Aspects relevant for upgrading Pyrolysis Liquids
It is not clear yet what upgrading routes (hydrogenation, hydrocracking, hydrodeoxygenation, and so on) are the best in hydroprocessing of pyrolysis oils, when the objective is to arrive at transportation fuels. Various conflicting results are reported in literature, specifically in terms of deoxygenation levels; they vary from 10 to 90 % at seemingly similar operating conditions and same type of catalysts, and for such discrepancies no explanations are presented. The lack of understanding is due to the complexity of the process itself, but also caused by the infancy of the pyrolysis process and products / analysis thereof. Pyrolysis oil obtained from varying resources and processes are completely different, the hydroprocessing itself is not well understood, and proper analysis techniques are missing.
It is generally accepted that the reaction conditions for conventional hydroprocessing of crude oil derivatives cannot be adopted as such for pyrolysis oils. In particular, the pressures should be higher. More importantly, pyrolysis oil cannot be treated directly at temperatures exceeding 300oC, because of a high charring tendency.
Various aspects play an important role in the hydroprocessing of such pyrolysis liquids:
The complexity of pyrolysis oil starts at the very beginning, viz. the biomass feedstock used to produce the oil. Biomass from different resources varies significantly in composition. The quality and composition of the pyrolysis oil depends strongly on the type of feedstock from which they are derived. As a consequence, each type of oil will behave differently in its further upgrading. In literature, a variety of (fractions from) bio-oils have been used. Earlier tests were carried out using oils derived from hydrothermal liquefaction,. The behavior of such oils in hydroprocessing reactions may be completely different compared to fast pyrolysis liquids. Obviously, understanding the nature of the feed and the processing technology and process conditions are essential to understand the literature results. However, the present analytical techniques are not fully sufficient to understand the hydrotreating of pyrolysis oils.
Charring is one of the crucial aspects in hydroprocessing pyrolysis oils. Without an active catalyst, significant charring of pyrolysis oil occurs due to repolymerization reactions. Active catalysts are sulfided NiMo/Al2O3 and CoMo/Al2O3, but also noble metal catalysts as Ru, Pd, Rh, Pt have been reported. However, catalysts are seen to deactivate after some operating time: values of a few hours are reported up to a few days. Different deactivation mechanisms are recognized, amongst others blocking of catalyst sites, irreversible poisoning of the catalyst (by nitrogen and sulphur compounds, water, etc), sintering or any other structural degradation of catalyst or support, coking and metal – alkali deposition, leaching of active metals to solution and so on. Information on the effect of ash, chlorine and /or nitrogen is not provided. It is reported that water in the pyrolysis oils has a strong negative effect on conventional alumina supported catalyst, because the alumina is partially converted into its hydroxide rendering the catalyst inactive). The charring is a function of type of feed, catalyst (structure) and operating conditions, and lignitic components as well as the carbohydrates are responsible. Substantial work is as of date dedicated to the development of new catalyst.
Crucial in the upgrading of such liquids is to find active catalysts able to withstand water at the extreme conditions of high temperature and pressure. Water present in the feed (and produced during hydroprocessing) will render most conventional supports for catalyst (e.g. silica, alumina) inapplicable as they can dissolve and rearrange. High pressures, and large residence time are required as well. One reason can be the limited solubility of hydrogen in pyrolysis oil, which may directly affect the reaction rate.
High reactor pressures (over 100 bar) seem necessary to avoid excessive charring due to the repolymerisation of the oil. It is unknown what (hydrogen partial) pressure is required to remain the hydrotreating activity. The effects of the presence of other gases than just hydrogen (e.g. CO, CO2 and CH4) on the process are not clear to date. It is also unknown why and how the hydrogen pressure actually affects the conversion, e.g. either by affecting the gas-to-liquid phase or internal catalyst hydrogen mass transfer or any other overall reaction rate limitation. Solubility data of hydrogen in bio-oil and in its hydrogenated products are not available.
Temperature and residence time
Already at low temperatures (< 250oC) and relatively short residence times (< 30 min), a considerable amount of hydrogen is consumed by the pyrolysis oil (up to 100 Nm</sup>3</sup>/t). However, a limited overall deoxygenation degree is achieved. The main reaction seems hydrogenation instead of hydrodeoxygenation, e.g the reactions of aldehydes and ketones to their corresponding alcohols. Actual hydrodeoxygenation of pyrolysis oils is achieved at temperature above 300oC only, which is then accompanied by large hydrogen consumptions (ranging from 200 up to 600 Nm3/kg oil). The product from this hydrotreatment can be distillable with limited or no coke formation. Studies including a systematic variation of the residence time of the liquid phase are limited. Hydrogenation of bio-oils at 200 to 250 oC may be achieved already at residence times varying from 15 up to 60 min, but no systematic studies on conversion rates are reported. Besides, at these temperatures limited overall deoxygenation is achieved. Effective deoxygenation down to oxygen contents < 10 wt% requires 1 to 2 hours residence time for the liquid phase, and significant higher temperatures of 350 to 400oC are required. The long residence time needed is reflected in the very low space velocities required for complete deoxygenation (liquid hourly space velocity of < 0.3 m3oil/(h m3 catalyst)). Apart from the preferred hydroprocessing reaction and un-preferred charring reactions, other (undesired) reactions can take place as well. The most prevalent is the methanation reaction, which is not preferred. It does not only cause temperature excursions as methanation of CO and CO2 is highly exothermicity (-207 kJ/mol CH4), but also consumes large amounts of hydrogen and carbon.
