Pretreating cellulosic biomass

Abstract

The present invention pertains to methods for pretreatment of cellulosic biomass for bioconversion into ethanol and other biofuels and wood-based chemicals, recycling of newsprint and other paper products, microfibrillation of cellulose for use as an additive in the food and cosmetic industries, manufacturing improved hardboard, and producing and improved “super” pulp while reducing chemical usage and spent liquor generation. In particular, the instant invention employs supercritical, critical or near critical fluids with and without polar cosolvents [critical fluid, SuperFluids or SFS] for the pretreatment of cellulosic biomass.

Description

GOVERNMENT SUPPORT

Research leading to this invention was in part funded with Grant No. 90-33610-5111 from the United States Department of Agriculture, Washington, D.C.

FIELD OF THE INVENTION

The present invention pertains to methods for pretreatment of cellulosic biomass for bioconversion into ethanol and other biofuels and wood-based chemicals, recycling of newsprint and other paper products, microfibrillation of cellulose for use as an additive in the food and cosmetic industries, manufacturing improved hardboard, and producing and improved “super” pulp while reducing chemical usage and spent liquor generation. In particular, the instant invention employs supercritical, critical or near critical fluids with and without polar cosolvents [critical fluid, SuperFluids or SFS] for the pretreatment of cellulosic biomass.

BACKGROUND OF THE INVENTION

Biomass resources are currently greatly underutilized in the United States. If effectively exploited, these resources can reduced our dependency on foreign oil while alleviating several environmental problems. In the recent state of the Union address by President Bush, both the Executive Office and the Democratic Congress concurred on the need for an alternative energy problem. Biomass resources have the potential to make a significant contribution to this program.

America's forests are prime candidates for improvement in biomass utilization. In the harvesting of a tree, approximately 40% of the biomass is either burned or treated as a waste problem. Thus, a substantial amount of wood material is not utilized for the more traditional structural uses of wood. This waste biomass can be utilized as the feedstock for fermentation processes which produce ethanol, industrial chemicals and ruminants for animal feed. The bioconversion of plant biomass is both technically and economically limited by the absence of high yield and low cost conversions technology for lignocellulosic materials. Ethanol production from plant substrates requires the development of an efficient pretreatment process for increasing the susceptibility of woody biomass to hydrolytic enzymes. Pretreated biomass substrates may then be hydrolyzed to glucose, and converted to ethanol by conventional yeast fermentation. Biomass produced ethanol can displace gasoline usage, thereby reducing oil consumption while lowering automobile-generated air pollution.

Paper waste, another underutilized biomass resources, is currently a burden on our overfull landfills. This biomass waste can also be pretreated, re-pulped and recycled; alternatively, pretreated paper waste can also be used as feedstock to bioconversion plants.

SuperFluids are a gentle and energy-efficient means of comminuting or disrupting woody biomass. SuperFluids offer several advantages as a form of wood pretreatment, and may also find utility as a form of wood pulping.

As shown in FIG. 1, a compound becomes supercritical at conditions that equal or exceed both its critical temperature and critical pressure. These parameters are intrinsic thermodynamic properties of all sufficiently stable pure compounds and mixtures. Carbon dioxide, for example, becomes a supercritical fluid at conditions that equal or exceed its critical temperature of 31.1° C. and its critical pressure of 72.8 atm (1,070 psig). In the supercritical fluid region, normally gaseous substances such as carbon dioxide become dense phase fluids that have been observed to exhibit greatly enhanced solvating power.

At a pressure of 3,000 psig (204 atm) and a temperature of 40° C., carbon dioxide has a density of approximately 0.8 g/cc and behaves much like a nonpolar organic solvent such as hexane, having a dipole moment of zero Debyes. A supercritical fluid uniquely displays a wide spectrum of solvation power, as its density is strongly dependent upon temperature and pressure. Temperature changes of tens of degrees or pressure changes by tens of atmospheres can change a compound's solubility in a supercritical fluid by an order of magnitude or more. This unique feature allows for the fine-tuning of solvation power and the fractionation of mixed solutes. The selectivity of nonpolar supercritical fluid solvents can also be enhanced by the addition of compounds known as modifiers (also referred to as entrainers or cosolvents). These modifiers are typically somewhat polar organic solvents such as acetone, ethanol and methanol.

In addition to their unique solubilization characteristics, supercritical fluids possess other physicochemical properties that add to their attractiveness as solvents. They can exhibit liquid-like density yet still retain gas-like properties of high diffusivity and low viscosity. The latter increases mass transfer rates, significantly reducing processing times. Additionally, the ultra-low surface tension of supercritical fluids allows facile penetration into microporous materials such as cellulosic biomass, increasing saturation as well as expansion and extraction efficiency and overall yields. While similar in many ways to conventional nonpolar solvents such as hexane, it is well known that these fluids can extract a different spectrum of materials than conventional techniques. Product volatilization and oxidation as well as processing time and organic solvent usage can be significantly reduced with the use of supercritical fluids.

A material at conditions that border its supercritical state will have properties that are similar to those of the substance in the supercritical state. These so-called “near critical” fluids are also useful. To simplify the terminology, materials that are utilized under conditions, which are supercritical, near critical, or exactly at their critical point with or without polar cosolvents or entrainers are jointly referred to as critical fluid, SuperFluids™ or SFS.

The basic steps of the inventive critical fluid-biomass pretreatment (CBP) process are: (a) milling of biomass into specific size and, if necessary, the controlling of moisture content; (b) preheating the milled biomass to operating temperature utilizing the heat generated from compressing the critical fluid); (c) exposing the biomass to the critical fluid to permeate the wood matrix; and (d) rapid depressurization or explosive decompression, leading to expansion of solvent within the woody biomass and its defibration. With either a critical fluid or low boiling liquid, specific volume changes of two or three orders of magnitude are attainable with relatively small changes in pressure. For example, the expansion factor (ratio of volumes at operating to standard atmospheric conditions) of saturated steam at 1,000 psig is approximately 60 whereas supercritical carbon dioxide has an expansion factor of 360 at a pressure of 1,200 psig and 32° C. Thus, supercritical CO2 at these conditions may be at least six times more effective than steam in the defibration of wood by an explosive decompression mechanism. This enhanced expansion force enables the method to become effective and applicable to many wood types.

The inventive critical fluid pretreatment process is envisioned to utilize relatively low cost or potentially no cost critical fluids such as carbon dioxide at moderate conditions of temperature, pressure and residence time in a rather simple equipment configuration. The primary aim of the proposed CBP technique is to mechanically defibrate biomass and paper substrates in order to improve the accessibility of hydrolytic enzymes to cellulose fibers. However, in addition to the critical fluid, these solvents can be tailored to break down polymeric lignin into its aromatic constituents, solvate the reaction products, and thus separate lignin from the desirable cellulose and hemicellulosic constituents of wood. Critical fluids may thus have utility as a pretreatment as well as a pulping technique through defibration as a result of explosive decompression at a lower operating temperature, and fractionation as the result of critical fluid reaction and solvation effects. The CBP process may also find utility in paper recycling processes by mechanically pretreating newsprint while chemically enhancing the deinking process.

SUMMARY OF INVENTION

The present invention improves cellulosic biomass pretreatment by using near-critical, critical and supercritical fluids with or without polar cosolvents which have the combined capability to defibrate biomass at low operating temperatures and to fractionate wood into its constituents.

Experiments were conducted with white pine and critical carbon dioxide at different conditions of time, temperature, and pressure. Several CBP-white pine experiments were also conducted with polar entrainers in SFS CO2, other critical fluids (nitrous oxide, propane, ethylene, and Freon-22) and saturated steam at 240° C. At a lower level of effort, several CBP and steam explosion pretreatment experiments were conducted with a typical hardwood—red oak, and a representative paper waste—newspaper strip of the Wall Street Journal. In order to evaluate explosive decompression of a critical fluid, several preliminary microfibrillation experiments were conducted in a critical fluid disruption apparatus. Finally, we conducted several experiments with a critical fluid solubilization apparatus to confirm that at the low temperatures utilized, critical fluids do extract lignin and/or other aromatic wood constituents.

Experimental results and analytical data indicate the critical fluid biomass pretreatment (CBP) process is technically feasible, and has several advantages over conventional steam explosion pretreatment. These conclusions are based on: (1) high conversion efficiencies (CBP was 60% more effective than steam explosion in pretreating white pine, 300% better in pretreating newsprint and just as effective in pretreating red oak); (2) biomass recovery yields (between 95 and 99%) which are much higher than steam explosion yields (often less than 80%); (3) relatively low operating conditions of temperature, pressure and time.

