Preparation and characterization of novel positively charged copolymer composite membranes for nanofiltration

Abstract

A novel copolymer poly(ether ether ketone) bearing pendant tertiary amine groups (TAPEEK-x) was synthesized by 3,3′-dimethylaminemethylene-4,4′-biphenol (DABP), 4,4′-bisfluorodiphenylketone (BFDPA) and bisphenol A (BPA). After reaction with iodomethane, the copolymer turned to poly(ether ether ketone) with pendant quaternary ammonium groups (QAPEEK-x). Afterward, QAPEEK-x copolymer was dissolved in formic acid and coated on PAN ultrafiltration membrane as a composite membrane. The NF membrane preparation conditions, such as drying temperature, formic acid concentration, copolymer concentration, feed concentration, and tolerance in chlorine solution were investigated. The experiments showed that the membrane whose coating polymer concentration was 0.5 wt% in 88% formic acid had the best flux of 12.6 L m⁻² h⁻¹ at 0.4 MPa and best rejection of 99.4% for 500 ppm MgCl₂. In addition, the membrane showed excellent tolerance ability of chlorine with a little decrease of MgCl₂ rejection. To some extent, this study provides a new insight to improve membrane separation capability and stability in the desalination process.

---

Introduction

Nanofiltration (NF) is a relatively pressure-driven and low energy consumption separation technology that has developed rapidly in the last decade.¹ Generally, nanofiltration membranes separation ability is in the intermediate range between ultrafiltration and reverse osmosis, whose pores size are 0.5–2.0 nm in diameter and nominal molecular weight cut off (MWCO) values are 200 to 1000 Da.² The main features of nanofiltration membranes include its high water flux at low operating pressure and high rejection of small organics or multivalent ions, which to some degree greatly reduce operating energy consumption, capital and costs. Due to those attracting advantages, there is undoubtedly that nanofiltration will become a promising and potential technology that can be applied in many fields such as water softening,^3–5 retention of dyes,^6–8 and waste water reclamation.⁹

According to membrane surface charge, nanofiltration membranes can be positively charged, negatively charged and neutral.¹ But their common ground lies in its active layer that is used for separation depending on steric hindrance effect¹⁰ or Donnan exclusion theory¹¹ and is the core content of all the composite membranes. Up to now, most NF membranes are usually produced by polyamide,¹² sulfonated polysulfone¹³ and sulfonated poly(ether ether ketone),¹⁴ but they are mostly negatively charged and just have good rejection capacity for some anions like SO₄²⁻.¹⁵ From this point of view, developing a new type nanofiltration membrane with high rejection to multivalent cations like Mg²⁺, Ca²⁺ is especially important.¹⁶ Positively charged NF membranes have been developed in recent years due to its high rejections for multivalent cations.¹⁷ For example, Deng and Xu prepared a high flux positively charged nanofiltration membranes by UV-initiated graft polymerization of methacrylatoethyl trimethylammonium chloride onto polysulfone membranes.¹⁸ The rejection of MgCl₂ is 94.8% and the water flux is 20.3 L m⁻² h⁻¹ at 0.2 MPa. Moreover, Gao et al. prepared a positively charged nanofiltration membrane with a flux of 20.6 L m⁻² h⁻¹ through the crosslinking of copolymer poly(2-methacryloyloxy ethyl trimethylammonium chloride-co-2-hydroxyethyl acrylate) and glutaraldehyde.¹ Obviously, those experimental results sufficiently verify the capacity and validity of nanofiltration membranes. However, till now, to achieve the high performance of water flux and salt rejection in meanwhile is relatively difficult.

Poly(arylene ether ketone) bearing pendant tertiary amine groups (TAPEEK-x) was synthesized in our previous work.¹⁹ The NF membrane displayed the high water flux but mediate rejection for MgCl₂. Therefore, in present work, we synthesize a new copolymer and explored better coating conditions to improve the composite membranes rejection capacity. Firstly, we choose bisphenol A as a comonomer whose copolymer chemical structure is shown in Fig. 1. Furthermore, we used a copolymer with low ion exchange capacity as active layer to insure high salts rejection¹⁸ and increase the coating time to insure the active layer flawless. Under such condition, the influence of drying temperature, polymer concentration, feed concentration, concentration of formic acid and tolerance in chlorine solution were investigated. Through all kinds of studies on coating conditions, we aimed to obtain a new type of nanofiltration membrane with good performance in water flux and rejection.

