investigated. The membrane morphology was uniform with an equal distribution of pores. Elemental analysis revealed the deposition of inorganic monovalent and divalent salts in addition to a few metal oxides. Sodium and calcium were the most abundant cations and chlorine was the most abundant anion. Calcium formed large aggregates which on long-term operation will clog the pores and damage the membrane and is, therefore, recommend being removed in a pretreatment step. Surface topography showed high variation in features height which enhanced the roughness. The average linear roughness parameter was 0.7 μm. AGMD process using PTFE membrane showed an excellent solids rejection above 99% and COD minimization up to 70%. The membrane flux increased on increasing the membrane pore size and salt rejection was not affected considerably. The initial water flux was 4 l/m2 h and the flux stabilized after 3 h at 3.3 l/m2 h. The membrane distillation process is not energy intensive and can be driven by waste heat or renewable energy sources which renders the process applicable in ships, ports, or coastal treatment facilities. The process showed promising results in handling such a challenging feed and further elaboration of the process parameters and membrane materials is recommended.

5 Recommendations

Membrane distillation is a recent technology that is still at its early stages. It is therefore recommended to test the AGMD system at different operating conditions (temperature, flow rate, etc.). Different membrane materials can yield distinctive results; PTFE membrane showed outstanding results however, modification to the intrinsic membrane characteristics (e.g. porosity, hydrophobicity, etc.) can improve the process sustainability. The AGMD system offers low thermal conduction losses, yet this comes at the expense of reducing the water flux. The air gap configuration should be compared with other configurations such as direct contact or vacuum membrane distillation to assess the most efficient configuration. The behaviour of the system varies greatly with the wastewater to be treated. Collecting different samples of bilge water from different locations and perform a comparative study should provide insights into the effect of feed characteristics on the process effectiveness.