The primary potential application for the critical fluid pretreatment process is pretreating biomass waste for bioconversion into ethanol and other wood-based chemicals. The use of waste biomass as a cheap raw material for the production of ethanol by bioconversion processes could significantly impact the manufacturing cost of gasohol, and help meet President Bush's goals of reducing auto emissions and air pollution over the nation's largest cities by 50% within the next ten years. There are several other potential applications such as recycling of newsprint and other paper products, microfibrillation of cellulose for use as an additive in the food and cosmetic industries, manufacturing improved hardboard, and producing and improved “super” pulp while reducing chemical usage and spent liquor generation.

Experimental results and analytical data indicate the critical fluid biomass pretreatment (CBP) process is technically feasible, and has several advantages over conventional steam explosion pretreatment. These conclusions are based on the following:

• • CBP was shown to be, under certain conditions, much more effective than steam in pretreating typical softwood and a representative paper waste (about 60% more effective in pretreating white pine and more than 300% better in pretreating newsprint).
  • CBP was just as effective as conventional steam explosion in pretreating representative hardwood—red oak, improving its enzymatic hydrolysis six-fold.
  • Biomass recovery yields are very high (between 95 and 99%) for conditions which favor enzymatic hydrolysis. These yields are much higher than steam explosion biomass yields which are usually less than 80%. Steam explosion results in the solubilization of pentosans and oligomers, improving susceptibility to hyrolytic enzymes, but also reducing yield. Biomass is also lost as the result of pyrolysis gases, acetic acid from acetyl groups, formic acid, furfual and other compounds which evaporate with water.
  • The biomass loss to the critical fluid can be recovered in a commercial process and utilized to form useful byproducts (this was demonstrated by an online high pressure UV monitor).
  • The optimal residence time for CBP pretreatment, under certain conditions, was estimated to be between 1 and 5 minutes. For example, the enzymatic hydrolytic efficiency for SFS CO2 pretreated newspaper strips was more than 300% greater than untreated material for a residence time of only 1 minute at 2,000 psig and 60° C. It should be noted that this residence time can be further reduced by dynamic mixing and explosively ejecting the SFS treated biomass materials. This compares well with steam explosion for which residence times between 1 and 10 minutes have been used.
  • Lower pressures appear to work better than higher ones for the CBP process, presumable because the crystallinity of the cellulosic materials is adversely affected by high pressures. For white pine shavings, SFS CO2 was most effective at the lowest pressure (2,000 psig) tested. For newspaper strips, the enzymatic hydrolytic conversion efficiency is very favorable at 2,000 psig and very poor at 10,000 psig.
  • We anticipate that the maximum pressure requirements should be less than 3,000 psig for carbon dioxide because both effects of decompression energy and biomass salvation should reach points of diminishing returns beyond a reduced pressure of about 3.0 (refer to Figure x in which density at low temperatures asymptotically increases with pressure). Lower pressure requirements lower initial capital cost, and reduce equipment wear and operating costs.
  • Lower CBP pretreatment temperatures were more effective than higher ones, presumable because critical fluid density and explosive decompressive forces decrease with temperature. For SFS CO2 pretreated white pine shavings and cubes, the optimum operating temperature was around 60° C. For newspaper strips hydrolytic enzymatic susceptibilities decreased dramatically with increased temperature.
  • Low CBP operating temperatures are very favorable since steam explosion suffers several drawbacks—low biomass yields and high energy costs—because of the higher temperatures utilized. Additionally, steam explosion leads to loss in brightness and deterioration of fiber and paper properties (due to partial cellulose depolymerization)—factors that make steam exploded wood fibers unsuitable for paper making.
  • As a result of relatively low operating temperatures, CBP pretreatment has an additional advantage over steam explosion which requires a post-treatment washing step to remove furfural and other biomass degradation products that are inhibitory to the action of hydrolytic enzymes.
  • Critical carbon dioxide proved rather effective in pretreating woody biomass, improving the enzymatic hydrolysis efficiencies of white pine, red oak and newsprint by more than 300% at relatively low temperatures and pressures. This result was quite encouraging because CO2 is non-toxic, non-flammable, relatively environmentally benign, and readily available (CO2 will be present onsite as a byproduct of an ethanol fermentation process).
  • The enzymatic hydrolysis efficiency of white pine pretreated by critical carbon dioxide was significantly improved (approximately doubled) by the addition of 2 mole % ethanol as a polar entrainer. This is a rather convenient finding since ethanol, as well as CO2, will be available onsite at biomass to energy plants. There is however, likely to be some trade-off between using ethanol as a polar entrainer to improve CBP's pretreatment capability and selling ethanol as a product.
  • The moisture content of “green”, never-dried red oak chips improved the efficiency of the CBP pretreatment process with supercritical carbon dioxide at 10,000 psig and 200° C. This improvement could have been caused by carbon dioxide dissolving in the interstitial water, associating to form carbonic acid, and inducing mild hydrolysis. This result is quite advantageous over steam explosion because the heating of moisture originally present in the wood is potentially the greatest contributor to steam consumption, requiring more than twice as much steam as does dry wood.
  • In terms of initial capital, as well as operating and maintenance costs, the economic trade-offs are between the high temperature requirements of the steam explosion process and the high pressure requirements of the critical fluid pretreatment process. The CBP process will be operated at the minimum pressure requirements and pressure drops. While the former impacts vessel wall thickness, the latter will impact both the capital and operating costs of the compressor. Indeed, if the CBP process can be operated rather isobarically, the costs of the CBP process would be significantly impacted. Continuous flow operation would also significantly impact process economics. In a continuous flow process, the size of a CBP pipe reactor will be determined by the residence time requirements and throughput capacity. Since the residence time requirements have been shown to be low and competitive with steam explosion, capital cost of the CBP process at low pressures (less than 3,000 psig) will be competitive with that of steam explosion.

The developed CBP process would be more effective, selective, flexible, generally applicable, and energy efficient than existing processes. The real benefits of the CBP process are expected to arise from its technical edge over existing methods of pretreating biomass.

The CBP process has been demonstrated to be more effective than steam explosion in pretreating white pine and newsprint, and just as effective as steam explosion in pretreating a representative hardwood—red oak. Furthermore, the adiabatic expansion of the critical fluid will lead to a cooling of medium, thus reducing thermal and oxidative degradation of cellulose and hemicellulose and improving overall carbohydrate yield. The effectiveness of the proposed CBP process is also positively impacted by the capabilities of critical fluid solvents to remove lignin and other polymeric resins which bind the wood fibers. The accessibility of the fibers to cellulase and their susceptibility to enzymatic hydrolysis would thus be improved; also, the strength, integrity and brightness of the resulting fibers would be improved, making the CBP process a good candidate for pretreating wood and reducing downstream pulping and bleaching requirements. There are several potential applications of the critical fluid biomass pretreatment process;

• • CRITICAL FLUID PULPING to produce high quality paper is the largest and most difficult marketplace to penetrate, Critical fluids can potentially create a super fiber while reducing chemical usage and spent pulping liquor generation, minimizing the environmental impact of the pulp and paper industry on society at large.
  • MANUFACTURING OF IMPROVED HARDBOARD is quite possible since the CBP process can displace conventional chemical-mechanical techniques, and offers several advantages over steam explosion pulping. In this application, the CBP process would depend equally on the chemical salvation and mechanical defibration capabilities of critical fluid pretreatment.
  • RECYCLING NEWSPRINT AND OTHER PAPER PRODUCTS . . . . The CBP process was demonstrated to be very effective in pretreating newsprint with excellent yield (almost 100%). The possibility exists that the very apolar critical fluids can be used for the deinking of newsprint, and that this capability can be enhanced by the use of polar entrainers and fatty acid or surfactants.
  • PRETREATING BIOMASS WASTE for bioconversion into ethanol and other wood-based chemicals. The use of waste biomass as a cheap raw material for the production of ethanol by bioconversion processes could significantly impact the manufacturing cost of gasohol, and help meet President Bush's goals for reducing auto emissions and air pollution over the nation's largest cities by 50% within the next ten years.
  • MICROFIBRILLATION OF CELLUOSE for use as an additive in the food and cosmetic industries. This application constitutes a niche-market which is sufficiently large to justify evaluation. We anticipate that the CBP process would be primarily mechanical in this application

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a depiction of a supercritical fluid phase diagram.

FIG. 2 is a process flow diagram of the critical fluid pretreatment (CBP) apparatus.

