OCR Text |
Show TJU -, ^ : 30 -, > yi - 40 - - 20 - 2 -1 • 5 H 5? Yields, w 1 _ 3 05-j 0 f 1 00-^ Cellulose __ 0 • \^ y • . 1 MP • .1 3 101 K/s ^= Cellulose. &^**"a V fr^ a / / a 7 * / '4 -'*'_>- '^^g^^lL 0 1 MPa • • • ^ a 10° K/'s " - Tar Gas I 1] Char j - • C,H4 C,H, •C2H, 400 500 500 700 Temperature, C 800 400 500 600 700 Temperature, C 800 igure 4 Application of the reaction mechanism in bio-FC to the product distnbution from the devolatilization of pure cellulose at roughly 1000 °C/s to the stated temperatures with no isothermal reaction period. Cooling rates were assigned as 200 3C/s (Hajaligol. 1982). In clockwise order, the panels show the major products, oxygenated gases, CMS. are nvdrocarbons Grasses, that tend to have the highest ash contents, also tend to have the lowest P V M values. But the assigned cellulose fractions of all biomass forms tend to be widely variable. Cases with high cellulose fractions can exhibit very low and very high P V M values. The assigned compositions of the ngnm-like components also vary significantly. The H-numbers range from 8.4 to 12.5, whereas the O-numbers range from 1 to 6 (per 9 C-atoms). A second aspect of our two-component hypothesis for piomass compositions is illustrated in Fig. 4. All the parameters in Bio-FC's reaction mechanism for cellulose were assigned by fitting the dataset reported by Hajaligol et a (1982) This dataset was selected because the mass and elemental closures are satisfied within tight tolerances, so the product distnbution is complete, and the cellulose was ash-free The heating rate of 1000 °C/s is also relevant to suspension finng applications, although the thermal histories n these tests are not simple. The power to the wire mesh neater was turned off as soon as the ultimate reaction temperature was achieved, then the samples cooled at approximately 200 5C/s. Devolatilization dunng these extended cooimg periods is definitely not negligible, expeciaiiy since the devolatilization rates with biomass are -eiativeiy very fast. This mechanism is able to correlate the entire product distribution within experimental uncertainty throughout all stages of cellulose pyrolysis. This includes the transient mass partitioning among char, tar, and noncondensible gases; the transient yields of CO, H 20, and CC2 . the transient yields of the oxygenated oils, and the C2-C3 hydrocarbon gases. The transient yields of C H 4 and H : from this dataset are descnbed within similar tolerances but were omitted due to lack of space. Baseline values for the rate parameters for the lignin-like component were assigned with a similar procedure using the analogous dataset 'or milled wood lignin reported by Nunn et al. (1985) After the reaction mechanisms for both piomass components were fully specified, the impact of ash eataiys.s was addressed in a calibration procedure. Given detailed mineral compositions, ash catalysis could be implemented mechanistically based on several studies that point to a predominant impact of calcium and alkali cations (Raveendran, 1995; Nik-Azar, 1997; Jensen '998, Unfortunately, such detailed information is almost never reported when biomass is utilized m utility applications Furthermore, the reported database on detailed mineral compositions in vanous forms of biomass shows dearly that the amounts of the most active cations do not correlate with the total ash levels within useful quantitative tolerances (Raveendran, 1995). W e devised the following calibration procedure to circumvent these ambiguities. After the two component proportions and compositions have been assigned, tne P VM of the whole biomass is regarded as an indication of the impact of mineral catalysis. Rate parameters in the charnng mechanism for both components are adjusted to match the |