Fig. 1 Chemical structure of copolymer TAPEEK-x and QAPEEK-x.

Experimental part

Materials

PAN ultrafiltration membrane (MWCO = 50 000 Da) was purchased from Beijing Separate Equipment Co. Ltd. 3,3′-Dimethylaminemethylene-4,4′-biphenol (DABP) was synthesized as we reported before.²⁰ 4,4′-Bisfluorodiphenylketone (BFDPA) were purchased from Aldrich. Bisphenol A (BPA) were purchased from Aldrich. Polyethylene glycol (PEG), formic acid whose water content was 2% and N,N-dimethylacetamide (DMAc) were purchased from Sinopharm Group Chemical Reagent. All other reagents were obtained from commercial sources and used as received.

Synthesis of copolymer TAPEEK-x and QAPEEK-x

We synthesized all kinds of copolymer with different number of tertiary amine groups by controlling the molar ratio of DABP and BPA. TAPEEK-x where x represents the feed percent of DABP was synthesized and the concrete procedure was shown in Fig. 1. DABP (18.01 g, 60 mmol), BPA (9.13 g, 40 mmol), BFDPA (21.82 g, 100 mmol), Cs₂CO₃ (65.1 g, 200 mmol), DMAC (240 mL) and cyclohexane (100 mL) were added to a flamed-dried 500 mL three-necked flask in condition of a nitrogen and mechanical stirrer and a Dean–Stark trap. The reaction bath was heated to 120 °C for 10 h until water was removed from mixture by azeotropic distillation. After cyclohexane and water had been distilled off, the temperature was raised gradually to 130 °C to react at this temperature for 30 h to obtain a viscous solution. After cooling to room temperature, the solution was filtered and poured into 80 °C stirred deionized water to remove salt and solvent. The fiber-like precipitate was washed with hot water and ethanol five times prior to being dried under vacuum at 100 °C for 24 h to produce the final product.

QAPEEK-60 was prepared as follows. 5 g TAPEEK-60 and 10 mL DMAC were added into flask, then 3.78 g iodomethane was added and the mixture was stirred at 30 °C in the dark for 12 h. The mixture was poured into ethanol and washed with deionized water and ethanol several times. At last, the polymer was dried in vacuum oven for 12 h at 80 °C.

Fabrication of composite membrane

The active skin layer of the composite membrane was prepared by coating QAPEEK-x on the porous PAN ultrafiltration membrane. Firstly, PAN ultrafiltration membrane were immersed into 10% glycerol–H₂O solution for 4 min to insure that the support membrane pores can be occupied by glycerol molecules. The ultrafiltration membrane were wiped with tissue paper and mounted on a clean glass plates. Subsequently, different concentrations solution in formic acid was poured on PAN ultrafiltration membrane. After 5 minutes, the excess solution was drained off by keeping the membrane vertically. Eventually, the membranes were dried in oven for 10 min at 70 °C.

Characterizations

¹H NMR spectra were measured at 300 MHz on an AV300 spectrometer. The mechanical properties of membranes were measured by means of tensile test in dry state at room temperature, 30% humidity using an Instron-1211 mechanical testing instrument at a speed of 5 mm min⁻¹. The intrinsic viscosities were carried out by an Ubbelohde capillary viscometer in DMAC at 30 ± 0.1 °C with a polymer concentration of 0.5 g dL⁻¹. Water contact angle were measured on a Dataphysics OCA 20 contact-angle system at ambient temperature.

The membrane morphology

The membrane morphology was characterized by a scanning electron microscope (SEM, XL 30 ESEM FEG, FEI Company). In order to obtain a clear pore structure of the cross-sections, the composite membrane were freeze-dried in freeze dryer and broke the membranes in liquid nitrogen.

Atomic force microscopy (AFM) images were prepared by a SPI 3800/SPA 300HV (Seiko Instruments Inc., Japan) with tapping mode. Approximately, 1 cm² composite membrane samples were fixed on a specimen holder and 8 μm × 8 μm of the membrane were scanned through a commercially available SiN₄-cantilever with a spring constant of 2 N m⁻¹.

