Mild Hydrotreating of Bio-Oil - American Chemical Society

---

ARTICLE pubs.acs.org/EF

Mild Hydrotreating of Bio-Oil: Effect of Reaction Severity and Fate of Oxygenated Species Richard J. French, Jim Stunkel, and Robert M. Baldwin* National Bioenergy Center, National Renewable Energy Laboratory, 1617 Cole Boulevard, Golden, Colorado 80401, United States

bS Supporting Information ABSTRACT: Bio-oil derived by fast pyrolysis of biomass represents a potentially attractive source of hydrocarbon transportation fuels. Raw bio-oil however is completely unsuitable for application as a fuel due primarily to high organic oxygen content, which imparts a number of undesirable properties including high acidity and low stability. These problems can be overcome by catalytic hydrodeoxygenation (HDO); however, removing oxygen to very low levels by hydrotreating carries a strong economic penalty. Mild hydrotreating (where moderate levels of deoxygenation take place) coupled with coprocessing in a petroleum refinery represents an alternative to deep hydrotreating, which may help improve the economics of manufacture of hydrocarbon transportation fuels from biomass. This study reports on the effect of reaction severity on the quality of bio-oil produced via mild hydrotreating in a semibatch reactor using conventional hydroprocessing catalysts. Detailed speciation of oxygen functional groups in distillate and bottoms products has been carried out, and the fate of organic oxygen as a function of reaction severity has been determined. The results indicate that acceptable refinery blendstocks and perhaps final fuels can be produced by mild hydrotreating.

’ INTRODUCTION Concerns over global climate change and economic and sociopolitical issues associated with energy security and wealth transfer have highlighted the need to develop renewable and sustainable technologies for the manufacture of liquid transportation fuels. The Energy Independence and Security Act of 2007 (EISA) responds to these concerns by establishing goals for renewable fuels production including 36 billion gallons of renewable fuels in 2022 with 21 billion gallons of this amount to be advanced biofuels. Cellulosic-based transportation fuels must represent at least 16 billion gallons of the advanced biofuels.1,2 It is estimated that the United States alone can sustainably produce a billion tons of biomass per year that could be used for this purpose.3 Biomass-derived fuels can, in principle, be “carbonneutral” where the carbon dioxide released from the process and final products is compensated for by the carbon captured in the next year’s crop. A recent standard (RFS2)4 requires renewable fuels production and use to substantially reduce emissions of greenhouse gases as compared to conventional fuels. If biomass is heated rapidly (ΔT/Δt ≈ 1000 K/s) in the absence of oxygen to temperatures in the range of 400 650 °C (a process known as fast pyrolysis), a combustible liquid (bio-oil) retaining ca. 75% of the energy content of the feed material is produced in high yield. Although superficially resembling a heavy fuel oil, this liquid contains about 50% oxygen, 15 30% water, and has many undesirable physical and chemical properties when compared to petroleum liquids. Bio-oil is corrosive, only partly volatile, and largely immiscible with hydrocarbons.5 8 Bio-oil can be converted to a gasoline- or diesel-like liquid by catalytic hydroprocessing using catalysts and conditions that are very similar to those used in petroleum hydrodesulfurization, hydrotreating, and hydrocracking processes. A recent review highlights the history and current status of this technique.7 When this approach was applied to fast pyrolysis oil, the oil was found to r 2011 American Chemical Society

coke severely in the original single-stage process;6 hence a twostage process was developed.9,10 In this improved process, the oil was stabilized at a lower temperature (150 280 °C) before it was fed to a high temperature reactor (350 400 °C) where the majority of the oxygen removal took place. Standard petroleumindustry hydrotreating catalysts were used including both nickel molybdenum (NiMo) and cobalt molybdenum (CoMo) on γ-alumina support. Recently, economic studies have been conducted of this process.11,12 Jones et al.11 assumed the bio-oil was produced in a circulating fluidized bed at a feed rate of 2000 bone-dry metric tons/day of wood chips at a cost of $50.7/ton. The oil was then hydrotreated in two stages at temperatures of 240 370 °C and 2015 2500 psig pressure to produce 44% hydrotreated oil containing 1.5% oxygen. Hydrogen consumption was assumed to be 5 wt % of the feed. This product oil was then hydrocracked as necessary and separated into gasoline and diesel streams. The projected minimum fuel-selling price was $2.04/gal ($1.34/gal ethanol equivalent), which could be reduced to $1.74/gal ($1.14/gallon ethanol equivalent) by colocating the plant with a petroleum refinery to take advantage of the low-cost hydrogen and large-scale hydrocracking capacity in the refinery. While these projections are promising, earlier studies concluded that the process was too expensive to compete with low-cost oil because of the large amount of hydrogen consumed, low product yields, low quality products that would require further upgrading in a refinery, and the corrosiveness of the raw oil.12 14 Another recent study15 was carried out to obtain an assessment of the process in light of incremental improvements to the bio-oil catalytic hydrodeoxygenation process, changes in refining Received: March 25, 2011 Revised: June 10, 2011 Published: June 13, 2011 3266