FIG. 3 is a process flow diagram of the critical fluid disruption (CFD) apparatus.

FIG. 4 is a process flow diagram of critical fluid solubilization apparatus.

FIG. 5 is a process flow diagram of steam explosion pretreatment apparatus.

FIG. 6 is a glucose standard curve for a dinitrosalicylic acid (DNS) reagent assay.

FIG. 7 is a graph of the effect of residence time on the CBP of kiln-dried white pine cubes with SFS CO2 at 10,000 psig and 200° C.

FIG. 8 is a graph of the effect of pressure on the pretreatment of kiln-dried white pine cubes by supercritical carbon dioxide at 200° C. for a residence time of 15 minutes.

FIG. 9 is a graph of density of CO2 as a function of pressure.

FIG. 10 is a graph of the effect of pressure on CBP of kiln-dried white pins shavings with SFS CO2 at 200° C. for 15 minutes.

FIG. 11 is a graph of the effect of temperature on the critical fluid pretreatment of kiln-dried white pine cubes for SFS carbon dioxide at 10,000 psig for a residence time of 15 minutes.

FIG. 12 is a graph of the effect of temperature on the critical fluid pretreatment of white pine shavings for SFS carbon dioxide at 10,000 psig for a residence time of 15 minutes.

FIG. 13 is a bar chart of the effect of critical fluid type on the CBP of kiln-dried white pine cubes at 10,000 psig and 200° C. for 60 mins.

FIG. 14 is photomicrographs of pulp before (FIG. 14a) and after treatment. FIG. 14b for the SFS CO2 comminuted pulp, FIG. 14c for homogenized pulp (two passes), and FIG. 14d for SFS NO2 comminuted and homogenized pulp (one pass).

FIG. 15 is a UV trace of SFS CO2 phase in contact with white pine at 2,300 psig and 60° C. in CBP-135

FIG. 16 is a UV trace of SFS CO2 phase in contact with white pine at 4,300 psig and 60° C. in CBP-136

DETAILED DESCRIPTION

The critical fluid biomass pretreatment (CBP) apparatus used to perform the supercritical fluid pretreatment experiments is shown in FIG. 2. The apparatus was designed so that wood chips can be readily introduced and recovered after explosive decompression of the critical fluid/biomass mixture. The experimental apparatus consists of a pipe reactor with a sample container, and air driven compressor and a critical fluid recycle loop. The CBP apparatus is rated for continuous operation at 10,000 psig and 200 C. The entire apparatus, shown in FIG. 2, is contained in a steel framed, polycarbonate (Lexan) barrier that allows the operator to freely interface with the machine.

The 21-inch long CBP reactor is made of ¾″ Schedule 160, seamless Hastelloy C-276 pipe with an internal diameter of 0.612″ and a volume of 100 cc. Both end caps of the reactor can be readily removed to allow insertion of a sample container holding the biomass material. The sample container is a 0.062″ wall 316 SS tube with 150 1/16″ diameter holes drilled uniformly over its surface area. The sample container has two aluminum press fit end-caps with 10 circumferential holes and one center hole, all 1/16″ in diameter. Within the reactor, the end cap heat losses can be minimized and the sample container can be maintained at an isothermal temperature. The pipe reactor and sample container with biomass material are heated by two 600 W electric heating elements. Temperature is sensed with and controlled via a thermocouple. which is inserted into the center of the pipe reactor with the sensing element placed about ¼″ above the sample container. The type K thermocouple is connected to a microprocessor-based temperature PDI controller with a self-tuning optimizing algorithm (Schuntemann, Model RCQ 5200).

The bottom of the pipe reactor is connected to a blowdown chamber via a high pressure ball valve, PV-7, and back to the inlet of the compressor, P-1. The blowdown chamber is a 1-liter Hoke cylinder, rated for 1,800 psig at 200° C., which is connected to a 1,500 psig relief valve and vented to the atmosphere, the compressor, P-1, is an air driven liquid pump with a nominal compression ratio of 122:1 and a three way spooling valve (Haskel, Model No. DSF-122). For critical fluids which are gaseous at room temperature, the unit can be readily connected to a single-ended diaphragm compressor (Superpressure—previously Aminco, Model No. J-46-13411), which can compress a gas up to 10,000 psig at a flowrate of 40 standard liters per minute.

In a normal operating mode, a weighed sample of biomass is introduced into the sample chamber which is then inserted into the pipe reactor. The critical fluid reactor is then heated to operating temperature and pressurized with a selected critical fluid. After the system has reached a stable operating temperature and pressure, the critical fluid is recirculated through the biomass sample by closing PV-4 and opening PV-5. Several experiments were carried out in the static mode without the circulation loop activated. Polar cosolvents or entrainers can be introduced through their displacement and/or solvation into the critical fluid. This is accomplished by closing PV-4 and opening PV-3. After a pre-specified residence or soaking time, the system is blown down by opening the ball valve, PV-7, as rapidly as possible. The sample container is then removed, the sample weighed, observations recorded, and the recovered biomass is stored at 4° C. prior to further analysis.

The critical fluid disruption apparatus is shown in FIG. 3. This pilot unit is rated for continuous operation at 10,000 psig and 200° C. The CFD reactor is heated with an electric band heater rated for 1,400 watts; cooling can be accomplished with coils inside the reactor. The contents of the CFD reactor can be mixed by a magnetically driven impeller with a torque of 15-in-lb, and a maximum, no-load speed of 3600 rpm. The blowdown chamber has a capacity of 10 liters, and is located directly below the dynamic mixing chamber. The blowdown chamber contains a pressure relief valve rated at 1,500 psig. A low-pressure trap is also located in a vent line. Cleaning solutions are supplied by the wash pump (P-2); the cleaning solvents include 0.1N NaOH, a 10% Clorox solution, and 95% methanol.

The critical fluid solubilization apparatus is shown in FIG. 4. The top part of the apparatus is a high-pressure critical fluid recirculation loop which is rated for continuous operation at 5,000 psig and 100° C. This high-pressure loop consists of a supply or soaking chamber (SCH), a circulation pump P-2, a solids chamber and several valves and connecting lines. P-2 is a variable speed (0 to 9,000 rmp), a high pressure (5,000 psig at 150° C.) gear pump capable of a flowrate of 300 ml/min at a pressure differential of 10 to 20 psig; the soaking chamber is a Hoke cylinder with a volumetric capacity of 150 ml; and the solids chamber is a reversed flow, in-line filter with a 90 micron stainless steel (SS) filter cup. The high pressure loop also has an online, high pressure UV/VIS detector (Isco, Model V-4). The high pressure loop interfaces with a low pressure half of the apparatus which is made up of a 500 ml decompression chamber (DC), two back pressure regulators (BPR-1 and BPR-2), a low pressure trap (LPT-1), and several two way valves and connecting lines.

The steam explosion apparatus is show schematically in FIG. 5. Water is fed with a high pressure, positive displacement, variable stroke pump to the pipe reactor and sample chamber via a 36 KW radiant heater. A PID controller is used to turn this radiant heater on/off, depending on the temperature sensed in the reactor by a thermocouple in the CBP reactor. The steam is condensed downstream of the reactor by cooling water (CW) and its pressure controlled by a back pressure regulator (BPR). Blowdown is achieved by rapidly opening a full bore ball valve.

In CBP experiments, the critical fluid was allowed to permeate the woody biomass matrix for a defined residence time at a pre-specified temperature and pressure. The critical fluid mixture was then rapidly decompressed, leading to the expansion of the solvent within the woody biomass and its defibration. Decompression, achieved by rapidly opening a full bore ball valve, occurred in approximately 1 second for most of the conditions tested. Decompression was thus rapid but not instantaneous. Also as previously discussed, the biomass material was retained in a perforated sample chamber, which allowed for the rapid exhaustion of the critical fluid but not for the physical disengagement of the biomass material.

As direct consequence of our experimental design, CBP effects are primarily the result of the internal expansion of the critical fluid within in the pores of the woody biomass matrix. All CBP treated biomass samples did in fact retain their general morphology, shape and size; there was no evidence of physical disruption to the naked eye. Defibration and enzymatic susceptibility can be further enhanced by ejecting the critical fluid saturated biomass out of the CBP reactor, and impinging the ejected material onto a series of disruption bars or through a disc refiner.

