Organic solvent nanofiltration (OSN) is starting to carve out a role in a number of process applications, including polymeric impurity removal, monomer/dimer separation, catalyst recovery and recycle, and color elimination. The technology requires less than 10% of the energy needed by thermal separation techniques such as distillation, and can provide ultrapure products and help increase product yield.
At the forefront of OSN developments is Evonik Industries, Essen, Germany. Its efforts really took off in March 2010 with the acquisition of Membrane Extraction Technology (MET), London, U.K. — a company spun off in 1996 by the department of chemical engineering at Imperial College London that boasts a range of proprietary technologies and offers DuraMem OSN membranes.
Andrew Livingston, Evonik MET's chief innovation officer and head of the department of chemical engineering at Imperial College, estimates the U.S. process industries alone could save between 3 and 5% in energy costs if membranes were used to extract organic solvents.
The DuraMem membranes reportedly offer long-term stability in polar solvents, including aggressive ones such as those in the polar aprotic family — for example, acetone, tetrahydrofuran and dimethylformamide. The molecular cut-off range is 150–900 Da, while maximum recommended temperature is 50°C and typical operating pressure is 20–40 bar. The membrane comes in two formats: flat sheets for lab trials, and spiral-wound modules for pilot trials and production.
"We call DuraMem our first generation (G1) solution. Then we started to get requests from very specific customers for apolar solvent solutions. So we specifically developed our PuraMem series for this. This is our G2 solution," explains Yuri Bouwhuis, Evonik MET's managing director.
Also available in flat sheets (Figure 1) and spiral-wound modules, PuraMem has a molecular cut-off range of 280–600 Da and the same temperature and pressure ranges as DuraMem. It's particularly aimed at solvents such as toluene, heptane, hexane, methyl-isobutylketone, ethyl acetate, oils and apolar high boilers.
"Customers are now looking for even-lower molecular weight cut-off points, below 280 Da, or a higher flux than the current PuraMem series can offer. These will be our G3 solutions," he adds.
The company currently has 30 applications up-and-running worldwide.
The results at one pharmaceutical manufacturer typify the sorts of savings possible with OSN, says Bouwhuis. That company makes a very expensive active pharmaceutical ingredient (API), worth in excess of €30,000/kg ($40,000/kg). The original purification strategy led to a loss of 10% of API product in each batch. "Using an Evonik OSN solution, we cut that to well below 1% and are saving the company over €7.5 million/yr ($9.9 million/yr)."
Another customer relies on OSN for polishing a product stream prior to chromatography, to optimize the use of the chromatography column and reduce the number of resin changes needed. A third customer relies on OSN for solvent recycling, allowing it to totally eliminate the need for expensive offsite solvent disposal. At one fine chemical manufacturer, OSN enables reuse of a very expensive catalyst, he adds.
Implementing OSN requires several steps. First, a laboratory-scale feasibility or proof-of-concept test evaluates membrane sheets' performance to identify the most suitable variant, for example, to separate two compounds or to concentrate one compound.
Then, a mini-module of the selected membrane usually is installed in a small pilot unit to establish both rejection and flux performance (Figure 2). The data collected enable adjustment of process parameters such as flow rate, pressure and temperature for full process optimization. "At this stage, we are able to make an economic assessment as to whether a process is economically viable," notes Bouwhuis. "As a rule of thumb, we consider a process to be viable when the payback time is equal to or less than one year."
These steps can be done either at the customer's site or at one of Evonik's regional labs in Europe, North America and Asia.
If using OSN makes economic sense, Evonik then usually partners with an industry-specific OEM, or a customer's preferred OEM, to design a suitable plant for the particular application.
BROADENING THE APPEAL
OSN isn't the holy grail of separation, Bouwhuis stresses. "Can we separate a solvent from a solvent at the moment? No. Is this a typical modular plug-and-play solution? No. It really is a lot more complex than plugging a cartridge in. The company's new app, which will shortly be launched under HP Polymers, will include a decision tree so that prospective customers can determine if OSN can work as an enabling technology for them."