Cellulose Acetate-Based Membranes Recovered from Black-and-White Cinematographic Films for the Simultaneous Removal of Nitrate and Phosphate Anions from Water by Nanofiltration

1. Introduction

Membranes and membrane processes involving cellulose derivatives as membrane materials are today a common presence in technologies for obtaining drinking water and treating aqueous effluents [1,2,3].

The cellulose (C) (Figure 1a) is a membrane material with excellent properties [4], but one of its derivatives, cellulose acetate (Figure 1b), still finds wide use today [5,6,7,8].

The success of these raw materials in various technological achievements has led to an unprecedented development of research to obtain membranes with high selectivity, increasingly higher productivity and long life [9,10].

In the last years the cellulose acetate has once again become the main actor in research to obtain membranes with industrial but also domestic or small-scale applications [11,12,13].

The disadvantages of cellulose acetate, compared with synthetic polymers, have become advantages in the era of clean and cleaning technologies:

• the selectivity of the artificial polymer and the productivity of the membranes have been largely improved by the contribution of nanomaterials: allotropic forms of carbon, metallic, oxide or composite nanoparticles [14,15,16,17,18,19].
• lower resistance to high pH allows recycling of used membranes by dissolution in strongly alkaline solutions, filtration followed by precipitation in water or aqueous solutions.
• biodegradability gives it increased attractiveness in the case of waste, which is no longer an environmental problem.

An aspect that initially did not constitute a problem, the accessibility and low price of cellulose acetate, has become in recent years an issue discussed by ecologists because the enormous quantities of cellulose acetate used worldwide have created the justified feeling that forests are disappearing also due to the production of membranes based on cellulose derivatives.

As expected, the researchers’ solutions came quickly and recently membranes obtained from newspaper waste [20], photographic and cinematographic films [21], or cigarette filters [22] have been promoted, and the ideas for recycling cellulose acetate seem to be in a continuous progress [23,24].

Obtaining drinking water in isolated inhabited areas involves the removal of inorganic micro pollutants from shallow groundwater: nitrate anion or/and phosphate anions.

While the phosphate anion is most likely to be found in the waters from these areas, the nitrate anion is also undesirable, which often occurs either from agricultural fertilizers or from acid rains [25,26].

The EU regulations are restrictive for nitrate anions because they are aggressive, especially for young children. The European regulations regarding the phosphate ion focus on reducing water eutrophication and come from several regulatory acts, not a single directive [25,26].

For both the nitrate and phosphate anions, a series of chemical or biological removal methods have been developed [27,28,29]. At the same time, membrane removal methods from water have been studied for these anions [30,31,32,33].

It should be emphasized that studies on many membrane methods have been developed for aqueous solutions containing either the nitrate anion or the phosphate anion. [34,35,36].

For the nitrate anion, nanofiltration processes through various types of membranes are well developed, some reaching the commercial stage [37,38,39,40].

This paper presents a method for the simultaneous removal of the proposed ions (nitrate and phosphate) by membranes. The developed method uses nanofiltration through cellulose acetate (CA)–silver nanoparticles (Agnp) composite membranes, obtained from cinematographic films.

The composite membrane (CA–Agnp–PP) was obtained by phase inversion, by contacting on a tubular propylene membrane (PP) the dispersion in DMF of the cellulose acetate (CA)–silver nanoparticles (Agnp) assembly.

2. Materials, Reagents and Methods

2.1. Materials and Reagents

The reagents used were the following: NaNO_3, and AgNO₃, are from Merck KGaA, Darmstadt, Germany, Anhydrous disodium phosphate (99%, Sigma Aldrich, Burlington, MA, USA), Dimethylformamide (DMF) (99%, Honeywell, Charlotte, NC, USA) hydrochloric acid (37%, Sigma Aldrich, Burlington, MA, USA) and sodium hydroxide (Merck, Rahway, NJ, USA).

The cellulose acetate used was waste from the film industry (containing silver nanoparticles) [21].

The polypropylene hollow fiber membrane support (PP) was provided by GOST (GOST Ltd., Perugia, Italy) and their characteristics and performance were previously presented in detail [41].

Rapid analytical tests performed using LCK349 Phosphorus total /Phosphate ortho and LCK339 Nitrate (Hach Lange GmbH, Düsseldorf, Germany) are specific to the analysis of test components in synthetic solutions prepared in the laboratory (phosphates, nitrates).