One hundred and thirty-six (136) critical fluid biomass pretreatment (CBP) experiments were conducted. Slightly more than 50% of these experiments were conducted with white pine shavings, cubes and chips. Most of these experiments were conducted with supercritical carbon dioxide at different conditions of time, temperature and pressure. Results are thus biased towards the pretreatment of a softwood with SFS CO2. Several CBP-white pine experiments were also conducted with polar entrainers in SFS CO2, other critical fluids (nitrous oxide, propane, ethylene, and Freon-22) and saturated steam at 240° C.

At a lower level of effort, several CBP and steam explosion pretreatment experiments were conducted with a typical hardwood—red oak, and a representative paper waste—newspaper strips of the Wall Street Journal. In order to evaluate explosive decompression of a critical fluid, several preliminary microfibrillation experiments were conducted in a critical fluid disruption apparatus. Finally, we conducted several experiments with a critical fluid solubilization apparatus to confirm that at low temperatures utilized, the critical fluid does extract lignin and/or other aromatic wood constituents.

In the examples to follow, the enzymatic conversion efficiencies of untreated biomass raw materials are first presented because they are considered base cases against which the critical fluid biomass pretreatment (CBP) and conventional steam explosion processes can be compared. In order to evaluate the CBP process, we investigated the influence of several process variables and material characteristics on biomass recovery yields and enzymatic conversion efficiencies. Several of these process parameters, such as: (i) residence time; (ii) pressure; (iii) temperature; (iv) critical fluid type; (v) polar co-solvent or entrainer; (vi) biomass type; and (vii) moisture content are discussed in the examples below. Thereafter, CBP results are compared to steam explosion, critical fluid pulping and solvation experiments are presented, and conclusions are drawn.

EXAMPLES

Example 1

Biomass Raw Materials

Representative hardwood (red oak), soft wood (white pine) and waste paper (newsprint) were tested. Kiln-dried, “green” oven dried and “green” never-dried red oak and white pine shavings, cubes and chips were tested. White pine and red oak cubes were prepared by cutting kiln-dried boards with an 8″ Craftsman table saw ripped with the grain; these strips were then cross-cut with a Delta bandsaw to produce cubes which are approximately 6 mm on each side. Shavings (approximately 25 mm×6 mm×3 mm) of white pine and red oak were prepared by planting kiln-dried boards on a 6″ Delta joint-planer. “Green” never-dried white pine and red oak chips were prepared making transverse cuts across the diameter of a log from a freshly felled tree. The transverse cuts were made with a newly sharpened chain saw blade to produce thin chips approximately 10 mm×6 mm×1 mm. The “green” never-dried chips were prepared with the courtesy of Professor Robert P. Ordway, Loudon, N.H. Newspaper strips (6 mm to 10 mm wide) were cut from the Wall Street Journal.

Example 2

Analytical Techniques

The effectiveness of critical fluid pretreatment as a pretreatment for woody biomass was evaluated from the efficiency with which that biomass can be converted into reducing sugars by enzymatic hydrolysis. This analytical technique was selected as a key “acid-test” method because enzymatic effectiveness is very dependent on the surface area of cellulose fibers that can be accessed by hydrolytic enzymes.

The enzymatic hydrolysis assay was conducted with two enzymes—cellulase from a selected strain of the fungus Trichoderma reesei (Celluclast 1.5 L from Novo Biolabs, Danbury, Conn.) and cellobiase from a selected strain of Aspergillus niger (Novozyme 188, also from Novo Biolabs). A mixture of enzymes is required to first convert insoluble, crystalline cellulose into a soluble reactive form, then to convert the soluble cellulose into cellobiose and reducing sugars such as glucose, and finally the cellobiose to glucose.

Enzymatic hydrolysis efficiency is a function of enzyme mixture and activity, temperature, pH and time. The temperature and pH were optimally fixed at 50° C. and 4.8; the hydrolysis time is fixed at three hours to allow significant conversion within a reasonable time frame. The activity of the enzyme mixture was fixed at 80 International Units (IU)/g of dry substrate. The activity of a 10:1 mixture of cellulose:cellobiose was determined by a filter paper (FP) assay.

In the FP assay, a 9.09 dilution of the enzyme mix in a 0.1 M sodium nitrate buffer (pH=4.8) was added to 50 mg of Whatman No. 1 ashless filter paper, vortex mixed and incubated for one hour at 50° C. One mL of the supernatant solution was added to three mL of dinitrosalicylic acid (DNS) reagent, and placed in boiling water for five minutes. After boiling, the resulting solution is diluted with 16 mL of distilled deionized (DDI) water, and absorbance at 550 n1 is measured against a blank in order to negate the effect of any residual sugars in the enzyme mix. The DNS reagent was made up of 1 wt % dinitrosalicylic acid, 0.2 wt % phenol, 0.05 wt % sodium sulfite and 1 wt % sodium hydroxide. The 3,5-dinitrosalicylic acid is reduced by glucose and other reducing sugars to 3-amino-3-nitrosalicylic acid while the aldehyde groups are oxidized to colored carboxy groups which can be measured at 550 nm. Phenol is used to increase the amount of color present while the bisulfite serves to remove any oxygen present and stabilize the color produced. Alkali is necessary for the reducing action of sugars on DNS; the activity of the enzyme mix, which is optimum at a pH of 4.8 to 5.0, falls off rapidly at pHs less than 4.5 and greater than 6.0.

The amount of reducing sugar produced from enzymatic hydrolysis of 50 mg of Whatman No. 1 filter paper was determined to be 1.71 mg from a glucose standard curve, shown as FIG. 6, established for the DNS reagent used in both the filter paper and the cellulose enzymatic hydrolysis assays. The enzyme mix used has a FP activity of 1.71, which is equivalent to 0.32 IU/ml for the 1:9.09 diluted enzyme mixture. The 10:1 mixture of cellulase:cellobiase was thus determined to have a FP activity of 15.54 and 2.87 IU/ml.

The enzymatic hydrolysis assay was conducted by adding a sufficient quantity of diluted enzyme mixture to around 0.5 g of substrate to achieve and activity ratio of 80 I.U. per gram of dry biomass substrate. The mixture was then incubated at an oscillation speed between 140 and 160 rpm (sufficient for good mixing but not high enough to cause enzyme denaturation as could be seen by foaming) for 3 hours in a water bath at 50° C. After incubation, the mixture was centrifuged for 10 minutes at 10,000 rpm or 30 minutes at 3,000 rpm to separate out unreacted biomass matrix. The supernatant (approximately diluted) was then analyzed for reducing sugar content using the DNS reagent technique described above.

Microscopic observations of cellulose fibers were made with a trinocular phase contrast compound microscope with 15× eyepieces, and 100× and 40× phase lenses (Alphaphot YS Series manufactured by Nikon, Tokyo, Japan). Samples were stained with safranin dye to facilitate microscopic observations. Photomicrographs were taken with a 35 mm camera, Model No. 4004 AF manufactured by Nikon.

Example 3

Untreated Biomass Raw Materials

The average enzymatic conversion efficiency of untreated white pine at the end of three hours was determined to be 3.84%+0.50% as shown in Table 1.

The results in Table 1 are based on the Trichoderma reesi cellulose assay. This assay was tuned to correct several inaccuracies, such as time-sensitive blanks, unearthed during our measurements. The enzymatic conversion efficiency of 3.84% is averaged from four of the six values listed in Table 1 (each of these six values was measured in duplicate; the highest and lowest of these values were thrown out). Note that the white pine average in Table 1 is based on biomass materials of widely different shapes and sizes with a broad range of moisture content. The enzymatic conversion efficiencies have all been corrected for moisture content, i.e. they are on a dry weight basis. It should also be noted that the average enzymatic conversion efficiency is very similar to a 5% yield obtained for untreated wheat straw with an enzyme loading of 80 IU/g after three hours of hydrolysis at 50° C. The enzymatic conversion efficiencies of several other biomass materials are listed in Table-2.