The water flux and solute rejection

The water flux was measured by a cross-flow filtration apparatus. The flux was calculated by equationwhere F is the water flux (L m⁻² h⁻¹) V is the volume of pure water (L), A is the effective membrane area (cm²), t is the operation time (h). The solute rejection of the NF membrane was calculated as follows:where C_p and C_f are the solute concentrations in the permeate and feed solution. Salt concentration was measured by electrical conductivity (Elmetron conductivity meter CC-501, Poland). The composite NF membrane were characterized by water flux and rejection. The effective membranes area was 23.7 cm². The membranes were carried out with pure water at 0.5 MPa for 4 h and the salt solution was evaluated at 0.4 MPa in the feed temperature at 25 °C. The effect of pressures ranging from 0.4 to 0.8 MPa, drying temperature, polymer concentration, water content of solvent and feed concentration was tested by 500 ppm MgCl₂.

The membrane molecular weight cut off (MWCO)

The membrane molecular weight cut off (MWCO) was carried out by filtration of different molecular weight polyethylene glycol (PEG) solutions (MW 200–1000 Da) at same concentration. The concentrations of PEG in feed and permeation were measured through a total organic carbon analyzer (TOC-VCPH SHIMADZU). All the measurements were based on three samples and adopt it average values.

The chlorine tolerance test

The composite membranes were exposed to sodium hypochlorite (NaOCl) solution with concentration of 500 ppm for an hour so as to keep consistent with PI composite membrane with same testing condition.²¹ Likewise, all the membranes were tested in 0.4 MPa at 25 °C. Water flux and rejection was measured including before and after the NaOCl treatment to evaluate the chlorine tolerance of the membranes.

Results and discussion

Synthesis of copolymer TAPEEK-x and QAPEEK-x

The TAPEEK-60 copolymer was synthesized by 3,3′-dimethylaminemethylene-4,4′-biphenol (DABP), bisphenol A (BPA) and 4,4′-bisfluorobenzophenone as indicated in Fig. 1. The intrinsic viscosity was measured by an Ubbelohde capillary viscometer is 0.42 dL g⁻¹ as show in Table 1. Its chemical structure was confirmed by ¹H NMR spectroscopy with CDCl₃ as the solvent. Fig. 2 (a) shows the spectra of the protons for the TAPEEK-60. The peaks at δ 3.5 ppm and δ 2.3 ppm are the chemical shifts of protons on methylene and methyl groups, respectively. The area ratio of H in methylene and methyl groups is close to 1 : 3, just meeting the composition of the tertiary amine groups on TAPEEK-60.

Table 1 Viscosity, contact angle and mechanical properties of TAPEEK-60 and QAPEEK-60

Sample	η (dL g^−1)	Water contact angle (°)	Tensile strength (MPa)	Tensile modulus (MPa)	Elongation at break (%)
TAPEEK-60	0.42	55.68	58.9	1200	8.1
QAPEEK-60	1.78	51.70	28.8	460	8.6

---

Fig. 2 The 1H-NMR spectra of polymers (a) TAPEEK-60 and (b) QAPEEK-60.

QAPEEK-60 can be obtained by reaction of TAPEEK-60 and iodomethane. As shown in Fig. 2(b), the mole ratio of H in ¹H NMR spectroscopy between methylene and methyl groups becomes 2 : 9. Therefore, the TAPEEK-60 copolymer can be completely converted to QAPEEK-60. Judging from Table 1, through quarter of amination reaction, the copolymer features produced some changes. For example, copolymer viscosity turns to 1.78 dL g⁻¹ due to the strong electric charge effect. What's more, the hydrophilicity was improved, with contact angle changing from 55.68° in TAPEEK-60 membrane to 51.70° in QAPEEK-60 membrane. In same condition, the composite membrane state with 0.5 wt% copolymer concentration has a lower the contact angle of 48.5° due to the different membrane morphology.

Morphology studies of composite membrane

Fig. 3 shows the cross-section and surface SEM images of different concentrations from 0.2 wt% to 0.5 wt%. All the surface of the composite membrane was smooth and dense, and there are no visible pores and defects. In the cross-section images, the active layer and the support layer with finger-like macrovoids were easy to distinguish. With the increase of polymer concentration from 0.2 wt% to 0.5 wt%, the thickness of active layer becomes from 200 nm to 900 nm.

Fig. 3 Cross-section and surface SEM images of different concentrations.

