Thatcher W. Root

Adsorptive membrane chromatography for purification of plasmid DNA.

Adsorptive membranes were investigated for the downstream processing of plasmid DNA by quantifying both separation efficiencies and adsorption uptake with the anion-exchange membranes. Separation efficiencies of the 10-ml Mustang-Q were measured using pulses of 6.1-kilo base pair plasmid DNA and lysozyme tracers, and comparing the responses for both conventional and reverse-flow operation. The plasmid exhibited nearly 200 plates/cm, almost as high efficiency as the protein despite the large difference in size. This behavior contrasts strongly with typical behavior for spherical porous particle packings, which predicted large decreases in efficiency with increases in tracer size. Batch adsorption isotherms for the 6.1-kilo base pair plasmid on small sheets of anion-exchange membranes at various ionic strengths showed high capacities for very large biomolecules. The maximum binding capacity for the membrane unit was calculated as 10 mg plasmid/ml, an order of magnitude greater than typical values reported for porous beads.

Production of Biofuels from Cellulose and Corn Stover Using Alkylphenol Solvents

Technically viable and economic conversion of lignocellulosic biomass into chemicals and fuels is an important challenge. One effective strategy is to first partially remove oxygen to produce reactive intermediates, denoted as platform molecules, followed by conversion of these molecules into desired products. These platform molecules have fewer functional groups than carbohydrates in biomass (e.g. , xylose, glucose), offering the potential for selective catalytic upgrading processes. One such platform molecule, which is receiving considerable attention in the literature, is levulinic acid (LA). From LA a variety of fuels and chemicals can be made, such as valeric acid esters, methyltetrahydrofuran (MTHF), and LA esters and ketals. Another valuable product is g-valerolactone (GVL), obtained from the reduction of LA. GVL can be used directly as a fuel additive, or as a precursor for fuels and chemicals. It is possible to achieve LA yields greater than 50 % by using aqueous solutions of sulfuric acid for cellulose deconstruction; 12] however, the commercial production of LA and its derivatives presents serious challenges. First, LA must be separated from the mineral acid to recycle the acid catalyst and avoid negative effects in downstream processes. Second, LA is produced in a low concentration, and its purification/recovery is expensive. Finally, the solvents normally used (e.g. , water) have lower boiling points than LA, which means that to recover the product by distillation the solvent needs to be evaporated. In this Communication, we show that alkylphenol solvents (i.e. , substituted benzene compounds containing a hydroxyl group and alkyl groups, R ; see Figure 1) can be used to selectively extract LA from aqueous solutions of sulfuric acid, thereby eliminating the previously mentioned drawbacks. In addition, alkylphenol solvents extract GVL from water with a higher partition coefficient compared to LA. Accordingly, the GVL concentration in the alkylphenol solvent can be increased by the conversion of LA to GVL combined with the recycle of this stream for successive extractions, thus facilitating recovery of GVL from the solvent by distillation. Importantly, we show that a carbon-supported Ru–Sn catalyst can be used for the selective reduction of LA to GVL by H2 in the presence of alkylphenols, without hydrogenation of the solvent. Finally, the aqueous phase containing sulfuric acid after extraction of LA by the alkylphenol solvent can be recycled for subsequent cycles of cellulose deconstruction, providing an effective strategy for sulfuric acid management in the process. An advantage of using alkylphenol solvents is that the alkyl groups strongly modify the physical properties of the solvent. For example, while phenol is relatively soluble in water, the presence of a longer alkyl chain increases the hydrophobicity of the molecule, decreasing the solubility of the alkylphenol in polar media (e.g. , water) and increasing the boiling point, which facilitates purification of the product at the top of a distillation column without need to evaporate the solvent. The alkyl chain length can be chosen to be sufficiently long to have a high boiling point, while keeping a good balance between hydrophilic and hydrophobic groups to extract LA and not sulfuric acid. Table 1 shows that with 2-sec-butylphenol, the partition coefficient for LA extraction remains at a value of approximately 2 (entries 1–3), while the partition coefficient for extraction of formic acid (FA) increases when the concentration of the products is increased. When the length of the alkyl chain of the alkylphenol increases, the polarity of the alkylphenol is reduced and the partition coefficient for LA decreased to 1.2 when using 4-n-pentylphenol (entry 4) and to 0.8 when using 4-n-hexylphenol (entry 5). The FA partition coefficient does not change significantly. GVL is more hydrophobic than LA and thus the partition coefficients are higher (entries 6–8), which allows the GVL concentration to be increased by successive recycling steps after hydrogenation of LA (Figure 1). However, as the amount of GVL in the 2-sec-butylphenol organic phase increases (entries 9–12), the LA partition coefficient decreases. Thus, the extent of solvent recycling prior to distillation represents a compromise between achieving high concentrations of GVL while also maintaining a high partition coefficient for LA extraction. Another advantage of the alkylphenol solvents is that the extraction can be carried out at elevated temperatures (entry 13), suggesting that the process could be carried out at the temperatures employed for cellulose deconstruction, thereby minimizing the need for heat exchangers, leading to energy and equipment savings. In addition, sulfuric acid was not detected in the organic phase for any of the entries in Table 1. Thus, it is possible to use the aqueous phase for multiple steps of cellulose deconstruction. After extraction of LA, the next step in the process is hydrogenation of LA to GVL. Previous literature reports that ruthenium on carbon (Ru/C) could be an effective catalyst. However, Ru/C hydrogenated the C=C bonds in 2-sec-butylphenol, forming butylcyclohexanol and butylcyclohexanone (corresponding to 0.3 % conversion of 2-sec-butylphenol at conditions listed in Table 2, entry 1). In addition, the Ru/C catalyst undergoes deactivation with time-on-stream in the presence of FA, even at [a] Dr. D. M. Alonso, Dr. S. G. Wettstein, Dr. J. Q. Bond, Prof. T. W. Root, Prof. J. A. Dumesic Chemical and Biological Engineering Department University of Wisconsin Madison, WI 53706 (USA) Fax: (+ 1) 608-262-5434 E-mail : dumesic@engr.wisc.edu Supporting Information for this article is available on the WWW under http://dx.doi.org/10.1002/cssc.201100256.

