Scalable Manufacturing of Layer-by-Layer Membranes for Water Purification

CHRISTOPHER M. STAFFORD
National Institute of Standards and Technology

Water is critical to world health, economic development, and security. This was highlighted recently when the Obama administration hosted the White House Water Summit to raise awareness of water availability concerns across the United States and to engage stakeholders in identifying long-term solutions for water production and management suitable for investment.

BACKGROUND

Water availability is not a new issue. The demand for clean water has risen dramatically since the Industrial Revolution and will continue through the Information Age and beyond. The world’s population has climbed to 7 billion, and as it expands further and water scarcity becomes a more widespread reality, it is imperative to think creatively about ways to safeguard access to clean water.

The obvious and most fundamental purpose of clean water is as a source of sustenance, to produce the food and water that every society needs to survive. Clean water is also vital to many of the complex processes that produce the technology that modern society demands and consumes. Many of those processes, however, introduce contaminants, such as heavy metals and other chemicals, into local water supplies.

For all these reasons there is a clear and growing need for technologies and processes that ensure water is clean, safe, and accessible (Shannon et al. 2008).

MEMBRANE TECHNOLOGY

Membranes and membrane technology, in particular polymer-based membranes, are key to the world’s water future (Geise et al. 2010). Membranes are capable of separating a wide range of contaminants from impaired water sources, from viruses and bacteria to heavy metals to dissolved salts.

Given that water covers 71 percent of Earth’s surface and 97 percent of that water is in the world’s oceans, an obvious focal point of research is desalination, the recovery of water from high-salinity water sources. This can be an energy-intensive process because of the high osmotic pressure of seawater: the average sea surface salinity is 35,000 g/L (for simplicity, let’s assume it is all sodium chloride), which generates an osmotic pressure (Δπ) of nearly 400 psi or 27.4 bar. Desalination is nonetheless highly attractive because of the volume of water available for recovery.

This paper focuses on membrane desalination via reverse osmosis, so a short introduction to reverse osmosis is warranted.

REVERSE OSMOSIS

In traditional osmosis, water flows across a semipermeable membrane from regions of low solute concentrations (in this example, pure water) to regions of high solute concentration (a concentrated salt solution), in effect diluting the solute and lowering the overall free energy of the system. The driving force for the flow of water is the osmotic pressure and is dependent on the concentration of solute molecules in the concentrated solution.

In reverse osmosis, pressure is applied to the high concentration region, which has to be greater than the osmotic pressure of the solution to drive water from regions of high concentration to those of low concentration (see Figure 1), again with the aid of a semipermeable membrane. This process generates purified water on one side of the membrane and a more concentrated salt solution on the other side.

The water flux (J_w) through the semipermeable membrane can be defined as:

where A is the membrane area, K_w is the permeability of the membrane, h is the membrane thickness, and (ΔP – Δπ) is the difference between the applied pressure and the osmotic pressure. From this equation, one can see that there is an inverse relationship between the applied pressure and the membrane thickness. Thus, a thinner membrane would be ideal as it would require less energy (pressure) to generate a given amount of water from an impaired water source of a given concentration of dissolved solutes (i.e., osmotic pressure).

PARADIGM SHIFT IN MEMBRANE TECHNOLOGY

The manufacture of today’s state-of-the-art reverse osmosis membranes is based on 1970s technology of interfacial polymerization of a selective layer directly on a porous support (Cadotte 1977, 1979).

In this process, polymerization of an aromatic triacid chloride (A) and an aromatic diamine (B) occurs at the interface of two immiscible liquids, where one liquid (typically the aqueous amine solution) is wicked into the porous support. The result is a highly crosslinked, aromatic polyamide (think crosslinked Kevlar or Nomex) membrane that selectively allows the passage of water and rejects salt. The chemistry easily lends itself to roll-to-roll (R2R) or web processing, can be performed over large widths of substrates, and produces a relatively low number of defects across the membrane surface.

Over the past 40 years, this membrane technology has slowly evolved through an Edisonian, trial-and-error approach. The process makes extremely thin (100s nm) selective membranes, but they are difficult to characterize because of high roughness and large heterogeneity. Thus, understanding of how these membranes work is insufficient to allow the rational design of next-generation membranes.

In 2011 my research team at NIST proposed a paradigm shift in how these types of membranes are fabricated, in which the selective layer is created layer by layer through a reactive deposition process. We anticipated the resulting mem-

branes to be smooth, tailorable, and exceptionally thin (10s of nm). The ability to tune the membrane thickness makes this process attractive due to potential energy savings from reduced pressure requirements.