Fig. 4 shows the three-dimensional AFM images of QAPEEK-60 composite membrane with different concentration. With the increase of polymer concentration from 0.2 wt% to 0.5 wt% and the thickness of active layer from 200 nm to 900 nm, the root mean square roughness (RMS) of composite membrane decreases from 8.6 nm to 4.5 nm. Although there is little variation in root mean square roughness, its reveals a smoother trend with the increase of copolymer concentration.²²

Fig. 4 AFM images of composite membrane with different concentrations.

Molecular weight cut off (MWCO) and pore size

Fig. 5 shows the retention curves of QAPEEK-60 composite membrane for PEG with molecular weight from 200 to 1000 Da. Herein, the composite membrane was coated with 0.5 wt% copolymer and in the curing temperature of 70 °C for 10 minutes. The molecular weight cut off is referred to the molecular weight of PEG when rejection rate turns to 90%. The pore sizes can be estimated through the relationship between Strokes radii and molecular weight cut off. Just as shown in Fig. 5, the MWCO of QAPEEK-60 composite membrane is 700 Da and Strokes radii is 0.64 nm,²³ which meets the range of nanofiltration.²⁴

Fig. 5 PEG retention of composite membrane for 1000 ppm PEG solutions at 0.4 MPa.

Nanofiltration membrane testing performance

Effect of drying temperature on membrane performance. As shown in Fig. 6, in same preparation condition above, the experimental result suggests that the performance of composite membrane changed with drying temperature. With the increase of temperature from 50 °C to 80 °C, the rejection of MgCl₂ was increased from 85.1% to 98.7%, while water flux decreased from 28.3 L m⁻² h⁻¹ to 6.75 L m⁻² h⁻¹. It can be explained that a high curing temperature promotes the volatilization of solvent and goes a further to form a more compact and dense active layer.

Fig. 6 Effect of drying temperature on membrane performance.

Effect of QAPEEK-60 concentration. In the preparation nanofiltration membrane, the supporting membrane was coated with different concentration of QAPEEK-60 solutions in the curing temperature of 70 °C for 10 minutes. Fig. 7 shows the effect of concentration on the rejection of MgCl₂ and the water flux. With the increase of QAPEEK-60 concentration from 0.2 wt% to 0.5 wt%, just as we expected, the rejection of MgCl₂ increased from 92.9% to 99.4%, which presented a rise trend. On the contrary, the water flux begins to decrease gradually with a flux from 17.7 L m⁻² h⁻¹ to 12.6 L m⁻² h⁻¹. This is because the increase of polymer concentration can result in a thicker coating layer and go a further to add water molecular hindrance. Therefore, the water flux was cut down correspondingly, which was completely shown in the morphology of different concentrations membrane cross-section (Fig. 3) whose active layer thick increases from 200 to 900 nm. Thus, with the increase of coating layer thick, the membrane surface becomes more and more dense and smooth, and according to Donnan exclusion theory, Mg²⁺ can be obstructed opposite of the thin-active layer due to its strong electrostatic repulsive interaction with positively charged membrane. However, in order to keep charge balance, Cl⁻ is also obstructed.

Fig. 7 Effect of copolymer concentration to water flux and rejection.

Effect of formic acid concentration and separation performance in different salts. In our experimental explore, we found that the concentration of formic acid can lead to some differences in flux and rejection. As shown in Fig. 8, the membrane whose coating polymer concentration was 0.5 wt% in 58% formic acid has a higher flux than membrane with same polymer concentration in 82% formic acid, while its rejection to some degree decreases a little. The main reason for this phenomenon lies in that formic acid can corrode the pores of PAN membrane and to some extent lead to supporting membrane pore collapse and shrink. Therefore, the concentration of formic acid is a important aspect to control water flux and rejection.

Fig. 8 Effect of water content of solvent and separation performance to different salts.

