Membranes for artificial photosynthesis Environment

Introduction

Artificial Photosynthesis

Artificial photosynthesis is a process for obtaining hydrogen and oxygen from sun and water and also hydrocarbons from carbon dioxide, just as plants do^[1].

Photosynthesis itself is a very complicated process involving many complex relations with macromolecules. A new two-stage technology called artificial photosynthesis has been designed to shorten this process and gain more control over it. In the first stage water, with the input of sunlight, produces oxygen, protons and electrons. In the second stage protons and electrons are used to store this energy in the form of hydrogen. Other products like methane or methanol can also be generated. If hydrocarbons are sought some carbon input is needed, such as atmospheric carbon dioxide^[2].

This whole process can be described as:

The half-reactions taking place in the anode and cathode are:

This reaction can be produced in a photoelectrochemical cell (PEC) also called a Grätzel cell [1, 3]. This cell uses photosensitisers attached to the surface of nanocrystalline TiO2 film. The arrival of a photon excites the complex and oxidises the water-oxidation catalyst. This catalyst, called «blue dimer», is usually a ruthenium complex [4], and is capable of attracting the four water electrons, splitting it into protons and oxygen. These electrons then flow through the TiO2 semiconductor to the platinum cathode, where a water molecule is formed for each pair of protons arriving through the proton-exchange membrane (Figure 1)[1]. This part of the project is being carried out in collaboration with the Institut Català d’Investigació Química (ICIQ).

Artificial photosynthesis cells and their advantages over photovoltaic cells

An artificial photosynthesis cell has two main advantages over conventional silica photovoltaic cells: lower manufacturing costs and better system performance. Both these advantages stem from the use of a catalyst, which drastically reduces the voltage input necessary for water electrolysis [2]. Moreover, this cell can be used not only to produce hydrogen but also to carry out chemical-reduction reactions [1]. With this technology humankind will have on hand a cheap and clean energy source that is not only carbon-free but may also reduces the amount of atmospheric carbon dioxide. This would favour a more sustainable lifestyle and help in the fight against environmental pollution; it could also be a solution to the fossil-fuel energy crisis we are currently facing.

The role of the membranes and the object of the study

This research centres on the development of biomimetic proton-exchange membranes to be used in future artificial photosynthesis cells[4, 5, 6]. The membranes have been produced by means of the phase inversion technique using a biomimetic liquid-crystal polymer previously synthesised in the laboratory. The various membranes obtained were tested in an experimental cell to assess their properties and check them against the commercial Nafion® membrane.

The Nafion® membrane

Nafion® 117 is a perfluorosulfonic acid membrane developed in 1966 and marketed by DuPont. It is 177.8 μm (0.007 in) thick. This polymer contains sulfonic acid, which drives the proton-hopping form of transport. This membrane has to be hydrated to be permeable, since the hydrophilic sulfonic acid groups line up in channels ranging in size from 10 to 50 Å. Through this channel pass the protons and cations but also the water and other undesired molecules like methanol. Furthermore Nafion® swells by water absorption, worsening its mechanical properties. The transport mechanism of Nafion® is therefore due to three factors: the active transport of protons, also called electroosmotic resistance; pore diffusion and conventional diffusion [7]. This membrane has poor selectivity, allowing through not only protons but also other undesired substances that might damage the electrodes or provoke undesired reactions (Figure 2[8]). These reactions might drastically reduce cell performance or produce unforeseeable effects on the cell, such as corrosion, overheating, explosions. But perhaps the biggest drawback is its price; the Nafion® is under patent and costs from 500 to 1000 dollars per m2 [8].

Nonetheless the Nafion® membrane is the only commercial proton exchange membrane used in large-scale industrial electrochemical systems [3]. It is therefore the worldwide benchmark for proton-exchange membrane fuel cells (PEMFC). It has also been thoroughly studied in the literature. For this reason the Nafion® membrane has been used as the crosscheck for gauging the performance of the new synthesised biomimetic polymers.

Biomimetic Membrane

A quick glance at the end bibliography shows that very little research has been carried out to develop new polymers that do not use sulfonic acid as proton carrier. As already noted, use of the sulfonic group is the main reason for the membrane’s poor selectivity. This is so because the membrane’s proton-transporting channels are big enough to allow the passage also of water, methanol and other species.

Biomimetic membranes involve a totally different concept from the commercial proton exchange membranes like the Nafion® membrane

A new approach has therefore been mooted, avoiding the use of sulfonic groups for proton transport and channel formation. With this new approach a different transport mechanism has to be used to increase proton selectivity; a new type of material therefore has to be designed to form the membrane.

