Soda- pulping Reaction of with free phenolic ß -O-4

ABSTRACT KEYWORDS The quinone methide from guaiacylglycol- ß-guaiacyl underwent Amines to the a-carbon with primary and secondary amines at Pulping 40°C. At pulping temperature, 170°C, only the primary amine adduct was Alkaline pulping detected. The quinone methide from guaiacylglycerol- ß-guaiacyl ether gave Quinones analogous adducts at 40°C, but no quinone methide-amine adducts were Delignification detected at 170°C. Instead, the major products were the vinyl ether and a substituted vinyl ether which resulted from a Mannich reaction of the vinyl ether, amine, and liberated formaldehyde.

John R. Obst Forest Products Laboratory, ForestService, U.S. Department of Agriculture, Madison,Wis. 53705

It has long been known that inclusion lowed to react with the quinone methide and can re-form the quinone methide in of certain amines in the soda pulping of generated from guaiacylglycol-ß-guaia­ soda liquor to ultimately yield the vinyl wood improves the selectivity of delig­ cyl ether, 1-(3-methoxy-4-hydroxy ether. The primary amine adducts are nification (1). More recently, the use of phenyl)-2-(2-methoxyphenoxy) ethanol more stable at high temperature than water-soluble primary amines has gen­ (I), and its diacetate. are those of the secondary amines. erated renewed interest in soda-amine Amines are known to undergo nucleo­ The reaction of amines with the vinyl pulping (2-4) and in the mechanism of philic addition to quinone methides (9). ether was unlikely, because the ether delignification. Wallis (5) has studied Reaction of the quinone methide (11) was formed in increasing amounts with the reaction of lignin-model compounds conveniently prepared from I diacetate time upon treatment of I with soda- with monoethanolamine (MEA), but the at 40°C (10) with primary and secon­ MEA (7). When the isolated vinyl ether significance of these reactions remains dary amines gave an adduct (III, Fig. 1) was treated with 33% n-propylamine to be demonstrated in alkaline-MEA in high yield (Table I). The vinyl ether and 0.5 N NaOH for 30 min at 170°C, no systems. Kubes et al. (6) considered a (IV) was the major product when a ter­ propylamine-quinone methide adduct delignification mechanism based on tiary amine was used. The n-propyla­ was detected. This lack of reversibility redox potentials of the spent pulping mine-quinone methide adduct was iso­ of the vinyl ether to re-form the quinone liquor but did not propose any chem­ lated and characterized by 1H and 13C methide is consistent with model stud­ ical mechanism. Obst and Sanyer (7) NMR, as well as by infrared and mass ies of kraft (11) and anthraquinone pulp­ have indicated that MEA may increase spectroscopy. Other primary and secon­ ing (12). delignification rate in a twofold man­ dary amine adducts were examined by The formation of a quinone methide­ ner: by reacting with quinone meth­ 1H NMR spectroscopy, and all spectra amine adduct (III) from I and primary ides, thereby preventing formation of were consistent with the general struc­ amines explains the low yields of guaia­ the vinyl ether with the beneficial effect ture III. col and vinyl ether previously reported of preventing lignin condensation re­ To more closely simulate wood pulp­ (7). Previous model studies showed simi­ actions; and by increasing the rate of ing conditions, model I was treated with lar reactions for both guaiacylglycol-ß­ cleavage of ß-ethers of etherified, or 0.33 N NaOH and various amines at ethers (I) and guaiacylglycerol-ß-ethers blocked, phenolic lignin moieties. This 170°C. A notable difference from the (V). Erythro-V triacetate gave a quinone paper presents the results of the reac­ results obtained at low temperature methide under mild conditions analo­ tions of free phenolic ß-O-4 ether lignin was that none, or only a trace, of the gous to that formed from I diacetate. models in soda-amine pulping liquors. secondary amine adduct was isolated This quinone methide also reacted with (Table II) and the yields of primary primary and secondary amines to give Resultsand discussion amine adducts were lower. When the an amine adduct (VI) (Table III, Fig. 2). n-propylamine-quinone methide ad- However, erythro-V treated with alka­ The reactions of quinone methides de­ duct (III, R = C3H7, R´ = H) prepared at li and amine at 170°C gave mainly IV rived from free phenolic ß-ethers in lig­ 40°C was treated with 0.5 N NaOH at and a C-5-substituted vinyl ether (VII) nin play a major role in determining 170°C for 30 min, only half of the adduct (Table III, Fig. 2) without any trace of delignification rates in soda, kraft, and was recovered. The diethylamine ad- quinone methide-amineadduct. The VII anthraquinone pulping (7). Also, the in­ duct (III, R = R´ = C2H5) treated under was formed by a Mannich reaction be­ creased delignification rate in soda-MEA the same conditions gave primarily the tween IV, formaldehyde liberated from pulping may in part be due to reaction vinyl ether and left only a trace of III. the terminal methylol group of the qui­ of the amine with quinone methides (7). Although amines are good , none methide, and amine. The VII was To test this proposal, amines were al­ they are also good leaving groups (9) also synthesized by treatment of either I