The purified water, characterized by 18.2 µS/cm conductivity, was obtained using a RO Millipore system (MilliQ^R Direct 8 RO Water Purification System, Merck, Darmstadt, Germany) [21].

2.2. Methods and Procedures

2.2.1. Obtaining Composite Membranes

The source of cellulose acetate containing silver nanoparticles was given by cinematographic films (exposed or unexposed to light). Cinematographic films were selected from eleven samples considered representative as raw material. To obtain a 2% cellulose acetate solution, 20 grams of photographic film, washed, dried and ground to submicron sizes in a colloidal ball mill (Retsch 100), were dissolved in dimethylformamide in 980 g of DMF when a gray solution (CA–Agnp–DMF) was obtained.

Solutions of 4% and 6% cellulose acetate in DMF were also obtained in the same way.

At the same time, concentrations of 7.5%, 11.5% and 16.1% Agnp in casting solution were obtained by using films from different batches. This parameter is random, studied, but not dependent on experimenter’s desire.

To create CA–Agnp–PP composite membranes, the phase inversion of the dispersed system (CA–Agnp–PP) was performed on and in the polypropylene (PP) hollow fibers of a nanofiltration module and the contact with deionized water circulating through fibers (Figure 2).

The CA–Agnp–PP composite membrane is obtained, which is thus prepared for nanofiltration of solutions containing sodium nitrate and sodium phosphate.

2.2.2. Nanofiltration of Phosphate and Sodium Nitrate Solutions

The aqueous solution subjected to nanofiltration was synthetically obtained by dissolving sodium phosphate and/or sodium nitrate in deionized water. The concentration of each species is 10ppm, 30ppm, or 50ppm.

Nanofiltration is performed by the module in Figure 3 in which the membranes were previously washed with deionized water. Complete washing was confirmed by checking, in the permeate, the removal of bromide ion (remaining from unexposed cinematographic films) with silver nitrate solution.

The result of the nanofiltration was presented in the form of the permeate flux (J) (equation 1) and the retention, or of the nanofiltration efficiency (R) (equation 2) [21,43]:

$$J={\displaystyle \frac{V}{S\bullet \Delta t}}\left(L/m^2·h\right),$$

(1)

where: J – permeate flux; V – permeate volume; S – membrane surface; Δt – operating interval.

$$\mathit{R}\mathrm{\%}={\displaystyle \frac{\left(c_0-c_f\right)}{c_0}}\bullet 100$$

(2)

where: R – retention; c₀ – concentration of feed solution; c_f – final concentration.

Analysis of nitrate and nitrate anions was performed with rapid tests LCK349 Phosphorus total /Phosphate ortho and LCK339 Nitrate (Hach Lange GmbH, Düsseldorf, Germany).

2.3. Equipment

The microscopy studies, SEM and HR-SEM, were performed on a Hitachi S4500 system (Hitachi High—Technologies Europe GmbH, Mannheim, Germany) [21].

For the comparative study of the scanning electron microscopy (SEM) and energy-dispersive spectrum for the characteristic X–Ray (EDAX) analyses, the membrane samples, subjected to the analysis, were visualized with the help of the FESEM–FIB workstation (scanning electron microscope with field emission electron and focused beam of ions), model Auriga (Carl Zeiss SMT, Oberkochen, Germany) [44,45].

The determination of the concentration of nitrate or phosphate ions was performed using a CamSpec M550 Spectrophotometer (Spectronic CamSpec Ltd., Leeds, UK) [46].

An atomic absorption spectrometer AAnalyst 400 AA spectrometer (PerkinElmer Inc., Shelton, CT, USA) with WinLab32—AA software (PerkinElmer), with a single-element hollow-cathode lamp was used for silver concentration. The experimental parameters were 328.1 nm wavelength and 0.7 nm spectral bandwidth at an operating current of 5 mA [47].

3. Results and Discussion

The black-and-white cinematographic films are now history and have become waste for two reasons:

• Physical-chemical wear of the cellulose acetate substrate during archiving (Figure 4);
• Translating information from motion pictures onto electronic media.

This waste is difficult to biodegrade due to the content of silver nanoparticles [21] (Figure 4). It becomes necessary for physically and/or morally worn black-and-white cinematographic films to be recycled. A bold idea would be to recover the silver and then reuse the cellulose acetate. This idea is not feasible because silver is in too small quantity to be technically and economically viable.