Ideally, fluorine should be avoided due to its high price and the difficulty of eliminating the membrane at the end of its life cycle due to its high toxicity. Also the new material must not absorb water since this leads to poor mechanical properties and low proton selectivity. This is important because Nafion® is useful for hydrogen fuel cells when used in humid gas phase but performs poorly in direct methanol fuel cells (DMFC) and in artificial photosynthesis cells, with liquid phase working. A new biomimetic membrane has therefore been developed using oxygen’s coordination capacity rather than cations to transport the protons through the dative bond along a polyether chain.

This membrane is based on observation of the self-assembly of the tobacco mosaic virus (TMV)[9].

Thus the tapered hydrophilic groups self assemble into columns, giving rise to ionic channels[10], which are then responsible for mechanical resistance and protection of the membrane from water absorption due to their hydrophobia. The centre of the column is thus formed by a continuous polar polyether chain acting as the proton carrier[5]. This combination should favour a much higher proton selectivity against other substances like water and methanol. The self-assembly can be seen in figure 3.

To ensure sufficient membrane permeability these channels should run perpendicular to the membrane surface. The membrane therefore needs to be as thin as possible to reduce the number of non-aligned ionic channels compressed between the two membrane surfaces.

Materials and Methodology

The biomimetic membranes are built up from a liquid crystal polymer obtained by chemical modification of the commercial polymer poly(epichlorhydrin) (PECH) with a tapered group comprising potassium 3,4,5-tris[4-(n-dodecan-1-yloxy) bencyloxy]benzoate [11].

A five-stage chemical synthesis process was used to obtain the polymer.

In the first stage p-(ndodecan-1-yloxy)benzyl alcohol is obtained from a solution of f 4-hydroxybenzyl alcohol (10 g, 0.081 mol), 1-bromododecane (20.0 g, 0.081 mol), anhydrous potassium carbonate (11.0 g, 0.081 mol) and 250 ml de acetone, using the catalyst 18-crown-6 ether.

In the second stage p-(n-dodecan-1-yloxy)benzyl chloride is obtained by reaction of the first-stage product with Tionyl chloride (6.93g, 0.058 mol) dissolved in p-(n-dodecan-1-yloxy)benzyl alcohol (17.02 g, 0.058 mol) and in 119 ml of dichloromethane (CH2Cl2).

In the third stage methyl 3, 4, 5-Tris[4-(n-dodecan-1-yloxy)benzyloxy]benzoate is obtained by reaction of the second-stage product with methyl gallate (methyl 3,4,5-hydoxy benzoate) (2.8 g, 0.015 mol), dry dimethylformamide (DMF 40 ml), potassium carbonate (6.5 g, 0.045 mol) and tetrabutylammonium hydrogen sulphate (0.1 g) dissolved in 4-(n-dodecan- 1- yloxy)benzyl chloride (15.5 g, 0.05 mol) and 60 ml of dry DMF.

The biomimetic membranes are made up by hydrophobic macromolecules that protect the nanochannels, which transport the protons by establishing hydrogen bridges with the ether groups inside

In the fourth stage the potassium salt Potassium 3,4,5-Tris[4-(n-dodecan-1-yloxy)benzyloxy]benzoate is obtained by reaction of the third-stage product t with a solution of KOH in C2H5OH (potassium hydroxide 34.3 g in ethyl alcohol 102 ml) dissolved in methyl 3,4,5-Tris[4-(n-dodecan-1-yloxy) benzyloxy]benzoate (10.2 g, 0.01 mol).

Finally, in the fifth stage the biomimetic liquid crystal polymer with a 67% degree of modification is obtained by reaction of the fourth-stage product with the poly(epichlorhydrin) (PECH) dissolved in THF. Its structure is shown in Figure 3.

Membrane production by casting

The first membranes were made with the polymer obtained from the former synthesis by means of an evaporation process called casting. The biomimetic polymer obtained by modification of the PECH was dissolved in chloroform and the polymer solution was placed in water to align ionic channels, at a ratio of 3 g of water to 5 g of polymer solution. The water aligns the columns, which are hydrophobic in their outer part and therefore try to minimise their contact with water. In the first membrane obtained, 6% in weight of the polymer was mixed in chloroform. In the second membrane obtained, 5% in weight of the polymer was mixed in chloroform. The two immiscible phases formed in the mixture are then placed in a container in an oven at 40 ºC. This evaporates off the water and solvent from the polymer solution, thus precipitating the membrane.