Protein retention in hydrophobic interaction chromatography : modeling variation with buffer ionic strength and column hydrophobicity

The variation in protein retention times with protein and surface hydrophobicity and mobile phase composition and concentration has been described with a simple thermodynamic model. Column capacity factors for two proteins have been measured as a function of mobile phase ionic strength for a series of columns with varying levels of hydrophobicity. Application of the model to these data suggests that the protein retention is dominated by the release of water molecules upon adsorption, which is consistent with the entropically driven nature of hydrophobic interactions. The calculated number of water molecules released agrees with estimates based on the reduction in hydrophobic surface area for adsorption.

Experimental Limiting Oxygen Concentrations for Nine Organic Solvents at Temperatures and Pressures Relevant to Aerobic Oxidations in the Pharmaceutical Industry

Applications of aerobic oxidation methods in pharmaceutical manufacturing are limited in part because mixtures of oxygen gas and organic solvents often create the potential for a flammable atmosphere. To address this issue, limiting oxygen concentration (LOC) values, which define the minimum partial pressure of oxygen that supports a combustible mixture, have been measured for nine commonly used organic solvents at elevated temperatures and pressures. The solvents include acetic acid, N-methylpyrrolidone, dimethyl sulfoxide, tert-amyl alcohol, ethyl acetate, 2-methyltetrahydrofuran, methanol, acetonitrile, and toluene. The data obtained from these studies help define safe operating conditions for the use of oxygen with organic solvents.

The effect of high intensity mixing on the enzymatic hydrolysis of concentrated cellulose fiber suspensions.

Enzymatic hydrolysis of lignocellulosic biomass in a high shear environment was examined. The conversion of cellulose to glucose in samples mixed in a torque rheometer producing shear flows similar to those found in twin screw extruders was greater than that of unmixed samples. In addition, there is a synergistic effect of mixing and enzymatic hydrolysis; mixing increases the rate of cellulose conversion while the increased conversion facilitates mixing. The synergy appears to result in part from particle size reduction, which is more significant when hydrolysis occurs during intense mixing.