Reactor types used in earlier research of pyrolysis oil hydrotreatment are batch autoclaves and packed bed reactors. Interpretation of experimental results derived from these reactors is hindered in several ways. Non-steady state processes (autoclaves) are never easily interpreted, especially if in the initial stage other side reactions take place (e.g. catalyst activation, methanation, and so on). Packed bed reactors are supposedly operated in some sort of a trickle flow regime, but the relatively large hydrogen to oil ratio can cause operating problems, like a reduced catalysts wetting, and enhanced evaporation of water, leading to undesired charring. For both, autoclaves and packed beds, mass transfer data at these extreme conditions are also lacking, and solubility data in feed and product are not available.
The viscosity of the hydrotreated product appears a strong function of the oxygen content (Elliott, 2007). Pyrolysis oils with high oxygen contents usually contain substantial amounts of water, rendering the oil quite fluent (20 – 50 cSt at 20oC). Upon oxygen removal (and a subsequent phase separation of the oil), the viscosity of the organic product is increased drastically (up to 100,000 cSt at 20oC). Only after a certain degree of hydrotreating, the viscosity is reduced again, ultimately reaching the value conventionally found for hydrocarbons when the oxygen is completely removed (around 1 cSt at 20oC).
Upon such processing, interesting options for the thus upgraded pyrolysis oils are either by a further refining treatment in conventional FCC processes, or by full deoxygenation to produce fully deoxygenated biofuels. To be successful upgrading has to be carried out with the objectives to:
- Increase energy density;
- Reduce viscosity;
- Improve miscibility with fossil fuels, and
- Reduce reactivity at ambient conditions (‘stabilization’).
Further use of upgraded pyrolysis liquids
The intention of using pyrolysis liquids is to arrive at products from pyrolysis liquids that have the potential to be co-fed in crude oil refineries. It is clear that pyrolysis oils need to be upgraded before co-feeding not only to reduce the oxygen completely but rather to primarily limit the number of the functionalities in which the oxygen is bonded to the molecules, e.g. as alcohols, as ketones, in ether structures or as an aldehyde. Secondly the objective is to arrive at products that can be mixed in the large transportation fuel pool, preferably by CO2 removal rather than H2 addition.
Two interesting deoxygenation routes can thus be foreseen, viz. hydrodeoxygenation (removing oxygen using hydrogen producing water), and decarboxylation / decarbonylation (removing oxygen through CO2 and CO). In refinery terms, these routes are either (further) hydrotreating or Fluid Catalytic Cracking (FCC). The decarboxylation / decarbonylation route will be economically more feasible than the hydrotreating route, as hydrogen consumption can be high. Instead of full hydrodeoxygenation the pyrolysis oil as such, a more interesting approach is first to transform the cellulose-derived fraction, preferably to alcohols in a ‘mild hydrogenation’ step. The use of such upgraded liquids in co-FCC processing was part of the FP6-project BIOCOUP and is now subject of the FP7 project FASTCARD.
The amount of deoxygenation for successful co-processing in refining processes is unclear. A dedicated parameter to assess the degree of mild hydrogenation for the bio-oil (and possibly its use as a co-feed) is its tendency to produce coke, via. the residue retained upon distillation (for example the ‘Conradson Carbon Residue’, ‘Micro Carbon Residue Testing’ or ‘ThermoGraphicalAnalysis’). In general, pyrolysis-oils show CCR values from 20 to 50 %. Depending on the severity of the treatment, products can be distilled with (almost) no coke formation.
- Appell H.R., Fu Y.C., Friedman S., Yavorsky P.M., Wender I., 1971, Converting organic wastes to oil: A replenishable Energy Source, Bureau of Mines Report of Investigations 7560
- Sheu Y.H.E., Anthony R.G., Soltes E.J., 1988, Kinetic-Studies of Upgrading Pine Pyrolytic Oil by Hydrotreatment. Fuel Processing Technology, 19(1), 31-50.
- Ardiyanti, A.R. et al., 2012, Catalytic hydrotreatment of fast pyrolysis oil using bimetallic Ni–Cu catalysts on various supports, Applied Catalysis A: General 449(0): 121-130