![[NF-2000 Database - AIR Program]](../images/Air.gif) |
AIR2-CT93-1436
Thermochemical Biomass Gasification: Upgrading of the Crude Gasification Product Gas |
 |
Proposal No: | AIR2-CT93-1436 |
| Date Prepared: | September 1999 | |
| Source: | Final synthesis report 1996 |
Introduction
Biomass can be used as feedstock for gasification processes. Producer gas obtained from these processes may contain significant amounts of tars, which hamper the application of the gas, for instance in gas engines. A maximum tar content in the gas of about 20-500 mg/Nm3 (depending on further applications) could be acceptable. It means that with initial tar contents between 2 and 100g tars/Nm3, tar conversions of at least 99% are required. This can be achieved by using various types of catalysts that in turn may be located in the same gasifier bed or downstream from the gasifier. From the various types of solids active or useful for hot gas cleaning, this project has concentrated on two of them that have already been shown to be beneficial. The solids used were calcined dolomites and related materials such as magnesite or limestone (80% of the project effort) and commercial (nickel based) steam reforming catalysts (20% of the overall effort).
Results
Both types of solids have been demonstrated to convert (destroy) tars present in the raw flue hot gas produced in biomass gasification when used downstream from the biomass gasifiers that are run well such as to produce 'a not very dirty' raw gas or a gas 'as clean as possible'. An optimised operation of a fluidized bed biomass gasifier can produce a gas with only 2g tars/Nm3. This appeared to be the upper limit above which nickel catalysts cannot be used since they are deactivated by coke formation.
As far as the quality the raw gas is concerned, the equivalent ratio (ER) appeared to be the most important factor in biomass gasification with air. This defines the temperatures of the bed and of the freeboard, the tar yield and the composition and calorific value of the fuel gas. On increasing ER from 0.20 to 0.45 the gas heating value decreases about 2 MJ/Nm3 and the tar yield decreases by about a 50wt%.
The H/C ratio in the flue gas is a measure of the steam content in the gas inside the gasifier. This appears to be an important variable in biomass gasification since increasing the ratio from 1.6 to 2.2 resulted in a 75 wt% decrease in tar, while the heating value of the gas increased by about one MJ/Nm3.
To obtain a high quality raw gas (maximum heating value and minimum tar content), the results obtained of this project recommend that the biomass is fed near the bed bottom, under the following conditions:
A small secondary air injection (10-20% of the primary air) in the freeboard and an addition (2-5 wt% respect to the biomass feed) of a calcined dolomite to the bed, mixed with the biomass, was found to improve the quality of the raw gas produced. However, under these conditions the loss of fines from the bed to the downstream cyclones or filters increased. This has to be taken into account in the design of the overall process.
Even using in-bed dolomite it was found quite difficult, and perhaps not possible, to reduce the tar contents in the raw gas below 0.5-2.0 g/Nm3. Hence, a hot gas catalytic cleaning is required to reduce tars below this level.
From studies of the gasifier performance, the gas quality, composition and dirtiness at the gasifier exit was known. Hence, so was the composition of the gas at the inlet of the secondary or downstream catalytic reactor. The detailed effects of various catalysts could thus be studied for known gas compositions.
The study of the effectiveness and usefulness of dolomites for hot raw gas cleaning was carried out with two types of gas atmospheres: synthetic ones and realistic ones (from biomass gasifiers). The synthetic gas mixtures used pure, targeted compounds known to be present in tars and hard to destroy (such as naphthalene and toluene). The effects of reaction conditions (such as gas composition temperature and residence time of the catalyst) on the rate of conversion of high-temperature tar was investigated using these catalysts. Such information is important for optimal design of reactors and for best choice of catalyst as well as defining pre-treating conditions and making possible improvement in the catalyst.
A study of reaction kinetics was used to formulate mathematical expressions that describe relationship between reaction rate with dolomite and with lime and local gas composition and temperature in catalytic reactors. These expressions are useful for modelling of reactors, in analyses of results from small experimental reactors, from pilot scale and full-scale reactors as well as for process design.
Naphthalene was used as a model substance in the detailed investigation of reaction kinetics. This compound is very suitable for this purpose, since it is the dominating troublesome component that remains when tar is not fully converted by the catalytic process. Benzene and toluene are also major residual components from tar, but these cause fewer practical problems since they condense as liquids, while naphthalene forms a solid that clogs filters and gas lines, even at short residence times. Hence, the decomposition of naphthalene will be limiting in sizing of reactors, and hence the kinetics used in reactor modelling were mainly that of naphthalene.
