From biomass to biofuels
The FP7 DIBANET (Development of Integrated Biomass Approaches Network) project focused on helping to reduce Europe’s dependency on the import of fossil fuels. The €3.7m scheme, which included partners from both Europe and South America, focused on developing the production of Ethyl-Levulinate (EL) from organic wastes and residues. According to the project’s website, EL is a ‘novel Diesel Miscible Biofuel (DMB) produced by esterifying ethanol with levulinic acid’.
In a summary report on the project, Dr Daniel Hayes of the Carbolea Group at the University of Limerick (UL) in Ireland, the organisation co-ordinating the project, outlined the background to the Biofine process.
The DIBANET proposal was focused on developing sustainable and profitable means for obtaining biofuels and platform chemicals from lignocellulosic biomass. It considered that acid hydrolysis was an effective tool for this, since it would allow for a wide variety of different feedstocks to be processed. It would also allow the process conditions to be engineered so that levulinic acid could be produced in high yields from cellulose/hexoses and furfural could be produced in high yields from hemicellulose/pentoses.
UL researchers have developed robust kinetic models for the acid degradation of biomass. These show the critical failings in the Biofine process that have meant it has never been successfully commercially deployed, despite being developed as far back as 1988.
The DIBANET proposal considered that an improved pre-treatment method might be important and suggested the use of ionic liquids. Theoretically, these could be of use since they allow for fractionation of the lignocellulosic polymers. However, ionic liquids were quickly rejected, as the project got underway due to poor efficiencies and high cost. The alternative (now patented) pre-treatment method that was subsequently developed has huge benefits. The pre-treatment allows the fractionation of cellulose from the lignin and hemicellulose. It can then be processed independently at conditions that are optimal for high levulinic acid yields whilst the same can be done for the hemicellulose, allowing for significantly increased furfural yields. Furthermore, the lignin is separated but kept intact (unlike in the Biofine process) and is of an organosolv-type quality and of high value.
By increasing the yields of high-value chemicals from hexoses and pentoses and by obtaining lignin as a separate saleable product, DIBANET vastly reduces the amount of low-value Acid Hydrolysis Residues (AHRs) that are produced. The models show that one tonne of Miscanthus will produce 151kg of AHRs when put through the DIBANET process, compared with 517kg when put through the Biofine process.
Fast pyrolysis experiments that used AHRs gave poor quality bio-oils at low yields, primarily due to the low hydrogen contents of these materials. On advice from an independent reviewer, this option for the DIBANET process chain was dropped midway through the project. The focus of the project had shifted substantially because of the significantly lower AHR yields that were achieved. It was a condition of DIBANET that all process needs would need to be met either from the residues of the process or from alternative renewable energy sources. Under a Biofine-type, level of production of AHRs it was found that the process energy needs of Biofine were so high that even AHRs equivalent to approximately 50% of the mass of the original biomass would not be sufficient. Indeed, the model suggests that an additional 572kg of biomass would need to be combusted, per tonne of biomass used in the Biofine process, to satisfy the energy requirements of that technology.
While the energy needs for DIBANET are significantly less than for Biofine, the AHR levels are so low that extra biomass is also required. In such a situation it is important that as much of the process energy requirements as possible are provided from the AHRs, so minimising the need for extra biomass. The most efficient way of providing the low-pressure steam and modest temperatures required in the DIBANET pre-treatment/hydrolysis processes is through the direct combustion of AHRs/biomass.
The separate gasification or pyrolysis of virgin biomass is not of relevance to this modelling deliverable or to the DIBANET project since the hydrolysis stage is the core concept and all other areas need to relate to that for a truly integrated process to be developed.
Therefore, the targets for the modelling of the DIBANET process chain shifted substantially from those considered at the proposal stage. Instead of concentrating on how to utilise the AHRs the focus shifted to considering what would be the most profitable outputs of the pre-treatment and hydrolysis processes and how these products should be utilised.
The DIBANET process can be separated into three distinct stages: (1) pre-treatment; (2) hydrolysis of cellulose for levulinic acid development; and (3) esterification of levulinic acid for EL production. It has been demonstrated that the pre-treatment process can be a financially viable technology in its own right, particularly when feedstocks with high pentose contents (e.g. bagasse) are used. Combining stages (1) and (2) increases capital costs but also increases the annual profits and the NPV. Combining all three stages can also be financially rewarding, providing that the value of the furfural and lignin co-products are fully exploited. Modelling has shown that the internal energy balance of the process is improved by combusting part of the lignin to make up the energy shortfall that results from the low amounts of AHRs that are produced. However, the modelling has also shown that it is economically far superior to sell this lignin and to purchase additional biomass to fuel the process instead (even with feedstocks, such as Miscanthus, that have higher costs).
The attractive financials of the DIBANET processes also mean that these can potentially be operated profitably at lower scales of operation than would be possible for many other lignocellulose-processing technologies. Indeed, there are possible scenarios in which a demonstration-scale plant could pay back the investment cost and provide a healthy NPV in addition to proving the process on an enhanced scale.
The end result has been the development of a core IP that has real potential for commercial deployment. This IP can also occupy an important niche in the biorefining sector in that it allows for the low-cost production of levulinic acid, something that has been long anticipated but not fulfilled to date.
Dr Daniel Hayes