Membrane production by phase inversion and glass modification

Phase inversion is a simple membrane-production process whereby the liquid crystal polymer is made to precipitate in an organic solvent, submerged in a bath of water. The precipitation occurs because the organic solvent leaves the polymer solution to dissolve in the bath water. The membrane thus precipitates out because the polymer is insoluble in water.

The membrane preparation method by phase inversion has clear advantages over the traditional casting technique. With the casting method there is a long wait until the water evaporates off and then another wait until the solvent evaporates off. With phase inversion, on the other hand, the membrane is formed instantly. It is also an easily controllable process widely used in industry and can also be used for forming a tailor-made membrane. This means that, depending on the precipitation conditions, two clearly differentiated zones can be obtained. The first is a dense zone or non-porous zone, which is what confers on the membrane the selective ion-carrying property through the biomimetic nanochannels. The second zone is porous, giving the membrane the necessary mechanical properties for its actual industrial application.

Artificial photosynthesis is a process for obtaining energy from water and sunlight, using a catalyst of ruthenium and titanium oxide, and then storing this energy in the form of hydrogen

To prepare the membrane the liquid crystal polymer was dissolved at 25% weight in the THF solution.

To ensure better alignment of the biomimetic channels and a solid surface as polar as possible, two glass supports were modified with the ions NH3+ (totally modified glass) and OH- (partially modified glass) according to the technique described in [12]. A more polar surface facilitates better alignment of the membranes, since the glass surface is hydrophilic (polar) whereas the coating of the biomimetic channel is hydrophobic (non-polar). A repulsion is therefore set up between them, enabling the channels to line up perpendicular to the glass, thus minimising their energy interaction with the surface.

Finally, using a Kpaint applicator, a 250 μm sieve was passed over the glass surface to create an even film of uniform thickness. After phase-inversion precipitation of the membrane, it was extracted and dried to remove any impurities. The characteristics of the membranes thus obtained are shown in Table 1.

Determination of proton permeability

The membranes’ coefficient of proton permeability, and also their permeability, were measured using the technique described in [13]. In this way membrane permeability was gauged by means of a testing cell. The feed cell A contains the HCl solution, while the stripping cell B contains the NaCl solution (Figure 4). By measuring the change of pH in the stripping cell, membrane permeability can be determined, keeping matter balanced in the cell. For the permeability experiments a 0.05 M NaCl solution and a 0.05 M HCl solution were used.

Thus, to determine the coefficient of permeability p (cm/s), the following equation has been used:

The flow of protons through the membrane J (mols/m2s) can be determined from the equation:

The permeability of membrane P (cm2/s) can be obtained from the following equation, where l is the membrane thickness.

Results and Discussion

Membrane production by casting

A membrane has been successfully obtained following this method. This method was dropped, however, on the grounds of irreproducible results. This method is also much more time-consuming than phase inversion, so it is difficult to control and cannot produce thin membranes. Even so, a membrane produced by this method has been used to compare its permeability with the phase-inversion membranes.

Membrane production by phase inversion

Membranes produced by this technique gave different permeability readings. The membrane obtained with the completely modified glass matched the permeability of the casting-produced membrane while the permeability of the membrane synthesised with partially modified glass was lower.

This method was also found to produce thin membranes with a dense non-porous layer. This is a very positive factor in terms of their application in artificial photosynthesis fuel cells, preventing the gaseous hydrogen from leaking out through the membrane, while also endowing the membrane with a certain degree of flexibility since the membrane is not completely dense.

The development of this membrane preparation method has enabled us to establish a standardised, rapid and efficient method for producing membranes with reproducible proton-transporting properties.

Proton permeability

The permeability assessment results obtained in the experimental cell are shown in Table 2. The membrane’s time to proton conduction is also given. The coefficient of permeability readings p (cm/s) are shown in Figure 5, while the permeability data P (cm2/s) are shown in Figure 6, excluding the experiments made with non-woven membranes since their thickness is uneven.

The membranes e1 obtained with totally modified glass show irregular conductivity and long times to proton conduction. Partially modified glass achieved better membrane alignment. This prompted development of the f1 membranes, showing a coefficient of permeability to match Nafion® and membrane 44-4. Nonetheless, since membrane f1 is thinner, its permeability is slightly lower than the latter two.

Swelling

One of the main drawbacks of the Nafion® membrane, apart from its high fluorine content, is the need of being moistened to conduct protons. This produces a swelling effect in the membrane, impairing its mechanical properties. It also increases the difficulty of operating with fuel cells, since the relative humidity on both sides of the membrane needs to be controlled to induce proton conductivity.