TABLE 1
ENZYMATIC HYDROLYSIS CONVERSION EFFICIENCIES OF
UNTREATED WHITE PINE
	MOISTURE	CONVERSION
	CONTENT	EFFICIENCY
BIOMASS TYPE	(%)	(%)
Kiln-Dried White Pine Shavings	6.02	5.60
Kiln-Dried White Pine Shavings	6.02	3.86
Never-Dried Green White Pine	63.10	3.35
Chips
Kiln-Dried White Pine Cubes	8.87	3.97
Kiln-Dried White Pine Cubes	8.87	4.19
Kiln-Dried White Pine Cubes	8.87	2.87
White Pine Average		3.84
(high and low discarded)

TALBE 2
ENZYMATIC HYROLYSIS CONVERSION EFFICIENCIES OF
VARIOUS UNTRETAEDBIOMASS RAW MATERIALS
	MOISTURE	CONVERSION
	CONTENT	EFFICIENCY (Ec)
BIOMASS TYPE	(%)	(%)
Kiln-Dried Red Oak Cubes	8.70	3.30
Oven-Dried “Green” Red Oak	2.01	3.53
Chips
Fresh-Cut Red Oak Chips	36.32	6.40
Newspaper Strips	5.84	15.39

The enzymatic conversion efficiency of newspaper strips appears quite high for an incubation time of three (3) hours. This conversion efficiency is similar to the 17% saccharification efficiency obtained for ball milled cellulose pulp in three (3) hours of hydrolysis at 50° C. Ball milling is considered one of the most effective “pretreatment” processes for woody biomass. This similarity suggests that, not surprisingly, the newspaper strips may have cellulose fibers that are very accessible to the hydrolytic enzymes.

Example 2

Effect of Residence Time on CBP

Residence time is a major operating parameter since it impacts process equipment size and process economics. The effect of residence time on the CBP of kiln-dried white pine cubes by supercritical carbon dioxide (SFS CO₂) at 10,000 psig and 200° C. is shown in FIG. 7 and detailed in Table 3.

The experiments listed were conducted with the apparatus shown in FIG. 7 in a non-recycle mode. The listed temperatures and pressures are time-averaged over each run's duration. The biomass recovery efficiency or yield, E_b, is corrected for moisture content before and after ever CBP experiment (for the experiments listed in Table 3, the average moisture content was reduced from 8.87% before to 2.80% after the experiments). The enzymatic hydrolysis conversion efficiency, Ec, is based on the dry weight of biomass recovered after CBP, i.e. Ec is corrected for moisture content and biomass material loss. These qualification, unless otherwise noted, are true for all the experimental data presented in this document.

The data in FIG. 7 indicates that an optimal residence time may lie between 1 and 5 minutes for the conditions listed in Table 3. This residence time is very attractive in that both the supercritical fluid and the biomass were stagnant during the saturation or residence time, and the CBP process was not optimized in terms of pressure, temperature and other operating variables. This residence time compares well with that of steam explosion for which residence times between 1 and 10 minutes have been used. While these experiments were conducted with sizeable cubes (approx. 6 mm of a side), biomass materials in commercial scale operations may be large and less uniform (23 mm long in the fiber direction, 19 mm wide and 4.3 mm thick).

TABLE 3
EFFECT OF RESIDENCE TIME ON CBP OF
KILN-DRIED WHITE PINE CUBES
(with SFS CO2 at 10,000 psig and 200° C.)
	PRESSURE	TEMP	TIME	Eb	Ec
RUN NO.	(psig)	(° C.)	(mins)	(%)	(%)
CBP-44	10,000	221	1	93.40	5.93
CBP-42	10,000	191	5	92.91	7.25
CBP-43	10,000	211	10	92.71	1.70
CBP-39	10,000	186	15	94.91	4.95
CBP-40	10,000	194	30	95.49	7.25
CBP-41	10,000	198	45	94.72	4.02
CBP-21	10,017	201	60	97.05	3.63
CBP-45	10,000	200	120	93.60	4.69

Optimum residence time will be governed by the rate of mass transfer into the biomass matrix. This rate would in turn depend on the diffusivity and interfacial tension of the critical fluid, as well as the pore size distribution and moisture content of the woody biomass. 6 mm thick wet chips are too large to be reacted uniformly in the times proposed for 250° C. and slow heating of the large green chips will result in uneven cooking at the higher temperature. Slow heat transfer into green aspenwood specimens heated in a steam gun has accordingly been reported earlier. The penetration of steam into moisture laden green chips is apparently restricted by capillary forces, and delayed by the rate at which interstitial water can be vaporized; the latter would be determined by conductive heat transfer coefficients which are low for lignocellulosic materials. In this regard, CBP would have another advantage in that critical fluids would dry out moisture laden biomass and improve saturation efficiency while steam explosion increases moisture content and decreases saturation efficiency.

CBP has an additional advantage in that its residence time for effective biomass saturation is not determined or restricted by thermal decomposition and pyrolysis rates. Residence time requirements can be further reduced by dynamic mixing and explosively ejecting the SFS treated biomass materials (in our experiments the supercritical fluid was rapidly decompressed but the biomass was retained within the confines of a sample chamber for ease of material recovery).

Example 3

Effect of Pressure on CBP

The effect of pressure on the pretreatment of kiln-dried white pine cubes by supercritical carbon dioxide at 200° C. for a residence time of 15 minutes is shown in FIG. 8 and detailed in Table 4.

Pressure and pressure drop are important variables in the critical fluid pretreatment process. Higher pressures are expected to drive more gases into the cells, and increase the pretreatment efficiency by increasing the PV work on decompression. Greater pressure drops on decompression should also improve defibration efficiency and enzymatic hydrolysis susceptibility.

Pressure and pressure drop were thus seen as very important variables in the CBP process and consequently, among the first to be tested to prove process feasibility. Pressure drop is also seen to be an important economic variable in that PV work can be conserved and recompression energy minimized by limiting the size of the pressure swing. From a practical standpoint, pressure drop may be limited by the need to avoid freezing during the expansion step. All experiments were conducted with maximum pressure drops—from operating pressure to atmospheric pressure; the reported pressures in psig are thus equivalent to pressure drops in psia.

TABLE 4
EFFECT OF PRESSURE ON CBP OF
KILN-DRIED WHITE PINE CUBES
(with SFS CO2 at 200° C. for 15 mins)
	PRESSURE	TEMP.	TIME	Eb	Ec
RUN NO.	(psig)	(C.)	(mins)	(%)	(%)
CBP-29	2,000	198	15	94.81	2.39
CBP-30	4,000	203	15	94.47	4.74
CBP-31	6,000	197	15	92.38	2.96
CBP-32	8,100	230	15	91.75	2.42
CBP-33	10,000	186	15	91.30	1.08

Surprisingly, the data in FIG. 8 suggests that the conversion efficiency peaks at around 4,000 psig for SFS CO₂ and white pine cubes. This data appears to be contrary to expectations since we anticipated that explosive pretreatment forces would increase with density and, of course, the pressure of the critical fluid as shown in FIG. 9. It should be noted that higher pressures would increase acidity, which could impact crystallinity and enzymatic conversion efficiency. There also exists the possibility that, at the high pressure tested, temperature adversely affects crystallinity and increases its resistance to hydrolytic digestion. A similar effect of pressure is shown in FIG. 10 and detailed in Table 5 for SFS CO₂ contacting white pine shavings at 200° C.

TABLE 5
EFFECT OF PRESSURE ON CBP OF
KILN-DRIED WHITE PINE SHAVINGS
(with SFS CO2 at 200° C. for 15 mins)
	PRESSURE	TEMP.	TIME	Eb	Ec
RUN NO.	(psig)	(C.)	(mins)	(%)	(%)
CBP-17	2,000	198	15	91.18	11.10
CBP-18	4,047	202	15	88.25	8.18
CBP-19	6,083	201	15	88.75	5.25
CBP-20	8,000	211	15	89.48	2.75

The data shown in FIG. 10 suggests that SFS CO₂ was most effective at the lowest pressures (2,000 psig) tested, and that SFS CO2 was almost three times more effective in pretreating white pine shavings than in pretreating white pine cubes. The former supports our thesis that pressure may be adversely impacting cellulose crystallinity whereas the latter suggests that sample preparation may positively impact the CBP process.

Example 4

Effect of Temperature on CBP

Temperature is another important process variable since the CBP process was developed as an alternative to the steam explosion pretreatment of woody biomass. Steam explosion suffers several drawbacks—such as low biomass yields and high energy costs—because of the high temperatures utilized. In general, biomass pretreatment processes including critical fluid pretreatment should be more favorable at low temperatures.

The effect of temperature on the critical fluid pretreatment of kiln-dried white pine cubes is shown in FIG. 11 and detailed in Table 6 for supercritical carbon dioxide at 10,000 psig for a residence time of 15 minutes.