Reprinted from Tappi, The Journal of the Technical Association of the Pulp and Paper Industry. Vol. 64, No. 10, October 1981, Copyright, 1981 by TAPPI, and reprinted by permission of the copyright owner. PULPING or IV with formaldehyde and amine in I. Products from the reaction of the quinone methide (II) soda liquor at 170°C. Both n-propyla­ from I diacetate with amines at 40°C mine-quinone methide adducts (III and Yield, % of theory VI, R = C3H7, R'= H) prepared from the acetates at 40°C gave only IV upon Recovered digestion with 0.5 N NaOH at 170°C. Amine Vinyl starting Guaiacol, resulting from ß-ether adduct ether material cleavage of V, also underwent a Man­ Amine (Ill) (IV) Guaiacol (I), % nich reaction with formaldehyde and amine. Although only one substituted Methylamine 86 0 Trace 0 guaiacol, 3-methoxy-4-hydroxybenzyl­ n-Propylamine 89 0 5.9 0 Dimethylamine 85 0 0.1 0 n-propylamine (VIII), was isolated in Diethylamine 84 0 . . . 3% yield, primary amines, guaiacol, and 4 Trimethylamine 0 60 0.4 27 formaldehyde may react to give other Triethylamine 0 71 . . . products (13). These condensation re­ 17 actions explain the low yields of IV and guaiacol reported previously (7). The difference in the reactions of I II. Products from the soda reaction of the quinone and V in soda-amine liquor at 170°C methide from I with amines at 170°C may be that the rate of formation of Yield, % of theory vinyl ether from the quinone methide of V is much higher than the rate of Amine Vinyl amine nucleophilic addition. It has been adduct ether shown (14) that vinyl ether formation Amine (III) (IV) Guaiacol from V is faster than from I, as loss of Methylamine 72 5 10 formaldehyde is more facile than pro­ n-Propylamine 63 11 6 ton removal from the corresponding Dimethylamine 5 63 17 quinone methides. The formation of Diethylamine 0 62 20 vinyl ether from the amine adduct VI Trimethylamine 0 74 14 most likely proceeds through the qui­ Triethylamine 0 69 11 none methide of V since digestion of the methyl ether of VI (R = C3H7, R' = H) with 0.5 N NaOH at 170°C did not yield treating both I and V with soda-amine creased quinone methide formation in the methyl ether of IV. liquor relative to the yields upon treat- the presence of amines. Solvent effects The lower yields of guaiacol upon ing with soda may be a result of in- have been shown (15) to have a major

1. Soda-amine (R = CH3, C2H5, C3H7; R' = C2H5, C2H5) reactions of guaiacylglycol-ß -guaiacyl ether (I).

100 Vol. 64, No. 10 October 1981 / Tappi III. Products from the soda amine reactions of effect on product distribution on treat- V triacetate at 40° C and V at 170° C ment of I with base. Tertiary amines Yield, % of theory have been used as solvents to promote quinone methide formation (9). Amine Vinyl Substituted adduct ether vinyl ether Amine (VI) (IV) (VII) Guaiacol Summary and conclusions V triacetate at 40° C Guaiacylglycol- ß-guaiacyl ether (I) di­ gested in alkali with primary amines Methylamine 81 0 0 0 at 170°C gave III via nucleophilic ad­ n-Propylamine 79 0 0 0 dition to the quinone methide. The qui­ Dimethylamine 82 0 0 0 none methide from I diacetate prepared Diethylamine 79 0 0 0 at 40°C underwent nucleophilic addi­ tion in high yield with both primary Vat 170°C and secondary amines. The quinone Methylamine 0 51 26 3.9 methide from guaiacylglycerol-ß-guai­ n -Propylamine 0 54 22 4.1 acyl ether (V) triacetate at 40°C reacted Dimethylamine 0 24 56 3.9 analogously to that from I diacetate. In Diethylamine 0 52 26 2.6 contrast, when V was digested in soda­ amine liquor at 170°C, no quinone