Unlike the Nafion® 117 membranes, therefore, which swell up, retaining 15% of water, the biomimetic membranes do not vary their weight or thickness under the same conditions. This behaviour can be explained by the biomimetic membranes’ low water interaction and different transport mechanism. Biomimetic membranes transport the protons through an ionic channel formed by ether groups with which the proton establishes hydrogen bonds; the Nafion® membrane, being a strong acid, works by dissociating the protons from the sulfonic acid and reacting strongly with the water to enable proton passage.

The biomimetic membranes have been produced by the phase inversion technique, obtaining a dense proton-transporting layer that prevents hydrogen leaks

Contact angle of glass support and membrane

The contact angle is the angle at which a water droplet interface meets a solid surface. Determination of the contact angle between the water and glass serves to assess the glass’s hydrophilicity, i.e., finding out which glass is most polar and hence has the best membrane alignment properties. The results are shown in Figure 7, and the contact angle readings in Table 3.

The figure and table show that the partially modified glass has a lower contact angle than the totally modified glass. This is so because the water has more affinity for the surface and therefore spreads out rather than shrinking; this makes the surface much more polar. A more polar solid surface has thus been developed to facilitate membrane alignment during the phase inversion process. This explains why the conductivity of membrane f1, obtained on the more polar support, is higher than e1 obtained on a less polar support.

The contact angle with the two sides of the membrane has also been determined, both that which is in contact with the glass during membrane precipitation and that which is in contact with the water. The membranes assessed were e1, formed with the totally modified glass, and f1, obtained with the partially modified glass. The results are shown in table 4.

This table shows that membrane f1, obtained with the partially modified glass, has a slightly lower contact angle than e1, although the difference is negligible. This would mean that membrane f1 might be more hydrophilic than e1. The better alignment of membrane f1 thus ensures that there are more ionic channels on the membrane face, whose lower part is hydrophile (polar) due to the ether groups making it up. At the same time there are fewer tapered non-polar groups, thus increasing the membrane’s hydrophilicity.

Thickness and morphology

To find out the membrane’s alignment each one was analysed under polarised light, a regular alignment showing up.

The membranes obtained can work for a minimum of 335 hours without stopping, offering a good proton conductivity without suffering structural damage or warping since no water is retained inside

A scanning electronic microscope (SEM) was used to determine the membranes’ thickness and morphology. The results reflect asymmetric membranes with a porous part and a dense part. This gives them good mechanical properties, enabling them to bend without breaking, while the dense part prevents large hydrogen gas molecules from passing freely through the membrane. This requirement is essential for any device that needs to be selective, like the artificial photosynthesis cell or fuel cells.

Analysis of the SEM images shows a mean membrane thickness of 20 μm (Figure 8). This figure is crucial because, working with such thin membranes, compound membranes can be built up merely by joining and adding on more layers.

The membranes were also analysed using an atomic force microscope (AFM) to delve into more detail of the membrane surface structure. These images show the same column aggregates before and after the experiment. (Figure 9). These aggregates have a mean size of 40 nm and could correspond to biomimetic column aggregates, thereby confirming the good chemical resistance of the membrane.

También se han analizado las membranas utilizando el Microscopio de Fuerza Atómica (AFM), con objeto de conocer más detalles sobre la estructura de la superficie de la membrana. En estas imágenes se pueden ver las mismas agregaciones columnares antes y después del experimento (Figura 9). Estas agregaciones tienen un tamaño medio de 40 nm y podrían corresponder a agregaciones de columnas biomiméticas, lo que confirmaría la buena resistencia química de la membrana.

Mechanical and chemical strength

The proton permeability experiments showed that synthesised membrane e1 can work for 335 hours on the trot without suffering any structural damage, even with strong constant agitation and in the presence of acids at pH 1.30 without suffering any sort of deterioration. A noteworthy feature here is the membrane thickness of 20 μm.

Transport Mechanism

Since there are no references to this transport mechanism in the scientific literature, a study was made of the proton transport mechanism through the membrane. To do so samples of the feed and stripping solutions were taken at the start and end of the e1 membrane permeability experiment. A reading was hence taken of the cation’s sodium concentration (Table 5) and pH concentration (Table 6).

By converting pH readings to molar concentration, a calculation can then be made of the change of proton concentration in the stripping solution. The calculation is made in the stripping solution because it is where the pH variation can be determined most progressively. The variation in the sodium ion concentration is measured in the feed solution, where the best analytical precision is to be obtained. The results are shown in Table 7.