This work indicated that the hydrocarbons that are formed in the decomposition of naphthalene, in addition to the main products CO, CO2, H2 and H20, are benzene, toluene and methane. Benzene and toluene were shown to be intermediate decomposition products from naphthalene. The rate of decomposition was shown to be proportional to the concentration of naphthalene, i.e. the overall reaction can be treated as a first order reaction with respect to naphthalene. It was shown that hydrogen and steam has an important retarding influence. Reaction rate in carbon dioxide-nitrogen mixtures, with no hydrogen or steam added to the inlet gas, was about 35 times higher than in a typical fuel gas at atmospheric pressure, In spite of this, increase in the partial pressure of carbon dioxide did not influence the rate of reaction. Carbon monoxide had no effect on the rate of reaction.
Under pressurised conditions at the temperature used (850-900 ºC), the calcium content of dolomite is present as carbonate rather than as oxide which is not active as catalyst. Temperatures at above 920 ºC would be required to obtain the oxide in a 10 bar process. Extrapolation of the results on the influence from the concentrations of main gas components indicated that such an increase of temperature would also be needed to compensate for the inhibiting effects of hydrogen and steam.
It was shown that an initial deactivation occurs when the dolomite or lime catalyst was exposed to the reaction mixture. This process took place at relative rates that were rather similar to published data on rates of reduction of BET surface area of calcium oxide in steam and carbon/dioxide. However, other effects can be involved. Step changes of gas composition and temperature caused temporary changes of reaction rate that possibly can be explained by a partial and reversible coverage of the catalyst surface by intermediate reaction products.
Four dolomites of different origin were compared for their ability to convert tar in pyrolysis gas produced from biomass. These were 3 Spanish dolomites (Navarra, Chilches II and Sevilla) and the Swedish dolomite Glanshammar. Considerable differences that affected their usefulness for gas cleaning were observed, with the Swedish dolomite having the highest activity. The rate constant estimated from tests at 850 ºC for the Spanish dolomites varied from 8.2 to 10.8 m3/kg h, while for the Swedish product the value was 15.6 m3/kg h.
A new catalyst was obtained from a cheap raw material. This catalyst had a good ability to decompose tar into useful gases at much lower temperature than those required for dolomite and lime. The ability of this catalyst was investigated using both naphthalene and pyrolysis gas. The catalytic activity at 500 ºC was of similar magnitude as that of dolomite at 900 ºC. However, temperatures above 600 ºC would be required to avoid carbon formation from carbon monoxide through the Boudouard reaction. Further work is needed to determine if this catalyst can be developed for commercial use.
Decomposition of ammonia in fuel gas was investigated using a nickel catalyst, an iron based catalyst and a dolomite. The nickel catalyst was very effective for decomposition of ammonia, while the other materials had low or negligible effect on ammonia in fuel gas.
Hydrogen sulphide caused deactivation of the nickel catalyst in decomposition of ammonia, while chlorine did not deactivate the catalyst. Steam concentration did not affect ammonia decomposition. Dolomite showed a high ability to decompose hydrogen cyanide.
Agricultural residues A further task was to study the catalytic decomposition of biomass tar derived from agricultural residues, analyse the gas and select target compounds from the complex biomass tar. Straw and Miscanthus were used as raw materials. After comminution of pellets, the raw materials were pyrolyzed at 700ºC. The tar-rich gases evolved, after addition of steam, were fed into a cracker where the amount of catalyst and temperature were varied. The resulting gases were cooled to condense the remaining tar, with permanent gases and hydrocarbons up to toluene directly analysed by GC. After the experiments run, neutral and aromatic compounds in the condensed tars were separated by solid phase extraction prior to analysis of target compounds by capillary GC. Basic compounds were separated from the organic extracts by extraction with sulphuric acid and then concentrated by reverse phase solid-phase extraction, prior to elution and analysis by capillary GC.
Experiments were performed at 700, 800 and 900 ºC, with a raw material input of between 100 and 200 g raw material per hour, with a nitrogen flow of either 28 or 56 Nm3/h and 24 to 54 g steam added per hour, followed by thermal and catalytic cracking, with varying amounts of dolomite as catalyst. Gas compositions after cracking corresponded closely to shift reaction-equilibrium compositions, indicating well-functioning experimental equipment.
It was found that the input concentration of tar to the cracker was greater than for actual gasification. This was a result of less dilution by nitrogen as well as the fact that some of the tars are converted within the gasifier, before entering the cracker.
The results obtained with straw and Miscanthus were compared with previously obtained results using mixed hardwood. The raw materials appeared to behave quite similarly though it appeared that tars from straw and Miscanthus decompose somewhat more readily than tar from mixed hardwood.
