The document summarizes research on enzymatic hydrolysis of cellulose in the ionic liquid NMMO. Key findings include: 1) A twin screw extruder was modified to produce high cellulose loading solutions of up to 13% in NMMO/H2O. Enzymatic hydrolysis reactions were then performed in the extruder. 2) Reaction conditions like cellulose loading, enzyme loading, pH, and agitation method significantly impacted sugar yields and conversion rates. The highest yield was obtained at 10% cellulose, 122 FPU/g enzymes, and acidic pH in a stirred tank reactor. 3) A stirred tank reactor provided better mixing than a shaker bath and achieved near complete conversion of cell

1. Enzymatic Hydrolysis of Cellulose in NMMO Brett RobbinsMaster’s ThesisFAMU-FSU College of EngineeringDepartment of Chemical and Biomedical Engineering

2. Traditional Biomass to Ethanol Schematic - DOE report

3. - DOE report

4. Pretreatment of Lignocellulose - DOE report

6. Strong intermolecular interaction

7. Very low vapor pressures

8. Easy to recycle due to phase separation

9. Can be modified for specific properties by changing cation or anion combination

10. Ionic properties tend to alter reaction kinetics- R. Rogers, 2002

12. Low vapor pressure solvent: Easily contained & recovered

13. Excellent solvent for dissolution of cellulose in monohydrate form

14. Environmentally friendly UN award for green process

15. Biodegradable solventDissolves cellulose by breaking the hydrogen bonds of cellulose layers - J. Collier and co-workers (2000, 2001) Commercially used as a cellulose solvent in the Lyocell Process

17. Initial reaction rates are higher in cellulose suspended and AA buffered solutions

19. Higher cellulose loadings could not be prepared in batch method mixing

20. Higher cellulose solutions produced a gelatin-like formation  Poor mixing

21. Higher reducing sugars yields desired for fermentation-S.Rama, R.Oyetunji (2009)

22. Goals of this work Objectives: * Perform in situ enzymatic hydrolysis of high cellulose loading in NMMO Modify the Extruder so that it is capable of making high loading cellulose solutions in NMMO/H2O Perform the Enzymatic Hydrolysis of cellulose using the twin screw extruder as a reactor Analyze how cellulose loading, enzyme loading, pH and agitation affect the initial reaction rate, reducing sugar yield, as well as overall conversion of cellulose to glucose

23. Approach Modify extruder in which higher cellulose loading solutions could be prepared Prepare 5, 10 and 13 wt% cellulose solutions Perform cellulose hydrolysis in extruder Does Reactive Extrusion show potential? Modify reaction conditions to maximize reducing sugar yields

25. The enzyme active site extends over 3 glucose monomer links

26. This allows for multiple breaks per a single absorption

28. Higher Cellulose Loadings Using the developed recycle system, we were able to produce 5, 10 and 13 wt% cellulose in NMMO

29. Solution Preparation NMMO near monohydrate Preparation Starting with 50/50 NMMO/H2O, evaporate water using rotovap for 5-8 hrs Cellulose/NMMO Solutions (5, 10 and 13 wt%) Dissolve measured pure cellulose in NMMO using extruder (make sure flow through recycle hose is established) Cellulose/NMMO Acetic acid Solution (no recycle system) Add NMMO/Cellulose solution to extruder Port 1 (85C for Zones 1-3) Add Acetic acid into feed port 4 (50C for Zones 4-10) Collect solution at exit and manually re-feed into extruder (repeat 5 – 8 times) For Shaker bath and STR Separate sample vials containing 10.7 g Cellulose/NMMO/AA Enzyme addition Add enzyme solution to sample following NREL FPU/g loading requirements In E + DI dilution samples, DI added directly into enzyme solution before adding to sample In E + AA and E + AA + 1 dilution samples, add AA and then enzyme solution to sample vial Take 0.5 mL samples at pre-designated time Boil samples to denature enzyme and stop reaction Determine reducing sugar concentration using DNS assay technique

