The environmental degradation caused by the consumption of fossil and the increasing energy demand is of great concern when taking in account social welfare. It is forecast that fossil fuel will not be able to defray society’s demands. Efforts are being placed in the study of lignocellulose-biomass conversion, which represents a possible solution to this problem since platform molecules can be obtained from it. These type of molecules can be used to produce fuels and fuels additive.

Hydrogenation reactions are industrially relevant in petrochemical and petroleum refining and in emerging biorefining. The reaction that was studied was the hydrogenation of levulinic acid (LA) to produce gamma-valerolactone (GVL), a chemical platform. Previous studies had conducted this reaction in a packed bed reactor. One of the problems presented in this kind of reactor is the mass transfer limitation, hydrogen delivery to the catalyst sites. Membrane reactors can lessen the mass transfer limitations by supplying hydrogen to the catalytic sites on the membrane surface. The main objective of this project was to develop a polymeric membrane for the hydrogenation of LA. Polytetrafluoroethylene (PTFE) was used as the support of the polymer Matrimid. Among the different types of membranes developed, we focus on the reaction conducted with a membrane that consisted of 0.25wt% and 1.0wt% Matrimid dissolved in methylene chloride (DCM) and spin-coated onto the PTFE membrane. Also, different types of solvents for the LA hydrogenation were evaluated such as glycerol, water and gamma-butyrolactone (GBL). At constant temperature of 70oC, glycerol did not show any yield of GVL after13-15 hours of reaction. When GBL was used as solvent for LA, a bronze colored liquid was obtained. Since this chemical has similar properties to GBL, it was not able to be analyzed by simple means. Water showed to be the most efficient solvent to produce GVL. Also, the PTFE permeability for water and GBL was evaluated. GBL has a lower vapor pressure in comparison to water, therefore allowing the reaction to be conducted at higher temperature without extreme permeability of the liquid phase, as expected due to its high boiling point. In the case of water, 70C was chosen as reaction temperature for further experiments, because produced a manageable water permeance through the PTFE membrane.

The hydrogenation of LA was conducted supplying hydrogen from the lower side of the membrane and from the upper side (in a continuous loop flow). Results have shown that the most efficient way to supply hydrogen is from below the membrane surface since supplying it from the upper side resembles a more traditional three-phase reactor such as a packed bed reactor.

Comparing the PTFE layer that was coated with 0.2wt% RuCl3 dissolved in ethanol (catalyst solution) with the one that consisted of PTFE with a polymeric layer of Matrimid, greater GVL results were obtained with the one that consisted of unmodified PTFE with catalyst. The presence of a polymeric layer on the PTFE lowered the GVL yield due to the inefficiency of delivery of hydrogen to the catalyst sites. An increase in pressure and temperature helped to lessen this effect.


Membrane Fabrication

Method 1. This type of membrane was made by crouching the pores of the PTFE. Catalyst solution consisted of 0.2wt% RuCl3 dissolved in ethanol. This solution was poured on top of the membrane and was then left to hydrogenate for two hours to ensure that the only substance on the surface was Ru.

Method 2. A polymeric solution consisting of 34g Matrimid, 74g THF, 74g GBL and 18g of butanol was poured on top of the PTFE layer and a casting knife was used to ensure a smooth and even surface. Then, it was submerge in water until the polymer harden. After the water treatment, a catalyst solution consisting of 0.2wt% RuCl3 dissolved in ethanol was poured on top and the membrane was hydrogenated for two hours.

Method 3. Two solutions of Matrimid dissolved in DCM were made. One consisted of 1wt% and the other of 0.25wt%. They were poured on the PTFE layer and were spin coated at a 2500rpm for 60s. The amount of 20uL, 50uL and 100uL were poured on the PTFE/Matrimid membrane. Each one of them were hydrogenated for two hours.

Method 4. This type of membrane was made by preparing a 20wt% solution of Matrimid dissolved in GBL mixed with a 0.5wt% RuCl3 catalyst solution dissolved in GBL. A casting knife was used to ensure a smooth, tin and even surface. The membrane was submerge in water until the polymer harden. It was then coated with 0.2wt% RuCl3 dissolved in ethanol and it was hydrogenated for two hours.

Method 5. This membrane consisted of PTFE coated with the catalyst solution which consisted of 0.2wt% RuCl3 dissolved in ethanol.

Calibration Samples

GBL calibration sample was prepared by mixing GBL with water. The fallowing approximately concentration were used: 0.10, 0.25, 0.5, 0.75, 1.0 and 1.5wt%. LA calibration samples were also prepared by missing LA with water using the fallowing concentrations: 0.25, 0.5, 0.75, 1.0, 1.5 and 2.0wt %.

Catalyst solution

The catalyst solution consisted of 0.2wt% RuCl3 dissolved in ethanol. The other catalyst solution consisted of 0.5wt% RuCl3 dissolved in GBL.

GBL and water permeability

GBL and water permeability was tested in the membrane reactor to determine which fluid is more convenient to use for the reaction. PTFE layer was placed in the membrane cell and was exposed to the liquid for approximately12 hours. Different temperature were tested, such as 60, 70 and 80C.

Solvents solutions

2.1 g of Levulinic Acid was dissolved in 70g of the fallowing solvents: water, GBL and glycerol.

Levulinic acid hydrogenation reaction procedure

PTFE layer was coated with the catalyst solution (or the PTFE cover with the polymer was coated with catalyst) and hydrogen was supply constantly for approximately 2 hours, until the fallowing reaction was completed:

2RuCl3(s) + 3H2 (g) -> 6HCl (g) + 2Ru(s)

Once Ru was the only element deposit on the PTFE layer, it was placed on the membrane cell. The reaction was carried out at 70oC for approximately 22-24 hours. Samples were collected at different hours in order to fallow the rate of the progress at which GVL was formed. Membrane developed in method 1, 2, 3, 4 and 5 were analyze in the reactor using the same procedure.


Researcher Biography

Thanks for visiting my web page!

Email: michelle.soto6@upr.edu or mcsoto6@outlook.com


Special Thanks To:

  • John P. Stanford:
  • Graduate Researcher and mentor, Department of Chemical Engineering.

    Thanks for your support and guidance.

  • Keith Rutlin:
  • Programs Administrator, Center for Sustainable Energy.

    Thanks for your help and assistance.

  • Dr. Mary E. Rezac:
  • Faculty Advisor, Department of Chemical Engineering.

    Thanks for this extraordinary experience.

    This material is based upon work supported by National Science Foundation Grant: REU Site: Summer Academy in Sustainable Bioenergy; NSF Award No.: SMA-1359082, awarded to Kansas State University.

    CSE Logo