TABLE 6
EFFECT OF TEMPERATURE ON DVD OF
KILN-DRIED WHITE PINE CUBES
(with SFS CO2 at 10,000 psig for 15 mins)
RUN	PRESSURE	TEMP.	DENSITY	TIME	Eb	Ec
NO.	(psig)	(° C.)	(gm/cc)	(mins)	(%)	(%)
CBP-97	10,000	26	1.08	15	100.00	3.74
CBP-34	10,000	54	1.01	15	95.56	9.72
CBP-35	10,000	60	1.00	15	97.21	10.84
CBP-36	10,000	111	0.87	15	96.95	4.11
CBP-37	10,000	146	0.79	15	96.06	3.09
CBP-38	10,000	170	0.76	15	91.31	5.71
CBP-33	10,000	186	0.73	15	91.30	1.08

In terms of enzymatic conversion efficiencies, the data in FIG. 11 indicates that the CBP process may work better at lower temperatures and appear to be an optimum at 60° C. for SFS CO₂ at the high pressures (10,000 psig) tested. The same is true for the uncorrected enzymatic conversion efficiency (the enzymatic conversion efficiency can be readily decoupled from the biomass yield by dividing E_c by E_b/100).

As shown in Table 6, SFS CO2 at 10,000 psig and 60° C. for 15 minutes, increased the enzymatic hydrolysis efficiency by almost 300% over the base case value 3.84%. It should also be noted that, with the exception of CBP-97, which was apparently unaffected by SFS CO2 at 26° C. and 10,000 psig, the biomass yield was also highest (97.21%) at 60° C. suggesting that SFS CO2 was least reactive with white pine at these conditions.

The impact of temperature on enzymatic conversion efficiencies could have been the result of reduced explosive pretreatment since the density of SFS CO2 decreased with increased temperature. As shown in Table 6, there appears to be some correlation between Ec and density at temperatures above the critical point of CO2, which is 31° C. It is, however, quite probable that hydrolytic and decomposition reactions occur at the higher temperatures, producing reaction byproducts that are inhibitory to hydrolytic enzymes. A similar effect of temperature on white pine shavings by SFS CO2 is shown in FIG. 12 and detailed in Table 7.

TABLE 7
EFFECT OF TEMPERATURE ON CBP OF
KILN-DRIED WHITE PINE SHAVINGS
(with SFS CO2 at 10,000 psig for 15 mins)
	PRESSURE	TEMP.	TIME	Eb	Ec
RUN NO.	(psig)	(° C.)	(mins)	(%)	(%)
CBP-12	10,000	53	15	92.58	3.26
CBP-13	10,000	61	15	93.13	12.51
CBP-15	10,000	105	15	9067	5.06
CBP-14	10,000	158	15	87.47	7.66
CBP-16	10,000	183	15	88.38	3.90

Again the highest biomass yield and hydrolytic conversion efficiency are obtained at a temperature of 60° C. The conversion efficiency is slightly higher (less than 1%), and the biomass yield is lower than those obtained for white pine cubes under similar conditions of pressure, temperature and time, suggesting the possibility that sample preparation could impact the CBP process.

Note that, as shown in Table 6, the biomass loss increases rapidly at temperatures above critical temperature of carbon dioxide which is, 31.0° C.

Example 5

Effect of Critical Fluid Type on Biomass CBP

Most CBP experiments were conducted with supercritical carbon dioxide. This critical fluid is a natural for this process because it is readily available and may be present onsite as a byproduct of an ethanol fermentation process. Supercritical carbon dioxide (CO2), with a near-ambient critical temperature, is the least expensive of all the SFSs to be tested. CO2 is non-toxic and non-flammable and, with the exception of contributing to the “greenhouse effect”, is environmentally benign. Perhaps the greatest asset and also potential drawback to the use of CO2 in an aqueous medium is the formation of carbonic acid. Acid hydrolysis is a well-known pretreatment process for woody biomass. Several critical fluids are compared to SFS CO2 in FIG. 13 and Table 8.

TABLE 8
EFFECT OF CRITICAL FLUID TYPE ON CBP
OF KILN-DRIED WHITE PINE CUBES
(SFS at 10,000 psig and 200° C. for 60 mins)
RUN		PRESSURE	TEMP.	TIME	Eb	Ec
NO.	CF	(psig)	(C.)	(mins)	(%)	(%)
CBP-21	CO2	10,017	201	60	95.24	3.56
CBP-22	N2O	10,025	201	60	95.24	5.03
CBP-23	C2H6	10,167	200	60	97.34	6.38
CBP-24	C3H8	10,044	201	60	98.12	1.69
CBP-25	Fr-22	10,333	210	60	96.73	5.33

It should be noted that the test conditions in Table 8, as revealed by the foregoing discussions in Examples 2, 3 and 4, are very non-optimal. Consequently, the enzymatic conversion efficiency for SFS CO2 treated white pine cubes (3.56%) is about the same as that for untreated white pine cubes (3.84%). We have assumed that these conditions are relatively non-optimal for the other critical fluids tested, and thus provide a convenient state at which the critical fluids can be compared. It would, however, be best to compare the effectiveness of critical fluids at conditions that are optimum for each critical fluid.

SFS N2O causes a weight loss very similar to that of SFS CO2. Both these fluids have similar molecular weights and critical properties. As such, they have similar densities at elevated pressures and temperatures. The similarity of biomass loss between SFS N2O and CO2 suggests that the solvation capacity of N2O at these conditions was more impacted by its dense phase behavior than by its polarity (N2O has a dipole moment of 0.2 Debyes whereas CO2 is nonpolar having a dipole moment of zero Debyes). Supercritical nitrous oxide, however, results in a conversion efficiency that is about 40% higher than that of supercritical carbon dioxide; this result suggests that its polarity may positively impact its pretreatment capability.

In a preliminary set of “scoping” experiments at similar conditions of time, temperature and pressure, supercritical propane (C3H8) was most reactive with white pine shavings resulting in a weight loss of 13.6%. Propane, which is slightly polar with a dipole moment of 0.084 Debyes, most likely extracted some wood based fatty acids. As shown in FIG. 13, n-propane was the least effective of the critical fluids tested with white pine cubes; this data suggests that CBP-24 with n-propane was severely mass transfer limited.

Supercritical ethane, with a chemical structure similar to propane, was the most effective critical fluid tested at 10,000 psig and 200° C. for 60 minutes. As shown in Table 8, supercritical ethane was approximately 80% more effective than SFS CO2 and almost four times more effective than supercritical propane. Supercritical ethane also caused a weight loss of 2.7%, which is about 50% greater than that caused by supercritical propane at the same conditions. These results suggest that ethane, which is a smaller molecule than propane, may not have been mass transfer limited.

Chlorofluorocarbons, which have exhibited promise in other critical fluid applications, may not be very applicable due to their widely accepted role in ozone destruction and the resultant trend in production phase-out. Freon-22 was, however, tested in order to evaluate the impact of dipole moment on wood solvation by critical fluids. Freon-22, chlorodifluoromethane (CHC1F2), has a dipole moment of 1.4 Debyes, which is somewhat less than that of water and a critical temperature (96.0° C.), which is very close to that of propane. At room temperature and atmospheric pressure, chlorodifluoromethane is a colorless, non-flammable, non-toxic gas. The wood solvation results (3.3 wt %) are moderate and less than that of supercritical dioxide. This result and that of nitrous oxide, discussed above, indicates that polarity does not play a significant role in the solvation of white pine constituents.

Ethylene (C2H4) was used to evaluate the effect of polarity on the CBP process since ethylene, like CO2, has a dipole moment of 0.0 Debyes. This critical fluid can also be used to evaluate this importance of the supercritical temperature effect because its Tc is about 20° C. lower than N2O and CO2. Unless it imparted some particular advantage, C2H4 would not be the solvent of choice because it is expensive and flammable. Examples of such advantages would be higher product recoveries at lower temperatures, less thermal denaturation, or changes in selectivity of extraction. Thus, the impact of operating temperature on biomass constituents may be more important to the solvation of wood by critical fluids than the actual critical temperature. At the high pressure and temperature tested in Table 8, supercritical ethylene polymerized into polyethylene causing the sample to gain weight after critical fluid pretreatment.

Example 6

Effect of Polar Entrainers on Biomass CBP

Polar cosolvents or entrainers have the potential to significantly improve enzymatic hydrolysis conversion efficiencies since they can be used in conjunction with critical fluids to selectively solubilize, weaken or separate lignin and other binding polymers from the hemi-cellulose and cellulose in woody biomass.

Most critical fluid solvents such as carbon dioxide are high volatility solvents that are relatively apolar. The solubility of polar compounds can be drastically enhanced by using low volatility polar cosolvents such as alcohols as entraining agents. The effect of three polar entrainers—ethanol, methanol and acetone—on the critical carbon dioxide pretreatment of white pine cubes is listed in Table 9.