The compositions of the three tars obtained after cracking at 900 ºC using 80 g: dolomite was compared. Total tar conversion was the same for all three raw materials, 96%, and the components were similar. However, naphthalene percentage (by weight) of total tar varies from 51% for hardwood, 62% for Miscanthus to 75% for straw. These differences were also seen without catalyst with the corresponding percentages being 9, 12-13 and 24. It may be concluded that the tars from straw are the easiest to decompose to naphthalene, with the hardwood tars being somewhat more difficult, with Miscanthus in between but closer to hardwood. In practice this may mean very little, since reduction of naphthalene levels is the limiting step in catalytic cracking, requiring by far the longest residence time.
Screening A small-scale installation was built for screening potential catalysts and to gain kinetic data. Initially, the aim was to study carbonaceous catalysts such as charcoal and cokes. However, following preliminary experiments that showed losses and other disadvantages of these materials, dolomite was tested in the adiabatic packed bed reactor used with periodic reversal of the feed flow. This gives a relatively low, inlet and outlet temperature, while in the middle of the reactor a high-temperature zone exists. The reverse flow operation mode enables clean producer gas to be obtained in a low energy-demanding process. However, to enable stable operation a small positive heat effect is required and the tar cracking process is endothermic. Therefore, a small amount of air is injected in the reactor resulting in spontaneous combustion. By controlling this secondary air flow, the temperature level in the reactor can be regulated, The reactor was operated successfully, but even with a very high tar conversion (> 99%) the outlet concentration was still somewhat too high (200 mg/Nm3).
An economical analysis of a total CHP - unit including the proposed tar cracker showed that additional investment costs for the tar cracker are relatively small. Furthermore, the increase in operational costs is only a few percent. Further economical optimisation need to be carried out at plant level.
Kinetic model A simple kinetic model was developed in order to study the effectiveness of different types of dolomites for hot gas cleaning from biomass gasifiers for overall tar elimination was developed. It introduced an overall kinetic constant (kapp) for the net tar removal that was a good index for the activity of the dolomite or nickel catalyst in terms of the tar elimination.
Dolomite phisico-chemical analysis Pore structure analyses were made for four different dolomites. Surface area (BET) determinations, mercury porosimetries and chemical analyses have been made. Pore size distributions and specific surface areas are similar for the different dolomites, all of which showed only large micropores, with little evidence of micro- or mesoporosity. The origin, quarry, composition and/or type of dolomite have only a small influence (10- 20%) on its activity for tar destruction.
Nickel catalyst A commercial steam reforming catalyst from BASF (GI-25S) was proved for tar elimination from the flue gas. Longitudinal profiles of temperatures in the fixed bed of the nickel catalyst were measured. However, it was not easy to get an isothermal fixed bed at the small pilot plant scale. This was taken into account in all analyses made in this part of the project work. For particle sizes greater than 1.0 mm (as is the case for the commercial catalyst pellets) internal diffusion is very important and controls the process. At high temperatures (>780ºC) no catalyst deactivation was detected in runs of 8 hours. This nickel catalyst was found to be very active. Tar contents of less than 10 mg/m3 were easily obtained with space times of 15.000 h-1, contact times of 0.1 seconds.
Air addition The effect of air addition on biomass tar conversion in catalytic packed bed crackers was studied by using an isothermal micro reactor as well as both a fluidised bed and fixed bed bench scale biomass gasification set up with down stream tar crackers. The micro reactor was used in experiments with artificial biomass producer gas containing naphthalene as a model tar compound. Experiments were carried out with inert silica and catalytically active calcined dolomite bed material both with and without air addition. Kinetic results were modelled assuming three parallel first order naphthalene decomposition reactions.
With secondary air the results show a significant increase in tar decomposition rate both with dolomite and silica. With dolomite, the naphthalene decomposition reactions on the catalyst surface dominate the overall conversion rate making the relative effect of air addition less significant. However, with dolomite, secondary air feeding has the positive effect of significantly decreasing the catalyst deactivation rate. Experimental results with real tar from the fluidised bed bench scale gasification set-up were in qualitative agreement with results from the micro reactor experiments. At a constant equivalency ratio (ER) for the overall gasification process the use of secondary air resulted in a sharp decrease in tar content in the exit gas stream of the dolomite cracker.
The experimental results showed that the addition of secondary air results in a sharp decrease in tar concentration in the outlet gas of tar crackers packed with dolomite and more particularly when packed with silica which generally is supposed to have no catalytic activity. In commercial scale gasifiers the ratio of secondary air and primary air will be limited by the resulting decrease in carbon burn out in the gasifier.
Conclusions
Overall, this project has generated more than two hundred conclusions useful for hot and catalytic gas cleaning or upgrading relevant to biomass gasifiers. Probably, it will take several years to distribute these results, for them to be understood and applied in commercial gasification plants.
© Copyright 2006 Policy Statements
Updated
by CPL Press:
03/07/2007
- biomatnet@biomatnet.org
 |
 |
 |
News |
 |
Events |
|
|