Frontier Orbital Interpretation of Gas Release Pathways in Lignin


Frontier Orbital Interpretation of Gas Release Pathways in Lignin...

0 downloads 55 Views 2MB Size

Article pubs.acs.org/EF

Frontier Orbital Interpretation of Gas Release Pathways in Lignin Thermolysis Preetinder S. Virk* and Michael T. Klein† Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, United States ABSTRACT: An attempt was made to interpret the experimentally observed gas release pathways in lignin thermolysis by the frontier molecular orbital (FMO) theory. A set of 20 monoaromatic molecules that mimic the moieties in Freudenberg and Harkin’s model of spruce lignin were thermolyzed at 250−550 °C and fractional conversions from 0.0005 to near unity. Guaiacol substrate, a model of the prevalent coniferyl alcohol monomer residues in lignin, thermolyzed by two parallel paths, namely, (R1) demethanation to CH4 and catechol and (R2) decarbonylation to CO and phenol, with their relative rates in the ratio of ∼5:1. Thermolyses of three substituted guaiacols, namely, 2,6-dimethoxyphenol, isoeugenol, and vanillin, also resulted in demethanation and decarbonylation, analogous to R1 and R2 for guaiacol, but vanillin decarbonylation, to guaiacol and CO, was more akin to and roughly 1000 times faster than the single decarbonylation path (R3) of benzaldehyde to benzene and CO. Pathway R1 was mechanistically interpreted as a thermally allowed pericyclic group transfer elimination of CH4. A FMO diagram for R1 suggested a dominant electronic interaction between highest occupied molecular orbital (HOMO) (methane) ↔ lowest unoccupied molecular orbital (LUMO) (o-benzoquinone), with an HOMO−LUMO “gap” of ∼11 eV. The demethanation kinetics of 2,6-dimethoxyphenol, isoeugenol, and vanillin were essentially the same as observed for guaiacol, implying that their substituents did not appreciably alter the energies of their LUMOs relative to the parent.





INTRODUCTION This work is an initial attempt to interpret the experimentally observed gas release pathways in lignin thermolysis by Fukui’s frontier molecular orbital (FMO) theory.1 A set of 20 monoaromatic molecules that mimic the moieties in Freudenberg and Neish’s model of spruce lignin,2 as elaborated by Harkin,3 was experimentally thermolyzed at T = 250−550 °C and fractional conversions from 0.0005 to near unity. The kinetics of substrate decomposition and gas and liquid product appearances were delineated, in hopes of deriving generalized pathways for CH4, CO, H2O, and CO2 formation. The chosen model substrates are shown in Figure 1, and their decomposition kinetics and pathways have earlier been summarized by Klein and Virk4 and detailed by Klein.5

EXPERIMENTAL SECTION

The experiments using guaiacol substrate, a model of the prevalent coniferyl alcohol monomer residues in lignin, will be presented first, as a prelude to the theoretical FMO interpretation. Figure 2, a plot of

Figure 2. Guaiacol thermolysis pathways. Special Issue: In Honor of Michael J. Antal Received: May 31, 2016 Revised: July 15, 2016 Figure 1. Lignin model compounds thermolyzed. © XXXX American Chemical Society

A

DOI: 10.1021/acs.energyfuels.6b01327 Energy Fuels XXXX, XXX, XXX−XXX

Energy & Fuels



normalized product appearances, shows that, at low fractional conversions, 0.0005 < X < 0.1, guaiacol thermolyzed by two parallel paths, namely, (R1) demethanation to CH4 and catechol and (R2) decarbonylation to CO and phenol, with their relative rates in the ratio of ∼5:1 (at much higher conversions, 0.5 < X < 1.0, not of the present interest, a solid coke product was observed and selectivity of the catechol product was diminished relative to CH4, CO, and phenol). Next, Figure 3 depicts the kinetics of guaiacol thermolysis, showing

Article

RESULTS AND DISCUSSION

The preceding experimental results suggest that demethanation of a guaiacyl moiety in lignin might reasonably be interpreted as a thermally allowed pericyclic 6e(2σ + 2π + 2σ) group transfer elimination of CH4, as depicted in Figure 5 for the general