This table shows that the change in molar concentration is the same for protons and for sodium ions. This means that for each proton passing from the feed to stripping solution a sodium ion passes from the stripping to the feed solution; this shows the proton transport mechanism to be counter-transport (Figure 4).

Conclusions

This research has proven the feasibility of producing an alternative to such a well-established commercial membrane as Nafion®. This study merges liquid crystal polymer materials science with nanotechnology and chemical engineering.

Sodium and pH analysis has shown the proton transport mechanism to be counter-transport

A methodology has been developed to obtain an aligned biomimetic membrane with a similar proton permeability to its commercial alternative. A methodology has also been established for ensuring reproducible and rapid membrane manufacture. All this has been achieved by chemically modifying the glass support where the membrane is precipitated to make it more polar and thus ensure better aligned membranes. The membranes obtained were also thin and uniform, with a dense layer and a porous layer. The membranes withstood mechanical stress and chemical agents for a long time. The hydrophilicity of the support and membrane has also been ascertained by measuring the contact angle. Furthermore, by means of the swelling experiment, a membrane has been obtained that, unlike Nafion®, absorbs absolutely no water.

This study lays down the foundations for a more in-depth study of the transport mechanism. A comparison of the variation over time of sodium-ion and proton concentration in the permeability cell shows this to be a counter-transport mechanism.

---

TO FIND OUT MORE

1. Concepcion J, Jurss J, Brennaman M, Hoertz P, Patrocinio A, Murakami Iha N, Templeton J, Meyer T. Making Oxygen with Ruthenium Complexes. Accounts of Chemical Research 42 (2009), 12, 1954–1965.
2. Brimblecombe R, Dismukes G, Swiegers G, Spiccia L. Molecular water-oxidation catalysts for photoelectrochemical cells. Dalton Transactions 2009 (2009), 43, 9374–9384.
3. Zaidi S, Matsuura T. Polymer membranes for fuel cells. Springer Verlag, 2009.
4. Francàs L, Sala X, Benet-Buchholz J, Escriche L, Llobet A. A Ru-Hbpp- Based Water-Oxidation Catalyst Anchored on Rutile TiO2. Chem- SusChem 2 (2009), 4, 321–329.
5. Ronda J, Reina J, Cádiz V, Giamberini M, Nicolais L. Self-organized liquid-crystalline polyethers obtained by grafting tapered mesogenic groups onto poly (epichlorohydrin): Toward biomimetic ion channels. Journal of Polymer Science Part A: Polymer Chemistry 41 (2003), 19, 2918–2929.
6. Gust D, Moore T, Moore A. Solar Fuels via Artificial Photosynthesis. Accounts of Chemical Research 42 (2009), 12, 1890–1898.
7. Schultz T, Zhou S, Sundmacher K. Current status of and recent developments in the direct methanol fuel cell. Chemical Engineering & Technology 24 (2001), 12, 1223–1233.
8. Mauritz K, Moore R. State of understanding of Nafion. Chem. Rev 104 (2004), 10, 4535–4586.
9. Percec V, Heck J, Lee M, Ungar G, Alvarez-Castillo A. Poly {2- vinyloxyethyl 3, 4, 5-tris (4-(n-dodecanyloxy) benzyloxy) benzoate}: a self-assembled supramolecular polymer similar to tobacco mosaic virus. Journal of Materials Chemistry 2 (1992), 10, 1033–1039.
10. Percec V, Cho W, Ungar G, Yeardley D. Synthesis and NaOTf mediated self-assembly of monodendritic crown ethers. Chemistry- A European Journal 8 (2002), 9, 2011–2025.
11. Ronda J, Reina J, Giamberini M. Self-organized liquid-crystalline polyethers obtained by grafting tapered mesogenic groups onto poly (epichlorohydrin): Toward biomimetic ion channels 2. Journal of Polymer Science Part A: Polymer Chemistry 42 (2004), 2, 326–340.
12. Yoshio M, Kagata T, Hoshino K, Mukai T, Ohno H, Kato T. One-dimensional ion-conductive polymer films: alignment and fixation of ionic channels formed by selforganization of polymerizable columnar liquid crystals. J. Am. Chem. Soc, ACS Publications (2006), 128, 5570-5577.
13. Zhang X, Benavente J, García- Valls R. Lignin-based membranes for electrolyte transference Journal of Power Sources, Elsevier (2005), 145, 292-297.

top