dx.doi.org/10.1021/ef200462v | Energy Fuels 2011, 25, 3266–3274

Energy & Fuels requirements (particularly increased hydrodesulfurization requirements), and changes in petroleum prices. This report suggested that costs could be significantly reduced by mild hydrotreating of the bio-oil and then coprocessing the partially deoxygenated products with petroleum-derived material in a refinery. Therefore, the authors recommended reducing the severity of HDO to leave about 7% oxygen in the bio-oil, thus avoiding hydrogenating aromatics while reducing hydrogen consumption, catalyst costs, and hydrotreater capital costs. The residual acidity of the oil could then be accommodated by diluting with crude oil or an internal refinery stream (naphtha, gas oil, etc.). Various blending strategies could be employed as shown in Figure 1. This simplified schematic shows material derived from mild hydroprocessing of bio-oil blended and coprocessed with petroleum-derived material using existing refinery unit operations to carry out additional deoxygenation required to maintain acceptable product quality for the final fuels. For this strategy to work, a number of important criteria must be met by the products from mild hydrotreating: (1) The acidity of the bio-oil must be reduced from the typical TAN (total acid number) value of over 100 to about 15, assuming that hydrotreated bio-oil would be blended in a 1:8 ratio (acidity of blend less than 2 mg KOH/g oil (15)).

Figure 1. Blending strategies for bio-oil in the petroleum refinery.

ARTICLE

(2) The hydrotreated bio-oil must be completely miscible with hydrocarbons. (3) The hydrotreated bio-oil must be highly volatile so that it is amenable to fractional distillation (some high-boiling residue is acceptable). If hydrotreated bio-oil is to be coprocessed in the refinery, information on the composition and concentration of oxygenates in the bio-oil-derived hydrotreated products is critical to ensure that product quality is maintained. Further, while considerable data exist on bio-oils that have been hydrotreated to 2% oxygen or less, there are little or no data on hydrotreated bio-oils containing ∼7% oxygen, and there is very little information concerning the relationship between total oxygen, oxygen functional groups, acidity (TAN), hydrocarbon miscibility, and hydroprocessing severity.

’ EXPERIMENTAL SECTION Reactor. Significant work on the hydrotreating of biomass pyrolysis oil has been done in trickle-bed reactors;16,17 model compound studies18,19 and some work on whole oil20 have been done in batch reactors. A semibatch configuration with the ability to sweep the reactor headspace was the configuration selected for this study (P and ID shown in Figure 2). This slurry reactor provides well-defined catalyst-feed contact and the opportunity to collect gas, condensate, and residual liquid samples (via an internal dip tube) throughout the experiment as conditions are changed. Because the underlying chemistry is the same, the type of products and changes in products with operating parameters will be qualitatively similar to those of continuous reactors. However, it is recognized that semibatch reactors have several inherent limitations including unsteady-state operation, continuously varying catalyst-to-oil ratio, and poorly defined temperature/time relationships due to lengthy heat-up periods between preheat and reaction. A simplified schematic is shown in Figure 3, indicating the major functional features of the semibatch system including the high-pressure condenser system used to obtain samples of the products swept from the reactor. Bio-Oil. The oil used was produced from white oak during 2008 by entrained flow pyrolysis at 550 °C in the NREL Thermochemical

Figure 2. Hydrotreater schematic. 3267

dx.doi.org/10.1021/ef200462v |Energy Fuels 2011, 25, 3266–3274

Energy & Fuels

ARTICLE

Process Development Unit. The residence time in the entrained-flow reactor was ∼0.5 s. Results of proximate, ultimate, and acid analysis of this oil are shown in Table 1. Because of the higher pyrolysis temperature employed, this oil has somewhat higher water content and lower organic oxygen content than is considered typical.5 Catalyst and Pre-sulfiding Conditions. Catalyst was provided by Grace Davison and is representative of a modern hydrotreating catalyst. It contains nickel-promoted molybdenum on a high-surface alumina support (Ni Mo/γ-Al2O3, 4.8% NiO, 25% MoO3). Ni Mo catalysts when compared to the other standard hydrotreating catalyst (Co Mo/γ-Al2O3) feature a higher rate of decarboxylation of organic acids, which may be advantageous in this work; however, they are also more active for aromatic ring hydrogenation, which could lead to

Figure 3. Simplified reactor system schematic.

Table 1. Oil Analysis component

wt %

water

32.8

ash