Figure 3. Guaiacol thermolysis kinetics. Figure 5. Demethanation of a guaiacyl moiety by 6e(2σ + 2π + 2σ) group transfer elimination.

each of the two parallel pathways discerned in Figure 2 to be approximately first-order in guaiacol over a roughly 10-fold range of the initial substrate concentration, 0.46 < Co (mol/L) < 3.07. Finally, with regard to experiments, Figure 4 is an Arrhenius plot for

case (G), the prototypical guaiacol case (P), and the model case (M) of H2 elimination from cis-but-2-ene, with Arrhenius parameters for the latter from Masson et al.6 FMO analysis of the preceding group transfer elimination is most conveniently affected by its microscopic reverse 6e group transfer addition, the transition state of which closely resembles that of a 6e(4π + 2π) Diels−Alder cycloaddition. Figure 6 is a

Figure 4. Arrhenius parameters for methoxybenzene thermolysis pathways. Figure 6. FMO interaction diagram for butadiene−hydrogen group transfer.

methoxybenzene thermolysis pathways, including also results for anisole substrate, which was more refractory than guaiacol and exhibited far lower selectivity to CH4, with a ratio (CH4/phenol) = 0.2 ± 0.1. Results in Figure 4 were obtained at fixed initial Co = 0.45 mol/L over a range of 300 < T (°C) < 550, which caused a 5 order of magnitude variation of the first-order guaiacol demethanation rate constant k1, leading to an Arrhenius expression log k1 (1/s) = (10.9 ± 0.5) − (43.7 ± 0.5)/θ

FMO interaction diagram for the butadiene−hydrogen group transfer, using molecular orbital pictures from Salem and Jorgensen7 and orbital energies from Pearson.8 Of the pair of frontier orbital interactions, the highest occupied molecular orbital (HOMO) (butadiene) ↔ lowest unoccupied molecular orbital (LUMO) (hydrogen) “gap” = 11.1 eV < HOMO (hydrogen) ↔ LUMO (butadiene) “gap” = 16.0 eV; therefore, the former will “control” the stabilizing E(FMO) < 0 and, hence, the observed electronic activation energy E*.

(1)

where θ = 0.00457T (K) is Benson’s scaled temperature for the activation energy E* expressed in kcal/mol. B

DOI: 10.1021/acs.energyfuels.6b01327 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels

Figure 7. FMO interaction diagram for o-benzoquinone−methane group transfer.

Along analogous lines, Figure 7 is a FMO interaction diagram for the o-benzoquinone−methane group transfer, using reasonable (but uncertain for o-benzoquinone, especially) values of E(HOMO) and E(LUMO). In this case, of the pair of frontier orbital interactions, HOMO (methane) ↔ LUMO (o-benzoquinone) “gap” = 10.8 eV < HOMO (o-benzoquinone) ↔ LUMO (methane) “gap” = 17.4 eV; therefore, the former will “control” the stabilizing E(FMO) < 0 and, hence, the observed electronic activation energy E*. Two curious consequences, not hitherto mooted in the literature, follow. First, the methanation here has an “inverse” electron demand relative to the hydrogenation; this would suggest that, because substituents generally perturb the HOMO far more than the LUMO, the methanation (and its microscopic reverse, demethanation) of guaiacyl moieties should be relatively insensitive to further substitution. Second, in both butadiene hydrogenation and o-benzoquinone methanation cases, the dominant gap is roughly the same at ∼11 eV; therefore, the FMO theory would predict that both reactions should have comparable activation energies and, hence, similar kinetics. Data summarized at the bottom of Figure 5 show that, whereas, at T = 400 °C, guaiacol demethanation is almost 5 orders of magnitude faster than cis-but-2ene dehydrogenation, the kinetics of the reverse methanation and hydrogenation reactions, estimated by microscopic reversibility, are, as predicted, of the same order of magnitude. Finally, Figure 8 presents a comparison between the demethanation and decarbonylation kinetics, at fixed T = 400 °C and initial substrate concentrations Co ∼ 0.5 mol/L, of guaiacol, three substituted guaiacols, namely, 2,6-dimethoxyphenol of Fukui type X (electron donating), isoeugenol of Fukui type C (conjugated), and vanillin of Fukui type Z (electron withdrawing), and also benzaldehyde (control). Demethanation kinetics of guaiacol, 2,6-dimethoxyphenol, isoeugenol, and vanillin by pathway R1 were virtually identical, with log k (400 °C) = −3.2 ± 0.1. This observed insensitivity to substituents is in accordance with the earlier theoretical interpretation of R1 as a thermally allowed group transfer elimination with HOMO (methane) ↔ LUMO (o-benzoquinone) dominant and LUMO energies a little altered by substituents. Turning to decarbonylation by pathway R2, guaiacol, 2,6-dimethoxyphenol, and isoeugenol exhibited almost identical kinetics, with log k (400 °C) = −3.8 ± 0.1. However, vanillin decarbonylation, to guaiacol and CO, was roughly 1000 times faster than decarbonylation of guaiacol alone and seemed more akin to the single decarbonylation path (R3) of benzaldehyde to benzene and CO.