Starting from the results achieved in our group on the recycling of cinematographic films as membranes for the removal of hydrogen sulfide and organic sulfur compounds [21], this time the goal was to obtain composite nanofiltration membranes based on cellulose acetate and silver deposited on polypropylene capillary membranes (hollow fiber) as a support, and to test them for the removal of phosphate and nitrate ions found simultaneously in dilute aqueous solution.

The work on nanofiltration of solutions containing phosphate and nitrate anions through silver nanoparticle-cellulose acetate-polypropylene hollow fibers composite membranes was carried out as follows:

• Obtaining and composition and morphological characterization of the membranes;
• Determination of process characteristics using pure water;
• Determination of retention and fluxes in nanofiltration of phosphate and nitrate solutions.

3.1. Morphological and Compositional Characterization of the Obtained Membranes

To obtain the composite membranes, a polypropylene support in the form of hollow fiber (Figure 5a) was used, which has a porosity specific to membrane contactors and can be easily assembled into a module (Figure 5b) that we chose so as to ensure a filtering surface of 1m².

The morphological appearance of polypropylene support membranes shows an outer fiber diameter of about 330 µm (Figure 6a) and an inner diameter or approximately 300 µm (Figure 6b). The specific porosity of the ultrafiltration (or membrane contactors) of the polypropylene hollow fiber membrane is presented in Figure 7a and Figure 7b.

The porosity of the hollow fiber polypropylene membranes indicate that it can be used for pertraction or in membrane contactors [47,48,49,50] (Figure 7a and Figure 7b).

Contacting the dispersion formed by cellulose acetate and silver nanoparticles through the support membranes with deionized water led to phase inversion, obtaining membranes with the morphology presented in Figure 8a, in which the deposition of the nanofiltration layer is observed (Figure 8b). The orange arrows highlight the deposited layer of cellulose acetate and silver nanoparticles.

The EDX analysis of the membranes aimed at the composition of the selective (surface) layer, but also of the polypropylene substrate. For the characterization of the membrane, the section in which the superficial nanofiltration layer is also highlighted was chosen (Figure 9a). The cross-sectional image shows the areas that were subjected to EDX analysis, as follows: (a) – the surface of the superficial layer; (b) – the surface of the propylene substrate section, and (c) – the surface of the superficial layer (the exfoliated part).

The analyses performed reveal that the surface of the superficial layer and the surface of the section of the superficial layer (the exfoliated part) have in composition (Table 1): silver, carbon, oxygen and sulfur, but also impurities, at trace levels, of sodium and silicon.

The surface (A) of the superficial layer is composed predominantly of silver, carbon and oxygen, which are the expected components of cellulose acetate coming from cinematographic films. The presence of sulfur is most likely related to bisulfite developing solutions, which lead to residual sulfur (most likely as silver sulfur). The presence of sodium is not unusual because it is a universal contaminant whose traces can be quite difficult to remove.

The section of the polypropylene support (B) indicates the presence of cellulose acetate in the pores of the support membrane, because along with carbon we also have an appreciable concentration of oxygen. It is worth noting that nanometric silver does not penetrate the pores of the polypropylene hollow fiber membrane, which is demonstrated by its absence in the spectrum, but also by the lack of sulfur that accompanies in the other two analyses.

The section of the superficial layer (exfoliated part) (C) has a qualitative composition similar to that of the surface of this layer but slightly modified quantitatively. It should be noted, however, the absence of sodium (most likely being eluted during the preparation of the membranes). The presence of silicon is accidental, most likely coming from the membrane cutting device during the preparation of the samples for scanning electron microscopy.

Following the EDX analysis of the cellulose acetate–silver nanoparticles–polypropylene (CA–Agnp–PP) composite membrane, the following can be stated:

• The composite membrane has a superficial layer composed of cellulose acetate and silver nanoparticles;
• The high concentration of silver shows that it is retained and distributed on the membrane surface during preparation;
• The superficial concentration of silver is very high compared to that in the solution obtained from cinematographic films, about 9ppm compared to cellulose acetate;
• The membrane support (polypropylene) also contains cellulose acetate in its pores, highlighted by the presence of oxygen;
• Silver nanoparticles do not penetrate the pores of the polypropylene support;
• The presence of residual sulfur from the cinematographic film processing process was found;
• The presence of trace impurities (sodium and silicon) is related to the sampling of the samples.