2. Soda-amine (R = CH3, C2H5, C3H7; R' = CH3, C2H5) reactions of guaiacylglycerol-ß -guaiacyl ether (V)

Tappi / October 1981 Vol. 64, No. 10 101 PULPING methide addition product was detected. NMR data of the acetylated adduct (VII) prepared with methylamine and Instead the major products were IV showed these major differences: δ 3.12 n-propylamine coincidentally gave cis- and a C-5-substituted vinyl ether (VII), (t, C-1 propyl methylene, J = 7 Hz); Hα at the same chemical shift as for IV. a Mannich reaction product of IV, 4.57 (d, Hß, J= 6); and 5.73 (t, Hα, J = 6). Guaiacol and IV yields were deter­ amine, and formaldehyde split from The Hα of the acetylated adduct also mined by gas chromatography. Yields the side chain of V. gave a poorly resolved, low-intensity from larger-scale reactions were deter­ The substituted vinyl ether could also triplet at δ 5.3, which results from re­ mined by column chromatography. be formed by addition of formaldehyde stricted rotation of the C-N amide bond. The substituted vinyl ether, VII, was 13 to the soda-amine digestion of both I The C NMR data (15.0 MHz, CDCl3) of also prepared in 75-85%yield from both and IV. Some of the guaiacol cleaved the adduct: δ 11.8 ppm (propyl methyl); I and IV by treatment with a twofold from V also underwent Mannich substi­ 23.1 (C-2 propyl methylene); 49.4 (C-1 excess of formaldehyde at 170°C in 40% tution. The difference in the 170°C propyl methylene); 55.9 (methoxyls); amine-cellosolve-1.0N NaOH (1:1:1). 1 soda-amine reactions of the free pheno­ 62.3 (Cα); 74.8 (Cß); 110.1-150(12 aryl The H NMR spectra were deter­ lic ß-ethers with (V) and without (I) the carbons). Mass spectrum: m/e (%); 331 mined on a Varian T-60 spectrophotom­ terminal methylol group leads to the (M+, 0.4); 195(12.7); 194(100); 152(3.2). eter with TMS as reference. The 13C strong suggestion that lignin-model The infrared spectrum of the acetylated NMR spectra were run on a Jeol FX 60 studies be performed with the more adduct contained two carbonyl absor­ spectrophotometer. Infrared spectra appropriate model, guaiacylglycerol­ bances: 1770 cm-1 (aryl acetate) and were run as films on sodium chloride ß-guaiacyl ether (V). 1645 cm-1 (amide). windows on a Beckman IR-12. Mass The results obtained explain the low Other amine adducts (III) and their spectral analyses by Raltech Scientific yields of guaiacol and vinyl ether from acetylated derivatives were prepared Inc., were performed on a Varian MAT soda-MEA digestion of I and V report­ and characterized by 1H NMR. Adducts 112 double-focusing mass spectropho­ ed previously (7). The suggestion that formed with primary amines and then tometer at 80 eV. amines could act as quinone methide acetylated gave 1H NMR spectra that scavengers-preventing some lignin had two resonances each for the amide Literature cited condensations and thereby promoting methyl, N-, and side-chain Hα re­ delignification–is therefore unlikely. sulting from restricted rotation of C-N 1. Wise, L.E., Peterson, F.C., and Harlow, A few lignin quinone methides, if they amide bonds. Side-chain and methoxyl W.M., Ind. Eng. Chem. Anal. Ed. 11: 18 1 are incapable of eliminating formal­ H NMR (60 MHz, CDCl3) of the acetyl­ (1939). 2. Chuiko, G.V., Chupka, E.I., and Nikitin, dehyde-for example, partially ether­ ated derivatives: III (R = CH3, R' = δ V.M., Bum. Prom. 8: 7 (1972). ified pinoresino structures-might CH3CO), 2.15 and 2.25 ppm (s, amide form quinone methide-amine adducts. methyl); 2.25 (s, acetate methyl); 2.8 3. Bolker, H.I., and Kubes, G.J., Cellul. Chem. Technol. 