Figure 8. Substituent effects on demethanation and decarbonylation pathways.



SUMMARY AND CONCLUSIONS (1) An attempt was made to interpret the experimentally observed gas release pathways in lignin thermolysis by Fukui’s FMO theory.1 (2) A set of 20 monoaromatic molecules that mimic the moieties in Freudenberg and Neish’s model of spruce lignin,2 as elaborated by Harkin,3 were thermolyzed at T = 250−550 °C. The kinetics of substrate decomposition and gas and liquid product appearances were delineated, with an attempt to derive generalized pathways for CH4, CO, H2O, and CO2 formation. (3) Guaiacol substrate, a model of the prevalent coniferyl alcohol monomer residues in lignin, thermolyzed by two parallel paths, namely, (R1) demethanation to CH4 and catechol and (R2) decarbonylation to CO and phenol, with their relative rates in the ratio of ∼5:1. (4) Pathway R1 was mechanistically interpreted as a thermally allowed pericyclic 6e group transfer elimination of CH4. (5) A FMO diagram for R1 revealed the dominant electronic interaction between HOMO (methane) ↔ LUMO (o-benzoquinone) and corresponding HOMO−LUMO “gap” of ∼11 eV. (6) 2,6Dimethoxyphenol, isoeugenol, and vanillin all exhibited essentially the same demethanation kinetics as guaiacol, implying that their respective type X, C, and Z substituents did not appreciably alter the energies of their LUMOs relative to that of the parent. (7) Decarbonylations of guaiacol, 2,6-dimethoxyphenol, and isoeugenol by R2 exhibited almost identical kinetics, but vanillin decarbonylated roughly 1000 times faster than guaiacol, by a path more akin to the single decarbonylation path (R3) of benzaldehyde to benzene and CO.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Present Address †

Department of Chemical and Biomolecular Engineering, University of Delaware, 250G Interdisciplinary Science and Engineering Building, 221 Academy Street, Newark, Delaware 19716, United States.

Notes

The authors declare no competing financial interest. C

DOI: 10.1021/acs.energyfuels.6b01327 Energy Fuels XXXX, XXX, XXX−XXX

Article

Energy & Fuels



REFERENCES

(1) Fukui, K. Theory of Orientation and Stereoselection; Springer-Verlag: New York, 1975. (2) Freudenberg, K.; Neish, A. C. Constitution and Biosynthesis of Lignin; Springer-Verlag: New York, 1968. (3) Harkin, J. M. In Oxidative Coupling of Phenols; Taylor, W. I., Battersby, A. R., Eds.; Marcel Dekker, Inc.: New York, 1967; pp 243− 322. (4) Klein, M. T.; Virk, P. S. Energy Fuels 2008, 22, 2175. (5) Klein, M. T. Model Pathways in Lignin Thermolysis. Sc.D. Thesis, Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA, 1981. (6) Masson, D.; Richard, C.; Martin, R. Int. J. Chem. Kinet. 1976, 8, 37. (7) Salem, L.; Jorgensen, W. L. The Organic Chemist’s Book of Orbitals; Academic Press: New York, 1973. (8) Pearson, R. G. Inorg. Chem. 1988, 27, 734−740.

D

DOI: 10.1021/acs.energyfuels.6b01327 Energy Fuels XXXX, XXX, XXX−XXX