TABLE 9
EFFECT OF POLAR ENTRAINERS ON CBP OF
KILN-DRIED WHITE PINE CUBES
(with critical carbon dioxide)
RUN	EN-	MOLE	PRESS.	TEMP.	TIME	Eb	Ec
NO.	TRAINER	(%)	(psig)	(C.)	(mins)	(%)	(%)
CBP-54	None	0.0	3,719	99	60	86.24	3.87
CBP-55	Methanol	2.6	7,080	94	60	94.42	14.68
CBP-	Methanol	2.4	5,440	53	15	96.29	6.05
101
CBP-99	Ethanol	1.7	4,630	47	15	95.47	22.20
CBP-	Acetone	1.3	5,300	51	15	98.94	4.24
103

The CBP apparatus, shown in FIG. 2, was operated in the full critical fluid phase recycle mode for all experiments conducted with polar entrainers (and controls such as CBP-54 in Table 9) in order to ensure uniform composition throughout the CBP reactor.

The data in Table 9 suggests that ethanol may be the most effective polar entrainer in improving the enzymatic hydrolysis efficiency of white pine. This is a rather convenient finding for the CBP process since ethanol as well as CO2 will be available onsite at biomass to energy plants. While CO2 is a waste product that can be utilized in a CBP pretreatment process, ethanol is a very valuable product so that its use must be minimized. There thus is likely to be some trade-off between using ethanol as a polar entrainer to improve CBP's pretreatment capability and selling ethanol as a product. From the data shown in Table 9, which shows a 100% improvement over the next best Ec reported for CBP-13 in Table 7, the trade-off would likely support the use of a polar entrainer. Quite contrary to its impact on SFS CO2, ethanol does not seem to have any influence on the effectiveness of critical nitrous oxide as show in Table 10.

TABLE 10
EFFECT OF POLAR ENTRAINERS ON CBP OF
KILN-DRIED WHITE PINE CUBES
(with critical nitrous oxide)
RUN		PRESSURE	TEMP.	TIME	Eb	Ec
NO.	ENTRAINER	(psig)	(C.)	(mins)	(%)	(%)
CBP-51	None	4,865	81	60	93.07	5.37
CBP-52	None	4,865	84	60	95.61	6.09
CBP-53	Ethanol	5,936	85	60	92.73	6.13

Example 7

Effect of Biomass Type on CBP

While most of our CBP experiments were conducted with white pine, a typical softwood, we conducted a preliminary evaluation of the applicability of CBP as a pretreatment process for typical hardwood (red oak), and a representative paper waste (newsprint). We also tested the CBP process for wet and dry versions of these “green” red oak chips in order to evaluate the impact of moisture content on the CBP process. The results of this investigation are presented below.

Typical Hardwood—Red Oak: The effect of several parameters on the critical fluid pretreatment of oven-dried “green” red oak chips is shown in Table 11 for supercritical carbon dioxide.

The data in Table 11 (with the exception of CBP-63 and CBP-66) suggests that the CBP process is rather effective against red oak, increasing its enzymatic conversion efficiency by about 300% from the base value of 3.53% for untreated red oak. Interestingly, the CBP process had an equivalent effect on white pine at some of the more favorable conditions tested (c.f. CBP-35 in Table 6 and CBP-13 in Table 7). The results at lower temperature and pressures in Table 11 (namely CBP-64 and CBP-98) are encouraging in that they suggest that CBP displays similar parametric characteristics with a typical softwood and a typical hardwood. As with white pine, supercritical carbon dioxide appears to perform better than supercritical propane in pretreating red oak as shown in Table 12.

TABLE 11
EFFECT OF SEVERAL PARAMERTS OF CBP WITH
CARBON DIOXIDE OF OVEN-
DRIED “GREEN” RED OAK CHIPS
	PRESSURE	TEMP.	TIME	Eb	Ec
RUN NO.	(psig)	(C.)	(mins)	(%)	(%)
Untreated	—	—	—	100.00	3.53
CBP-63	2,500	59	1	97.38	0.00
CBP-64	2,730	61	15	94.47	11.09
CBP-98	4,170	59	15	97.61	11.46
CBP-65	5,980	58	15	97.35	9.12
CBP-66	2,500	198	15	95.01	2.49
CBP-68	10,000	57	15	88.87	9.94
CBP-69	10,290	193	15	92.91	11.85
CBP-26	10,000	192	60	93.20	14.94

TABLE 12
EFFECT OF CRITICAL FLUID ON CBP OF
OVEN-DRIED “GREEN” RED OAK
CHIPS (with SFS at 10,000 psig and 200° C.)
RUN		PRESSURE	TEMP.	TIME	Eb	Ec
NO.	CF	(psig)	(C.)	(mins)	(%)	(%)
CBP-26	CO2	10,000	192	60	93.20	14.94
CBP-28	C3H8	10.050	199	60	95.56	12.92

Preliminary results with polar entrainers indicate that ethanol and methanol in supercritical carbon dioxide may have had an adverse impact on the CBP of kiln-dried red oak cubes as shown in Table 13

TABLE 13
EFFECT OF POLAR ENTRAINERS OF CBP OF
KILN-DRIED RED OAK CUBES
(with CO2 around 5,000 psig and 60° C.)
RUN	EN-	PRESSURE	TEMP.	TIME	Eb	Ec
NO.	TRAINER	(psig)	(° C.)	(mins)	(%)	(%)
CBP-100	EtOH	5,290	59	15	100.57	8.58
CBP-102	MeOh	4,710	45	15	98.33	7.78

It should be noted that the experiments listed in Table 13 were conducted at conditions that may or may not have been favorable for the CBP of red oak.

The impact of the initial moisture content on CBP of “green” red oak chips is shown in Table 14.

TABLE 14
EFFECT OF MOISTURE CONTENT ON CBP OF
“GREEN” RED OAK CHIPS
(with CO2 around 10,000 psig and 200 C.)
RUN	MOISTURE	PRESSURE	TEMP.	TIME	Eb	Ec
NO.	(%)	(psig)	(° C.)	(mins)	(%)	(%)
CBP-26	2.01	10,000	192	60	93.20	14.94
CBP-27	35.77	10,027	196	60	94.26	17.09

The moisture content of the oven-dried “green” red oak chips in CBP-26 increased from 2.01% before the experiment to 5.49% after the experiment whereas the moisture content in CBP-27 decreased from 35.77% to 6.84% after critical fluid pretreatment. It appears from the experiments listed in Table 14 that moisture content had no impact on biomass material loss and a positive impact on the critical fluid pretreatment of “green” red oak. The latter could have been caused by carbon dioxide dissolving in the interstitial water, associating to form carbonic acid, and inducing mild hydrolysis. This very exciting result that will be a function of pressure, temperature and time as well as biomass type and moisture content.

Representative Paper Waste—Newsprint: The effect of several parameters on the critical fluid pretreatment of newspaper strips is shown Table 15 for supercritical carbon dioxide. In general, the data indicates that the CBP process is more effective at lower temperatures and pressure, and that relatively mild conditions (2,000 psig and 57° C. for 1 minute in CBP-57) are sufficient to pretreat newsprint.

TABLE 15
EFFECT OF SEVERAL PARAMETERS OF
CBP OF NEWSPAPER STRIPS
(with critical carbon dioxide)
	PRESSURE	TEMP.	TIME	Eb	Ec
RUN NO.	(psig)	(° C.)	(mins)	(%)	(%)
Untreated	—	—	—	100.00	15.39
CBP-57	2,000	57	1	97.24	48.08
CBP-58	2,640	61	15	98.17	40.38
CBP-59	10,000	64	15	98.14	23.78
CBP-60	6,000	60	15	99.35	50.14
CBP-61	2,600	184	15	95.25	19.22
CBP-62	10,000	189	15	94.96	5.60

Example 8

Comparison of Steam Explosion and CBP Pretreaments

For comparative purposes, steam explosion experiments were conducted with saturated steam at 450 psig and 240° C. in the apparatus shown as FIG. 5. Exact comparison with literature reported data is difficult because the data is usually presented for complete hydrolysis of the biomass while our data is for only 3 hours of hydrolysis. Also, there is a wide variation of analytical techniques, e.g. different enzymes and concentrations, and often less than full disclosure of the basis of yields.