12: 621 (1978). The efficient reaction of amines with and 2.9 (s, N-methyl); 3.79 (s, meth­ 4. Julian, L.M., and Sun, B.C.H., Tappi formaldehyde and subsequent substitu­ oxyls); 4.45 (d, Hß, J = 5.8 Hz); 5.32 (m, 62: (8): 63 (1979). tion at reactive C-5 positions, especial­ Hα ), and 5.8 (t, Hα, J= 5.8); III (R = R' 5. Wallis A.F., Cellul. Chem. Technol. ly for secondary amines, may be bene­ = CH3), 2.27 (s, acetatemethyl); 2.41 (s, 10: 345 (1976). ficial to delignification by decreasing N-methyls); 3.76 and 3.81 (s, meth­ 6. Kubes, G.J., Fleming, B.I., MacLeod, the amount of lignin condensation. oxyls); 4.34 (m, Hα, Hß); III (R = R' = J.M., and Bolker, H.I., Tappi 61(8): 46 C2H5), 1.02 (t, ethyl methyl, J= 7); 2.27 (1978). (s, acetate methyl); 2.7 (q, ethyl meth­ 7. Obst, J.R., and Sanyer, N., Tappi 63(7): Experimental ylene, J = 7); 3.76 and 3.8 (s, methoxyls); 111 (1980). 8. Turner, A.B., Quarterly Rev. 18(4): 347 4.22 (m, Hα, Hß). The quinone methides from the ace­ (1964). tates of I and V were prepared at 40°C The triacetate of V at 40°C with pri­ 9. Wagner, H.-U., and Gompper, R., in in amine-dioxane-aqueous sodium hy­ mary or secondary amines and alkali “The Chemistry of the Quinonoid Com­ droxide (1:1:1) in a manner similar to gave erythro and threo amine adducts pounds,” (Saul Patai, ed.), Wiley, New one previously described (11). If two (VI) with and without the terminal ace­ York, 1974, p. 1145. phases resulted, additional water was tate. No separation of these four prod­ 10. Gierer, J., Lindeberg, O., and Noren, I., added until the solution was homo­ ucts was attempted, and yields were Holzforschung 33: 213 (1979). geneous. Reactions at 170°C were typi­ estimated from gravimetric recovery 11. Gierer, J., Lenz, B., and Wallin, N.-H., cally run with 40 mg of I, 2.0 ml of 40% and 1H NMR spectra. The absence of Acta Chem. Scand. 18: 1469 (1964). guaiacol was confirmed by thin layer 12. Obst, J.R., Landucci, L.L., and Sanyer, amine, and 1.0 ml of 1.0 N NaOH in N., Tappi 62(1): 55 (1979). stainless steel tubes tumbled in an oil chromatography. 13. Burke, W.J., Smith, R.P., and Weather- bath for 30 min. In the case of phase Dimethylamine and Vat 170°C gave bee, C., J. Amer. Chem. Soc. 74: 602 separations, less sodium hydroxide was the substituted cis- and trans-vinyl (1952). used or cellosolve was included as co­ ethers (VII, R = R' = CH3), which were 14. Gierer, J., and Ljunggren, S., Svensk solvent. Diethylamine in aqueous alkali also isolated by column chromatog­ Papperstid. 82(17): 503 (1979). gave one phase at 20°C, but when heated raphy on silica gel. The 1H NMR data 15. Fullerton, T.J., Svensk Papperstid. 78 δ (6): 224 (1975). in a sealed tube, two phases resulted at (60 MHz, CDCl3): 2.34 ppm (s, N­ about 90°C and persisted to 170°C (16). methyl); 3.6 (s, N-methylene); 3.9 (s, 16. Obst, J.R., Tappi 64(3): 171 (1981). Gas chromatographic analysis was per­ methoxyls); 5.48 (d, Hα, Jcis = 7 Hz); 6.2 The FPL is maintained at Madison, Wis., in formed on acetylated products (12). (d, Hα, Jtrans = 12.4); 6.4-7.5 (m, Hαcis cooperation with the University of Wisconsin. The quinone methide-n-propyla­ and trans, 6H aryl). Mass spectrum: + mine adduct (III, R = C3H7, R' = H) was m/e (%), 329 (M , 98); 284 (17); 255 (26); The author is grateful to L. L. Landucci for obtained as a light yellow oil after col­ 241 (36); 176 (49); 118 (39); 77 (73); 58 providing model I, J. R. Ralph for providing erythro-V and for 13C NMR and mass spectra umn chromatography on silica gel with (72); 44 (100). interpretation, L. C. Zank and M. F. Wesolowski chloroform as eluant. The 1H NMR data Compounds VI and VII prepared for instrumental analyses, and N. Sanyer for (60 MHz, CDCl3) were as follows: 6 0.91 with other primary and secondary discussion of this manuscript. ppm (t, propyl methyl; J = 7 Hz); 1.52 amines were characterized by 1H NMR This article was written and prepared by U.S. Government employees on official time and it is (sextuplet, C-2 propyl methylene, J = (60 and 270 MHz), and results consis­ therefore in the public domain. 7); 2.48 (t, C-1 propyl methylene, J = 7); tent with the proposed general struc­ 3.86 (s, methoxyl); 4.02-4.12 (m, Hα, tures were obtained. The 1H NMR spec­ Received for review Nov. 3, 1980. 1 H ß); and 6.88-7.1 (m, 7H aryl). The H trum of the substituted vinyl ethers Accepted April 14, 1981.

102 Vol. 64, No. 10 October 1981 / Tappi