• Energy Research

IMPROOF will develop and demonstrate the steam cracking furnace of the 21st century by drastically improving the energy efficiency of the current state-of-the-art, in a cost effective way, while simultaneously reducing emissions of greenhouse gases and NOX per ton of ethylene produced by at least 25%. Therefore, the latest technological innovations in the field of energy efficiency and fouling minimization are implemented and combined, proving that these technologies work properly at TRL 5 and 6 levels. The first steps to reach the ultimate objective, i.e. to deploy the furnace at the demonstrator at commercial scale with the most effective technologies, will be discussed based on novel pilot scale data and modeling results.

Abstract An Estonian pyrolysis shale oil has been characterized using comprehensive two dimensional gas chromatography (GC × GC) coupled to a flame ionization detector (FID) and time of flight mass spectrometer (TOF-MS). Oxygen containing compounds were abundant and the following classes were identified: benzenediols, hydroxybenzofurans, phenols, ketones, and naphthols. The large peak overlap observed for oxygenates with the hydrocarbon matrix using a non-polar × polar column combination made identification and quantification difficult. A methodology based on the more selective separation using a polar × non-polar and polar × mid-polar column configurations was developed and tested. The obtained bidimensional resolution confirms that the method provides an improved separation of unsaturated hydrocarbons, mid-polar and even highly polar oxygen containing compounds. The use of an internal standard methodology allowed complete quantification of the pyrolysis shale oil both by carbon number and chemical class.

Abstract The thermal decomposition of 2,5-dimethylfuran is studied in a bench-scale pyrolysis set-up equipped with a dedicated on-line analysis section including a GC × GC-FID/(TOF-MS). This analysis section enables both qualitative and quantitative on-line analysis of the entire reactor effluent with an unseen level of detail. The reactor temperature was varied from 873 K to 1098 K at a fixed pressure of 1.7 bar and a residence time of 300–400 ms, covering the complete conversion range of 2,5-dimethylfuran.The main products at low conversions are hydrogen, CO, methane, phenol, 2-methylfuran, 1,3-cyclopentadiene and a C 7 H 10 O isomer. At higher conversions increasing amounts of mono- and poly-aromatics such as benzene, toluene, indene and naphthalene are formed. These species appear to be secondary reaction products of 1,3-cyclopentadiene. At the highest temperatures more than 10 mol% of 2,5-dimethylfuran is converted into mono-, di-, tri- and tetra-aromatic products, which are known soot precursors. This high tendency to form molecular weight growths species even under diluted conditions can pose a threat for the use of 2,5-dimethylfuran as a fuel.

Abstract A fully controlled, continuously operated mini-plant has been designed and constructed based on auger reactor technology. Three types of biomass fast pyrolysis experiments were performed in this set-up, viz. non-catalytic, in situ catalytic fast pyrolysis and ex situ upgrading of non-catalytic fast pyrolysis vapours by means of a downstream, moving-bed catalytic reactor. Pine wood was selected as the reference biomass feedstock. The mini-plant enables variation of the catalyst loading and contact times while producing larger samples suitable for full characterization in continuous operation. Due to short catalyst residence times and the fact that the biomass fed to the reactor is always brought into contact with fresh catalyst (in case of in situ catalysis) or the pyrolysis vapours always contact with fresh catalyst in the moving bed catalytic reactor (in case of ex situ catalysis), catalyst deactivation and coking are prevented ensuring improved experimental repeatability in catalytic pyrolysis experiments. The performance of the system was verified by in- and ex situ application of a single type of heterogeneous ZSM-5 based acidic catalyst while the non-catalytic results were taken as reference. Catalytic fast pyrolysis results in more effective oxygen removal from the bio-oil in the form of water, and resulting in lower yields of the organic fraction. Moreover catalytic decarbonylation reactions gave rise to increased gas yields. With in- and ex situ catalysis, increases in the char yields were observed as well as coke deposition on the catalyst surface. GC × GC-FID and GC × GC-TOF-MS analysis of the produced bio-oils showed that the use of in- and ex situ catalysis causes conversion of high molecular weight compounds to lower ones. Disappearance of detectable sugars and aldehydes, a decrease of the yield of acids, formation of phenols, and favoured aromatics production were the other catalytic effects observed.

Abstract The importance of 1,3-cyclopentadiene (CPD) and cyclopentadienyl (CPDyl) moieties in the growth of polycyclic aromatic hydrocarbons (PAHs) was studied using new experimental data and ab initio calculations. The experimental investigation was performed in a tubular continuous flow pyrolysis reactor under both high ( 24 mol N 2 / mol CPD ) and low ( 5 mol N 2 / mol CPD ) nitrogen dilutions, covering a temperature range of 873–1123 K, at a fixed pressure of 1.7 bara. At the most severe conditions up to 84% of CPD is converted, and the amount of PAHs is more than 65 wt%. Major products observed during CPD pyrolysis were benzene, indene, methyl-indenes and naphthalene, in line with previous studies. On-line GC × GC-FID/(TOF-MS) also allowed to quantify minor species (methane, toluene, styrene, phenanthrene, anthracene, etc.), never reported before at this level of accuracy. The new experimental data have been used to further analyze the role of the successive interactions of CPD, indene, and naphthalene as well as the recombination and addition reactions of their resonantly stabilized radicals and refine their kinetics. The results of the modeling study are in good agreement with existing and new experimental observations.

Bio-oils produced by fast pyrolysis of lignocellulosic biomass have proven to be a promising, clean, and renewable energy source. To better assess the potential of using bio-oils for the production of chemicals and fuels a new comprehensive characterization method is developed. The combination of the analyical power of GC×GC-FID and GC×GC-TOF-MS allows to obtain an unseen level of detail for both crude and hydrotreated bio-oils originated from pine wood biomass. The use of GC×GC proves to be essential to capture the compositional differences between crude and stabilized bio-oils. Our method uses a flame ionization detector to quantify the composition, while GC×GC-TOF-MS is used for the qualitative analysis. This method allows quantification of around 150 tentatively identified compounds, describing approximately 80% of total peak volume. The number of quantified compounds in bio-oils is increased with a factor five compared to the present state-of-the-arte. The necessity of using multiple internal standards (dibutyl ether and fluoranthene) and a cold-on column injector is also verified.

To check if an unconventional fuel can be burned in an engine, monitoring the stability in terms of composition is mandatory. When the composition of a conventional fuel cannot be measured for practical reason, it can be approximated using the API (American Petroleum Institute) relations (Riazi-Daubert) linking the hydrocarbon group fractions with well-chosen properties. These relations cover only the paraffin (coupling iso and normal), naphthene and aromatic (PNA) groups as they were developed for conventional fuels presenting neglected amounts of olefins and oxygenates. Olefins and oxygenates can be present in unconventional fuels. This paper presents a methodology applicable to any unconventional fuel to build a model to estimate the n-paraffin, iso- paraffin, olefin, naphthene, aromatic and oxygenate (PIONAOx) composition. The current model was demonstrated for an automotive shredder residues (ASR)-derived gasoline-like fuel (GLF). The model was trained using real fractions measured with a comprehensive two-dimensional gas chromatography coupled with flame ionization detector (GC × GC-FID) technique. The lowest cumulated absolute error comparing with the confidence interval of the measured fractions was evaluated to be 12.4%. The model was tested for one fuel composition only, therefore, the error of the calculated fractions will be investigated with other fuels in future work.