3.2. Determination of Nanofiltration Characteristics Through CA–Agnp–PP Membrane Using Deionized Water

The determination of permeate flows as a function of the working pressure (Figure 10) was carried out at a flow rate of 120L/h of deionized water achieved with the recirculation pump (RP) (Figure 3). The working pressure is achieved using an air cylinder connected to the R1 tank, thus completing the supply water circuit. The nano-filtered water flow (permeate) was determined in the range of 5 to 25 atm. In Figure 10, we can see a significant increase int the flow in the range of 5–10 atm, after which the slope of the curve decreases. Increasing the pressure above 15atm is not desirable because it reduces the technical and economic interest. Implementing a process at high pressures raises technical problems that are not compensated by the increase in permeate flow.

The determination of permeate fluxes as a function of the flow regime was performed at a pressure of 15 atm (Figure 11). The flow regime was varied by changing the flow rate (Q) of feed water through the module fibers. The increase of permeate flux occurs differently in the range of 60–120 L/h when the slope is high, compared to the range 120–270 L/h when the slope of the curve is significantly reduced.

It can be said that in the first part of the flow interval, the flow transitions from laminar flow to turbulent flow. In the second part of the curve, the flow increase is much reduced because the turbulent regime is maintained within the same parameters.

3.3. Determination of Permeate Flow and the Retention for CA–Agnp–PP Composite Membranes

To determine the performances of the prepared membranes, two work directions were approached:

• varying the characteristics of the casting solution from which the membranes were prepared;
• changing the feed solution parameters.

The constant parameters in the study were the working pressure, 15 atm, and the feed solution flow rate, 120 L·m⁻².

Table 2 summarizes the results regarding permeate flux, nitrate anion and phosphate anion retention depending on the concentration of cellulose acetate (C_CA) and, respectively, silver particles (C_Agnp), in the casting solution.

The variation of cellulose acetate concentration to 2%, 4% and 6% (Table 2) led to a decrease in the flux value. Thus, from 15.2 L·m⁻²·h⁻¹ for the 2% CA membrane and 10.7 L·m⁻²·h⁻¹ for the 4%CA, it reached 4.4 L·m⁻²·h⁻¹ for the 6% CA membrane. The significant decrease in the flux value with increasing polymer concentration in the casting solution can be explained both by the increase in the density of the surface layer of the formed membrane and its thickness [46].

The retention of nitrate and phosphate anions increases with increasing cellulose acetate concentration in the casting solution. For the nitrate anion, the retention increases from 63.7% obtained with the 2% CA membrane, to 96.5% with the 6% CA membrane. At the same time, the retention of phosphate anions increases from 75.2% for the 2% CA membrane, to 99.0% for the 6% CA membrane.

Choosing the concentration of cellulose acetate in the casting solution will be a process of evaluating the need to obtain a high flux and a retention suitable to the needs. From the data obtained by us, the concentration of 4% CA in the casting solution is the one that achieves the flux–retention trade-off.

Increasing the concentration of silver nanoparticles (Agnp) in the casting solution (Table 2) causes a slight decrease of the flux from 11.3 L·m⁻^2·h⁻¹, for the membrane with 7.3ppm Agnp in casting solution, to 9.7 L·m⁻^2·h⁻¹ for the 16.1 ppm Agnp membrane. For the retention values, an increase of approximately 3% was observed for both the nitrate and the phosphate anions, simultaneously with the increase in the Agnp concentration in the casting solution.

For the Agnp concentration in casting solution, an increase would be desirable to ensure a higher retention of anions at relatively constant permeate flows. This parameter is determined by the quality of the raw material, which is not controllable from this point of view.

The modification of the feed solution parameters consisted of varying the pH and the anion concentration (Table 3).

The permeate flux decreases slightly with both the increase in pH and the anion concentration in the feed solution, being approximately 11 L·m⁻²·h⁻¹. The value obtained is specific to the nanofiltration of dilute aqueous solutions and depends on the hydration of the membrane, but also on the anions considered [51,52].

The variation in retention as a function of feed solution pH is much more relevant. Thus, for both the nitrate and for the phosphate anion at a concentration of 30ppm, the increase in retention is remarkable, being about 5%, with the increase in pH from 4 to 7. Further increase in pH does not cause a significant increase in the retention of the two anions, being within the experimental error limits of approx. 95% for nitrate and approx. 99% for phosphate.

Increasing the anions concentration in the feed solution from 10 ppm until 50 ppm, at pH=7, leads to a decrease in retention of approximately 3% for both anions.