In our original demonstration (Johnson et al. 2012), we used a solution-based deposition process in which we sequentially and repeatedly layered each reactive monomer (A + B) onto a solid substrate through an automated spin coating process. We observed growth rates of approximately 0.34 nm/cycle, where one cycle represents a single (A + B) deposition sequence.

The growth rate was shown to be dependent on monomer chemistry, spin conditions, and rinse solvents (Chan et al. 2012). Additionally, the layer-by-layer films are quite smooth, exhibiting a remarkably low root mean square (RMS) roughness of 2 nm compared to commercial interfacial polymerized membranes that exhibit an RMS roughness of 100 nm or more.

The fact that the films are relatively smooth and homogeneous has two compelling advantages: (1) it enables advanced measurements of the film structure via scattering- or reflectivity-based techniques, among others, and (2) it allows quantitative structure-property relationships to be developed as the film thickness is well defined. X-ray photoelectron spectroscopy and swelling measurements indicate that the crosslink density of the layer-by-layer membranes is comparable to that of their commercial counterparts, even though the layer-by-layer films are considerably thinner (Chan et al. 2013).

Other researchers have adopted this approach and verified that membranes produced using this layer-by-layer process indeed have viable water flux and salt rejection (Gu et al. 2013).

TECHNOLOGICAL CHALLENGES

One major drawback to the solution-based layer-by-layer approach is throughput: spin-assisted assembly is a relatively slow process and not easily scalable.

We have started to explore the use of a vapor-based approach, in which each monomer is deposited in the gas phase, similar to atomic layer deposition of metals and oxides (Sharma et al. 2015). Each monomer/precursor is (1) heated in order to build up sufficient vapor pressure of the precursor and then (2) metered into a rotating drum reactor through dosing ports with differential pumping and purge ports on either side (see Figure 2). Again, the number of cycles (or number of consecutive ports) determines the thickness of the resulting membrane.

This approach has many advantages—such as speed, safety, and scalability—over the solution-based approach. We have shown that we can deposit 20 layers of (A + B) per minute (3 s/cycle), compared to 1 layer of (A + B) every 2 minutes using the solution-based approach (2 min/cycle). The growth rate using the vapor-based approach (0.36 nm/cycle) is nearly identical to the solution-based approach, ensuring that the processes are similar. One key advantage of the

vapor-based approach is the potential for scale-up via continuous, roll-to-roll, or web processing.

But there are still many challenges yet to overcome, from membrane support design to membrane characterization. For example, the active layer must be coated onto a microporous support layer; thus a method for adequately preventing intrusion of the reactants into the underlying support must be devised. Also, the polyamide network topology needs to be optimized to allow the highest flux of water while maintaining adequate rejection of salt. This can be achieved through judicious monomer selection and deposition conditions.

A paradigm shift in manufacturing may lead to membranes and processes that are more energy efficient, offering one solution to the grand challenge of water security.

REFERENCES

Chan EP, Young AP, Lee J-H, Stafford CM. 2013. Swelling of ultrathin molecular layer-by-layer polyamide water desalination membranes. Journal of Polymer Science, Part B: Polymer Physics 51:1647–1655.

Geise GM, Lee H-S, Miller DJ, Freeman BD, McGrath JE, Paul DR. 2010. Water purification by membranes: The role of polymer science. Journal of Polymer Science, Part B: Polymer Physics 48:1685–1718.

Gu J-E, Lee S, Stafford CM, Lee JS, Choi W, Kim B-Y, Baek K-Y, Chan EP, Chung JY, Bang J, Lee J-H. 2013. Molecular layer-by-layer assembled thin-film composite membranes for water desalination. Advanced Materials 25:4778–4782.

Johnson PM, Yoon J, Kelly JY, Howarter JA, Stafford CM. 2012. Molecular layer-by-layer deposition of highly crosslinked polyamide films. Journal of Polymer Science, Part B: Polymer Physics 50:168–173.

Shannon MA, Bohn PW, Elimelech M, Georgiadis JG, Mariñas BJ, Mayes AM. 2008. Science and technology for water purification in the coming decades. Nature 452:301–310.

Sharma K, Hall RA, George SM. 2015. Spatial atomic layer deposition on flexible substrates using a modular rotating cylinder reactor. Journal of Vacuum Science and Technology A 33:01A132.