Fig. 8 shows its separation ability and water penetration ability for different inorganic salts. It can be seen that QAPEEK-60 composite membrane salts rejection order is MgCl₂ > CaCl₂ > NaCl > Na₂SO₄,^25–27 while its water flux for four kinds salt solutions has few changes. Therefore, this QAPEEK-60 positively charged nanofiltration membrane observes steric hindrance effect and electrostatic repulsive effect.^28,29

Effect of feed concentration and operating pressure. Fig. 9(a) shows the effect of feed concentration of MgCl₂ solution on nanofiltration performance. With the increase of MgCl₂ concentration from 0.5 g L⁻¹ to 3.0 g L⁻¹, QAPEEK-60 composite membrane rejection goes down steadily from 97.8% to 89.3%, which is mainly because more anions combine with positively charged membrane due to electrostatic interaction and thus go a further to decline composite membrane charge density. Meanwhile, as to water flux, there exists a polarization layer on the surface of nanofiltration membrane, resulting in the decrease of water flux slightly.^1,18,30

Fig. 9 Effect of feed concentration (a) and operating pressure (b) on water flux and salt rejection.

Fig. 9(b) shows the effect of operating pressure on QAPEEK-60 composite membrane. It can be seen from the figure that water flux increases almost linearly with the rise of operating pressure, while there is not apparent influence on salt rejection. Generally, according to solution-diffusion model,^31,32 the increase of pressure can produce a higher rejection. However, seeing from the result, this QAPEEK-60 composite membrane cannot obey this rule very much. That is because that we need to consider surface forces which can retain ions partially and lead to a long residence time of ion in the pore. When operation pressure goes up, the drag force increases because of the flux in the pore. However, considering its compact and dense membrane surface, MgCl₂ rejection just has slightly decline.¹⁷

Chlorine tolerance of the QAPEEK-60 NF membrane. As we all known, although PA membrane has relatively high flux and salts rejection, it cannot tolerate continuous exposure to water with only a few parts per million chlorine. However, QAPEEK-60 nanofiltration membrane doesn't have this amido bond, and to some extent, this composite membrane can endure sodium hypochlorite. In order to verify its tolerance ability of chlorine, we put a composite membrane with 0.5 wt% polymer concentration into 500 ppm NaClO solution for an hour to compare with common PA composite membrane. As shown in Table 2, through our testing, there was no obvious decrease of MgCl₂ salt rejection and water flux while PA membranes flux and rejection changed sharply. By comparison, this result suggests that QAPEEK-60 nanofiltration membrane do present an excellent tolerance ability of chlorine.

Table 2 Chlorine tolerance of the QAPEEK-60 NF membrane and PA composite membranes

	Before an hour to deal with		After an hour to deal with
Membranes	Flux (L m^−2 h^−1)	Rejection (%)	Flux (L m^−2 h^−1)	Rejection (%)
a Obtained from literature.^20
PA^a	22.1	98.0	48.6	67.1
NF (QAPEEK-60)	12.6	98.1	12.7	97.8

---

Conclusion

In conclusion, in order to obtain a nanofiltration membrane with high rejection and water flux, we used BPA as a monomer to copolymerize with DABP and BFDPA and obtained QAPEEK-x copolymer by quaternisation reaction. Afterwards, we chose QAPEEK-60 with low ion exchange capacity as coating polymer and improved coating condition to insure membrane active layer flawless. Through a series of experiments including drying temperature, polymer concentration, feed concentration, the concentration of formic acid, we found that the membrane with 0.5 wt% QAPEEK-60 in 88% formic acid had the best performance in nanofiltration test. The results exhibited the best rejection of 99.4% for MgCl₂ and high rejection for NaCl solution with a rejection above 75%. In addition, in the membrane chlorine-tolerance studies, QAPEEK-60 nanofiltration also showed excellent tolerance ability of chlorine with a little decrease of MgCl₂ salt rejection. Overall, QAPEEK-60 positively charged membranes had good separation performance and indicated its excellent effectiveness in desalination process.

Acknowledgements

This research was financially supported by the National Basic Research Program of China (no. 2012CB932802), the National Science Foundation of China (no. 51133008, 51203151, 21074133), the National High Technology Research and Development Program of China (no. 2012AA03A601), and the Development of Scientific and Technological Project of the Jilin Province (no. 20130204027GX).