Fully saturated steam was utilized because dry steam has been shown to form undesirable pyrolysis reaction byproducts. All steam explosion experiments were conducted for a residence time of 1 minute with saturated steam at 240° C. and 450 psig based on literature reported data (the sample were retained in the sample chamber for 5 to 10 minutes more during system heat up). It has been reported that enzymatic glucose conversions became essentially independent of pretreatment time as the steam pressure is increased to 350 psig. A more detailed study indicates that for the size of the biomass chips used in our experiments, the internal temperature of the chips would have reached 230° C. within three minutes for a reactor temperature of 240° C. After steam explosion and prior to analysis, the pretreatment biomass was thoroughly washed with water to remove furfural and other derivatives, which are inhibitory to enzymatic hydrolysis. It should be noted that the steam and the CBP pretreated biomass samples were not oven dried prior to analysis. It has been demonstrated that the drying process causes a collapse of the pore structure created by pretreatment. In Table 16, the results of steam explosion experiments are compared to the best CBP pretreatment experiments in terms of Ec′, enzymatic conversion efficiency based on total reducing sugars per unit of dry product.

TABLE 16
COMPARISON OF STEAM EXPLOSION AND CBP
PRETREATMENT
		STEAM
	UNTREATED	EXPLOSION	CBP	c.f.
BIOMASS TYPE	Ec′	Ec′	Ec	TALBE
White Pine Cubes	3.84	14.73	23.25	9
“Green” Red Oak	6.40	21.86	18.13	14
Chips
Newspaper Strips	15.39	15.29	50.47	15

Example 9

Critical Fluid Pulping

In addition to the mechanical force generated by the explosive decompression of a critical fluid, these solvents may be tailored to break down polymeric lignin into its aromatic constituents, solvate the reaction products, and thus separate lignin from the desirable cellulose and hemicellulosic constituents of wood. Critical fluids may thus have utility as a pretreatment as well as a pulping technique through defibration as a result of explosive decompression at a lower operating temperature, and fractionation as the result of critical fluid reaction and solvation effects.

In order to evaluate if explosive decompression of a critical fluid can enhance the pulping process, several preliminary microfibrillation experiments were conducted. Microfibrillated celluloses, having properties distinguishable from all previously known cellulose through a small diameter orifice in which the suspension is subjected to a high pressure drop (8,000 to 10,000 psig) in a homogenizer. Reportedly, the high velocity shearing action followed by a high velocity decelerating impact creates a substantially stable suspension after repeated passage through the small diameter orifice.

Microfibrillation experiments were conducted with high-pressure homogenization, critical fluid explosive decompression and a combination of these two techniques using a 0.2 wt % Domtar Northern softwood Kraft pulp. The high-pressure homogenization experiments, conducted with a APV Gaulin 15M unit, confirmed that several cycles (at least six) of room temperature homogenization or a lesser number of high temperature (80° C. to 90° C.) homogenization cycles must be used to microfibrillate pulp. Critical fluid decompression experiments were conducted in the critical fluid disruption apparatus shown in FIG. 3. These experiments were conducted in a batch mode with nitrous oxide and carbon dioxide at pressures ranging from 2,000 to 10,000 psig and temperatures ranging from 40° C. to 90° C.

While our analysis was impeded by fiber retention in the reactor after explosive decompression, these experiments indicated that N2O performed better than CO2, and high pressures and temperatures were more favorable. These experiments also indicated that microfibrillation was enhanced by explosively decompressing the critical fluid-biomass mixture through the Gaulin homogenization valve. These observations are based on the microscopic examination of the pulp before and after treatment. FIG. 14b for the SFS CO2 comminuted pulp, FIG. 14c for homogenized pulp (two passes), and FIG. 14d for SFS NO2 comminuted and homogenized pulp (one pass).

In order to confirm that, at the low temperatures utilized, the critical fluid does extract lignin and/or other wood constituents, several experiments were conducted with the critical fluid solubilization equipment shown in FIG. 4. In these experiments, white pine shavings were loaded into the solid example chamber and the system pressurized with supercritical carbon dioxide. After system stabilization at present operating conditions (e.g. 2,000 psig and 60° C.), the critical fluid recirculation pump, P-2, was turned on and the critical fluid monitored with an on-line, high pressure UV detector at 280 nm. Researchers have found that this UV wavelength was optimal for detecting lignin (or phenolic groups that make up the lignin molecule) solubilized in supercritical t-butanol. These researchers also utilized GC/MS and direct probe mass spectrometry analysis to confirm that interference from cellulose degradation products such as furfural and hydroxyl-methylfurfural was negligible.

The results of two critical fluid extraction experiments are shown in FIGS. 15 and 16. These experiments were both conducted with supercritical carbon dioxide at 60° C. for 1 hour. The sensitivity of both charts was set at 1.0 and zeroed with a 10% offset after system stabilization but prior to activating the critical fluid re-circulation loop. CBP-135 in FIG. 15 was conducted at a pressure of 2,300 psig and CBP-136 in FIG. 16 was conducted at a pressure of 4,300 psig. Both experiments indicate that a significant amount of lignin or other aromatic wood constituents such as tall oils, waxes, etc, was extracted into the circulating fluid, with the amount in CBP-135 being about 10% greater than in CBP-136. This result is in general agreement with some of our pretreatment data that showed higher enzymatic conversion efficiencies at 2,000 psig than at 4,000 psig. Although these conditions were relatively mild in temperature, the activity of carbonic acid at high pressures could have been significant. It should be noted that nitric acid in aqueous or alcohol solutions reacts rapidly with and dissolves lignin out of woody biomass.

While this invention has been particularly shown and described with references to specific embodiments, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.

Claims

1. A method of removing aromatic compounds from a cellulosic biomass having one or more aromatic compounds comprising the steps of combining a cellulosic biomass with a critical, near critical or supercritical fluid with or without a polar cosolvent; explosively decompressing said biomass to form a aromatic fluid and a disrupted biomass.

2. The method of claim 1 wherein said aromatic fluid is received in a recycling loop.

3. The method of claim 2 wherein said recycling loop removes said aromatic compound from said aromatic fluid to form a non-critical fluid and one or more aromatic compounds and directs said non-critical fluid to means for changing at least one of the group selected from temperature, pressure or constituents to make said non-critical fluid a near critical, critical or supercritical fluid with or without a polar cosolvent.

4. The method of claim 1 wherein said biomass is further processed with enzymes.

5. The method of claim 1 wherein said biomass is used as a feed component.

6. The method of claim 1 wherein said biomass is used as pulp.

7. The method of claim 1 wherein said biomass is converted to ethanol.

8. The method of claim 1 wherein said step of explosively decompressing said fluidized biomass comprised a 75 to 99% reduction in pressure over a period of time not exceeding 180 seconds.

9. The method of claim 8 wherein said reduction in pressure is in a period of time not exceeding sixty seconds.

10. The method of claim 9 wherein said reduction in pressure is in a period of time not exceeding one second.

11. The method of claim 1 wherein said step of combining the cellulosic biomass with a critical, near critical or supercritical fluid with or without a polar cosolvent takes place over a period of time of one minute to two hours.

12. A device for removing one or more aromatic compounds from a cellulosic biomass comprising:

A vessel having a chamber, a port, means for introducing a near critical, critical or supercritical fluid, means for explosively decompressing said near critical, critical or supercritical fluid with or without a polar cosolvent to form a aromatic fluid; said chamber for containing a cellulosic biomass having one or more aromatic compounds, said port having an open position for receiving a said cellulosic biomass having one or more aromatic compounds and a closed position in which said chamber is substantially closed, said means for introducing a near critical, critical or super critical fluid with or without a polar cosolvent in communication with at least one near critical, critical or supercritical fluid with or without a polar cosolvent, said means for explosively decompressing said critical, near critical or supercritical fluid capable of rapidly reducing pressure in said chamber;

A recycling loop in fluid communication with said means for explosively decompressing said near critical, critical or supercritical fluid with or without a polar cosolvent to receive said aromatic fluid, said recycling loop separating said aromatic compounds from said aromatic fluid to form a non-critical fluid and one or more aromatic compounds;

Critical fluid means for changing at least one of the group selected from temperature, pressure or constituents to make said non-critical fluid a near critical, critical or supercritical fluid with or without a polar cosolvent, said critical fluid means in communication with said recycling loop to receive said non-critical fluid and in communication with said means for introducing a near critical, critical or super critical fluid, said vessel receiving a biomass having one or more aromatic compounds in said chamber and forming a fluidized biomass with said near critical, critical or supercritical fluid with or without a polar cosolvent, and upon explosive decompression a aromatic fluid and a disrupted biomass, said aromatic fluid received in said recycling loop to form a aromatic compound and a non-critical fluid, said non-critical fluid received by said critical fluid means and processed to make a critical, near critical or supercritical fluid with or without a polar cosolvent to be received in said chamber.