The results presented are consistent with literature data regarding the importance of ion hydration in the nanofiltration process [52,53,54], but also of the hydration of cellulose derivatives [55].

To be recommended for applications, membranes must have the most stable behavior during processing. To determine variation of flow and retention over time, the following parameters were established:

• working pressure: 15 atm;
• the concentration of nitrate and phosphate anions is equal to 30ppm;
• pH of the feed solution is 7.0±0.2;
• the feed solution flow rate is 120 L/h;
• the membrane is obtained from a casting solution of 4% cellulose acetate and approx. 11.5ppm Agnp;
• experiments are performed daily, keeping the same membrane;
• the feed solution is refreshed daily (600 L/day).

From Figure 12, a permanent, but slow, decrease in the flow can be highlighted from approx. 10.2 L·m⁻²·h⁻¹ on the first day, to approx. 9.5 L·m⁻²·h⁻¹ on the fifteenth day of operation.

The determination of retention during nanofiltration of phosphate and nitrate solutions (Figure 13) was performed under the experimental conditions established for determining the flux variation over time.

Over the entire time interval, the retention is significantly higher for the phosphate anion than for the nitrate anion. The retention of both anions decreases from the first day when it is about 98.1% for the phosphate anion and about 93.2% for the nitrate anion to 95.0% for phosphate anion and 84.3% for nitrate anion on day 15 of the experiment. The error in determining retention is ±0.2% for the nitrate anion and of ±0.3% for the phosphate anion, and is given by both the sampling of the samples taken and the analytical errors of determination.

3.4. Application Perspectives of Nanofiltration with CA–Agnp–PP Membranes

Using nanofiltration for obtaining drinking water is recommended in isolated areas such as farms and households in locations where it is common to extract water from shallow depths. This water source may be contaminated from local sources (animals, birds, fertilizers), including nitrate and phosphate anions. Presumptive concentrations would be 30ppm from each anion in the shallow well feed water.

Drinking water needs for a family of five can reach about 50 L/day, including water for drinking, cooking and washing the face and mouth.

The results presented in the previous subsections show that membranes obtained from cinematographic films membranes (CA–Agnp–PP) and assembled into a module with a surface area of approx. 1 m² can meet these needs, when operating at a pressure of 15 atm and a feed water flow rate of 120 L/h, rejection of 98% for phosphate and 95% for nitrate, and a flux of 10 L·m⁻²·h⁻¹. The data presented shows that in an operating interval of 5 hours, the drinking water needs of the entire family of five can be covered.

The concentrate obtained from nanofiltration can be used as fertilizer for ornamental flowers or in the vegetable garden (Figure 14).

Of course, the application presented is only a suggestion, but it is important for locations such as: houses in mountainous and hilly areas, agricultural or livestock farms in the plains, river deltas or in case of flooding.

4. Conclusions

Finding sources of raw material for obtaining membranes also involves recycling various polymeric waste.

For one of the most important artificial membrane polymers, cellulose acetate, various wastes were used, including cigarette filters or cinematographic films.

This paper presents the production of nanofiltration membranes using black-and-white cinematographic films and their testing in the nanofiltration of waters contaminated with nitrate and phosphate anions in order to obtain drinking water for isolated areas.

The manufacturing of the nanofiltration membranes consisted in the first stage of solubilizing cinematographic films in dimethylformamide, when a dispersion of cellulose acetate and silver nanoparticles (CA–Agnp) is obtained. The obtained dispersion was transformed into membranes by contacting the dispersion through a hollow fiber propylene contactor (PP) with deionized water. The nanofiltration membrane (CA–Agnp) is obtained, which was studied in a module with a surface area of 1 m².

In the study, operating parameters were varied: working pressure, the flow rate and pH of feed water, but also parameters that determine membrane formation: cellulose acetate concentration and silver nanoparticle concentration.

The results show that under accessible working conditions: pressure of 15 atm and a feed water flow rate of 120 L/h, pH = 7, rejection of 98% for phosphate and 95% for nitrate, and a flow of 10 L·m⁻²·h⁻¹, in an operating interval of 5 hours, the drinking water needs of a family of five can be covered (50 L/day).

The presented study may constitute the premise for obtaining drinking water for isolated areas from shallow waters. The presence of silver nanoparticles in membranes may be a solution for eliminating biofouling, which would lead to a decrease in transmembrane flux.