Notes and references

1. Y. L. Ji, Q. F. An, Q. Zhao, H. L. Chen and C. J. Gao, J. Membr. Sci., 2011, 376, 254.
2. S. P. Sun, T. A. Hatton, S. Y. Chan and T. S. Chung, J. Membr. Sci., 2012, 401, 152.
3. S. Mondal and S. R. Wickramasinghe, J. Membr. Sci., 2008, 322, 162.
4. J. Schaep, B. Van der Bruggen, S. Uytterhoeven, R. Croux, C. Vandecasteele, D. Wilms, E. Van Houtte and F. Vanlergberghe, Desalination, 1998, 119, 295.
5. D. Nanda, K. L. Tung, Y.-L. Li, N. J. Lin and C. J. Chuang, J. Membr. Sci., 2010, 349, 411.
6. M. Riera-Torres, C. Gutiérrez-Bouzán and M. Crespi, Desalination, 2010, 252, 53.
7. I. Koyuncu, Desalination, 2002, 143, 243.
8. P. Zhong, N. Widjojo, T. Chung, M. Weber and C. Maletzko, J. Membr. Sci., 2012, 417, 52.
9. B. Van der Bruggen, C. Vandecasteele, T. V. Gestel, W. Doyenb and R. Leysen, Environ. Prog., 2003, 22, 46.
10. B. Van der Bruggen, J. Schaep, D. Wilms and C. Vandecasteele, J. Membr. Sci., 1999, 156, 29.
11. R. J. Peterson, J. Membr. Sci., 1993, 83, 81.
12. H. B. Park, B. D. Freeman, Z. B. Zhang, M. T. Sankir and J. E. M. Grath, Angew. Chem., 2008, 120, 6108.
13. H. M. Colquhoun, D. Chappell, A. L. Lewis, D. F. Lewis, G. T. Finlan and P. J. Williams, J. Mater. Chem., 2010, 20, 4629.
14. M. Dalwani, G. Bargeman, S. S. Hosseiny, M. Boerrigter, M. Wessling, N. E. Wessling and N. E. Benes, J. Membr. Sci., 2011, 381, 81.
15. J. Isaad, A. El Achari and F. Malek, Dyes Pigm., 2013, 97, 134.
16. H. F. Wang, Q. F. Zhang and S. B. Zhang, J. Membr. Sci., 2011, 378, 243.
17. S. Y. Cheng, D. L. Oatley, P. M. Williams and C. J. Wright, Adv. Colloid Interface Sci., 2011, 164, 12.
18. H. Y. Deng, Y. Y. Xu, Q. C. Chen, X. Z. Wei and B. K. Zhu, J. Membr. Sci., 2011, 366, 363.
19. J. Wang, J. L. Wang, H. F. Wang and S. B. Zhang, J. Appl. Polym. Sci., 2013, 127, 1601.
20. J. H. Wang, J. Wang, S. H. Li and S. B. Zhang, J. Membr. Sci., 2011, 368, 246.
21. L. Shao, X. Q. Cheng, Y. Liu, S. Quan, J. Ma, S. Z. Zhao and K. Y. Wang, J. Membr. Sci., 2013, 430, 96.
22. W. Richard Bowen, T. A. Doneva and H. B. Yin, J. Membr. Sci., 2001, 181, 253.
23. Q. F. Zhang, H. F. Wang, S. B. Zhang and L. Dai, J. Membr. Sci., 2011, 375, 191.
24. J. Schaep and C. Vandecasteele, J. Membr. Sci., 2001, 188, 129.
25. J. Schaep, B. Van der Bruggen, C. Vandecasteele and D. Wilms, Sep. Purif. Technol., 1998, 14, 155.
26. L. C. Li, B. G. Wang, H. M. Tan, T. L. Chen and J. P. Xu, J. Membr. Sci., 2006, 269, 84.
27. B. Van der Bruggen, J. Schaep, D. Wilms and C. Vandecasteele, J. Membr. Sci., 1999, 156, 29.
28. Y. L. Ji, Q. F. An, Q. Zhao, W. D. Sun, K. R. Lee, H. L. Chen and C. J. Gao, J. Membr. Sci., 2012, 390, 243.
29. G. T. Ballet, A. Hafiane and M. Dhahbi, J. Membr. Sci., 2007, 290, 164.
30. H. Sun, G. Chen, R. Huang and C. Gao, J. Membr. Sci., 2007, 297, 51.
31. C. C. Feng, J. Xun, M. M. Li, Y. Y. Tang and C. J. Gao, J. Membr. Sci., 2014, 451, 103.
32. S. G. Kim, J. H. Chun, B. H. Chun and S. H. Kim, Desalination, 2013, 325, 76.

---