30. Method of Assay: DNS 3,5-Dinitrosalicylic acid (DNS or DNSA) is an aromatic compound that reacts with reducing sugars and other reducing molecules to form 3-amino-5-nitrosalicylic acid, which absorbs light strongly at 540 nm. A reducing sugar is any sugar that, in basic solution, forms some aldehydeor ketone. Reducing sugars include glucose, fructose, glyceraldehyde, lactose, arabinose and maltose. - Miller, 1959 Increasing sugar concentration

31. Reactive Extrusion Increasing cellulose loadings yield higher reducing sugar concentrations Extruder shows potential for enzymatic hydrolysis

32. Approach Modify extruder in which higher cellulose loading solutions could be prepared Prepare 5, 10 and 13 wt% cellulose solutions Perform cellulose hydrolysis in extruder Does Reactive Extrusion show potential? Modify reaction conditions to maximize reducing sugar yields

34. Reducing Sugar yield does not accurately scale with initial cellulose concentrationIs there another way to increase conversion?

36. Long duration increases equipment cost

37. Higher cellulose loading causes foaming when air is taken up by the extruderCellulose NMMO 10% Acetic Acid Enzymes Low Density Low Viscosity Extruder Shaker Bath Higher cellulose loading solutions can be reacted in Shaker Bath

41. Also occurred in extruder

42. Low overall conversion 50-70%Possible inhibition or deactivation of enzymes

44. One order of magnitude increase in hydroxide ion concentrations

45. According to Genencor, the recommended pH range for AcceleraseTM 1000 is 4.8 to 5.2.

46. Sometime between 12 and 24 hours, high pH may cause inactivation of the enzymes

48. Samples with additional 10% acetic acid resulted in higher reducing sugar yields~2.5 mg/mL for 5 cellulose wt% ~4.5 mg/mL for 10 cellulose wt% Account for deactivation of enzymes, but high cellulose loading of 122 FPU/g may hide any substrate inhibition

50. Some inhibition occurs

52. May be due to an increase in ion release during the reaction causing enzyme deactivation earlier

53. If only the substrate inhibited enzyme activity

55. Particulate matter was seen settling on the base of the sample vial in higher Cellulose Loading samples

56. Settling may prevent enzymes in solution from reaching cellulose and cellobiose.Shaker Bath not providing Adequate Agitation

58. Reducing sugar yield does not plateau

59. High reducing sugar yields produced

61. STR provides better mixing for reactionSTR Shaker Bath Reducing Sugars (mg/mL)

62. Comparison with Literature Open Symbols are from Literature Closed Symbols are experiments Comparable to current work Literature works use ILs with regeneration, experiments combine pretreatment and hydrolysis steps

63. Conclusions ** In situ hydrolysis of high cellulose loading in NMMO/H2O solutions can be performed utilizing a twin screw extruder Modified Extruder is capable of making high cellulose loading solutions in NMMO/H2O Enzymatic Hydrolysis of cellulose can occur using a twin screw extruder as a reaction vessel Provides path to “consolidated bioprocess” Cellulose loading, enzyme loading, pH and agitation method are important factors in the initial reaction rate, reducing sugar yield, as well as overall conversion of cellulose to glucose 10wt% cellulose, 122 FPU/g, E + AA pH scheme in the STR provided the highest yield of reducing sugars (39.85 mg/mL) as well as cellulose conversion (95.55% assuming all glucose).

64. Recommendations In running the experiment in the extruder, make sure to maintain over flooded feed. This will prevent the extruder from pulling air into the extrusion system producing bubbles and reduce the pressure building capacity to push the solution through the recycle system. The hydrolysis may be run at 40 degrees Celsius, because it has been shown that higher pH is tolerated by the enzyme at low temperature. This can reduce the acid addition necessary for higher conversion. HPLC would be useful at identifying the final reducing sugar composition in solution for use in future fermentation reactions. The initial stirred tank reactor has shown to have promising results and should be investigated further.

65. Acknowledgements My Advisors, Dr Ramkrishnan and Dr. Collier for their guidance and insight in my research, without which I would not have been able to take this research the extent was able. Dr Paravastu for letting me use his laboratory and equipment to analyze my results. The FSU-FAMU College of Engineering Department of Chemical Engineering, for accepting me into the Master’s program and their financial support Bush Brothers and Sun Grant for their support of my research. Gary, Kimberly, Elizabeth and all members of the department who help with suggestions or assistance in this work.

66. Questions