Flow distribution in chromatographic columns

Abstract It becomes increasingly clear that flow non-uniformity frequently limits the performance of chromatographic columns, and that there are at least two major causes for this: non-uniform packing and inadequate header design. Static magnetic resonance imaging (MRI) of small columns in this laboratory has confirmed previous reports of packing non-uniformity and found void fractions to be smallest in the neighborhood of the wall while the central and upstream regions of the column are more loosely packed. Both static and flow MRI suggests that additional mal-distribution is introduced by non-uniform flow in headers, especially for short columns. Flow non-uniformity can reduce resolution directly, and it can also limit column capacity through perturbation-induced viscous fingering. A preliminary stress–strain analysis during column packing is reported which shows that the stress distribution depends strongly upon column aspect ratio (length/diameter) and is most favorable for low values. At the same time adequate flow distribution becomes more difficult as this ratio decreases, and this requires careful attention to the column headers. We therefore suggest a new strategy for header design and provide two specific examples which produce very nearly uniform residence time for all streamlines as well as uniform exit velocity.

Refining the scale-up of chromatographic separations.

The use of heavily loaded columns and complex processing conditions makes scale-up of chromatographic separations a non-trivial process. The wide ranges of process conditions that must be investigated demands that a large number of preliminary experiments must usually be made in small columns and laboratory-scale work stations. These preliminary data can be biased by improper column packing, poor distributors and dispersion in auxiliary apparatus, and it is important to understand these disturbing factors in detail. Moreover, it is precisely at this macroscopic level that our understanding of the chromatographic process is weakest, for large columns as well as small. This paper addresses three of these factors: Efficient elimination of peripheral effects and characterization of both header flow distribution and packing non-uniformity. This will be done using a variety of experimental and analytical approaches including nuclear magnetic resonance imaging, computational fluid dynamics and mass transfer, and careful experimentation.

Product Selectivity During Regeneration of Lean NOx Trap Catalysts

NOx reduction product speciation during regeneration of a fully formulated lean NOx trap catalyst has been investigated using a bench-scale flow reactor. NH3 and N2O were both observed during the regeneration phase of fast lean/rich cycles that simulated engine operation. Formation of both products increased with higher reductant concentrations and lower temperatures. Steady flow experiments were used to decouple the regeneration reactions from the NOx storage and release processes. This approach enabled a detailed investigation into the reactions that cause both formation and destruction of non-N2 reduction products. Pseudo-steady state experiments with simultaneous flow of NOx and reductant indicated that high concentrations of CO or H2 drive the reduction reactions toward NH3 formation, while mixtures that are stoichiometric for N2 formation favor N2. These experiments also showed that NH3 is readily oxidized by both NO and O2 over the LNT catalyst. These observations were incorporated into a schematic of the regeneration process that takes into account the spatial and temporal variations occurring within the catalyst.

Emergence of Ideal Membrane Cascades for Downstream Processing

An algorithm is developed for describing ideal membrane cascades for fractionation of binary and pseudo‐binary mixtures. It is shown that solvent management plays a key role in determining both purification and yield. Development of efficient diafilters is needed if membrane cascades are to achieve their full potential in competing with both chromatography and simulated moving bed operations in downstream processing of proteins. Such a replacement will also be important for fractionation of higher titers and larger substrates, such as plasmids, viruses, and even whole cells.

Characterizing the performance of industrial-scale columns.

The performance of a large commercial chromatographic column was investigated using a short pulse of a tracer and an extension of the reverse-flow technique. This technique permits separate determination of the unavoidable irreversible microscopic processes and the reversible effects of flow maldistribution, and allows for the separation of flow maldistribution in the flow distributors from flow maldistribution inside the packed bed. This analysis was performed on a 0.44 m Millipore IsoPak column using Cellufine GC 700, cellulosic-based media with an average particle diameter of 75 microm, for the stationary phase. The column efficiency was quantified by analysis of the effluent curve from a short pulse of a 5% aqueous acetone tracer. The study examined behavior of beds of different lengths (10-24 cm) and beds packed from different slurry concentrations (10-75% v/v). The slurry-packed columns were very uniform, and no significant macroscopic flow maldistribution was observed inside the column. The observed bed plate heights conformed to the predictions of available one-dimensional continuum models. Dispersion in the flow distributors was significant, corresponding to 15-25% of the intracolumn dispersion when the full 24 cm available bed length was used and a proportionally larger increase for shorter bed lengths. Thus, the headers are shown to produce a significant increase in the observed plate height.