'AGRICULTURE ROOM
REACTION OF HARDWOO D
LIGNIN WITH HYDROGE N
■
)I
1
UNITED STATES DEPARTMENT OF AGRICULTUR E
j'OREST SERVIC E
FOREST PRODUCTS LABORATOR Y
Madison, Wisconsi n
In Cooperation with the University of Wisconsi n
June 1938
SCHOOL Of i )ES~- Rr
OREGON STATE COLLEG E
CORVALLIS, OREGON
•
REACTION OF HARDWOOD LIGNIN WITH HYDROGENl
By
E . E . HARRIS, Chemist ,
JAMES D'IANNI, Graduate Student ,
and
HOMER ADKINS, Professor of Organic Chemistry ,
University of Wisconsin .
Lignin was found to react with hydrogen over copper-chromium oxide ,
absorbing one mole of hydrogen for each 25 g, of lignin . The reaction take s
place in dioxane solution at 250-260° under 200-350 atmospheres of hydroge n
during a period of about 18 hours . The mixture of products of the reactio n
is colorless or faintly yellow .
There was obtained from the hydrogenation of g 0 g . of hardwoo d
lignin 22 g . of methanol, 9 g . of an alcohol, b .p . 92-95° (7 mm .), about 3 g .
of a glycol, b .p . 107-110° (1 mm .), 20 g . of a glycol, b .p . 125-127° (1 mm .) ,
1g g . of a mixture of compounds ., b . p . 1300 (1 mm .) to 260° (1 mm .), 4+ g . o f
intermediate fractions and 5 g . of compounds boiling above 260° (1 mm .) .
The alcohol has been identified as 4-n propylcyclohexanol-l, (I) .
The analysis of the glycols for carbon and hydrogen corresponds to th e
OH
/
OH
OH
/
C3H7
CH2CH2CH2OH
C3H7
I
II
II I
formula C9H1 g02 . The one present in the larger amount contains a primar y
hydroxyl. group and has been oxidized to a keto acid, C9H1i03 . The glyco l
is stable toward hydrogenolysis so that it cannot be a 1,2- or 1,3-glycol ,
but the secondary hydroxyl may be in either of the three positions on th e
cycle . The secondary hydroxyl is tentatively placed in the 4-position a s
0
1
-Published in the Journal of the American Chemical Society, June 193 g
R1176
, ;C
Faye+G ST
G
a
"ORVI LLIS, OREGON
•1
in formula II, i .e ., it is 3-(4-h droxylcyclohexyl)-propanol-1 . The secon d
glycol has been characterized as -npropylcyclohexanediol-1,2 as s1aowh i n
formula III . These three compounds together with the methanol in th e
dioxane fraction constituted over 70 percent of the total weight of ligni n
subjected to hydrogenation .
The analysis for carbon and hydrogen of the mixture of product s
distilling above 130 0 (1 mm .), corresponds to the formula (C6 H11O ) n . Thi s
material does not contain methoxyl groups and the boiling range of the por tion distilled (i .e ., about 75 percent of the fraction) indicates that n has
a value of 3 to 54 The analysis fbr carbon and hydrogen corresponds almos t
exactly to an average composition of C 1 gH3203 or C2)H)42011. .
The products boiling above 130° (1 mm .) were dehydrated over
alumina at 400° and then hydrogenated over Raney nickel . A mixture of
hydrocarbons was obtained of the formula CnH2n_2 or CnH2n_ 4 boiling ove r
a wide range above 70° (1 mm .) . Most of these hydrocarbons distilled i n
the range 9 00 (1 mm .) to 1)E00 (1 rem.) . These facts indicate very clearl y
that there are in lignin, units containing more than nine carbon atoms ,
for the CnH2n_2 hydrocarbon 1,2-dicyclohexylethane containing fourteen carbo n
atoms, distilled under exactly the same conditions, boiled at 80-8 10 (1 mm .) .
It remains to be considered wh t structures fox lignin are i n
harmony with the results of hydrogenationjus suEmar
o
o na: A variety of
i
com.p ounds have been investigated with respect to their behavior towar d
Zy rogen over copper-Chromium oxide at elevated temperatures and pressures ,
i .e ., esters, alcohols, and glycols (Folkers, Wojcik, Connor, and Sprague) ,
sugars (Zartman), ethers (Van Duzee and Musser), furans (Burdick) an d
benzopyrans (Mozingo) .?
For the present purpose these studies may be summarized as follows :
an oxygen to carbon linkage is very labile toward hydrogenolysis over copper chromium oxide at 250° if it is (a) in the 1,2-position with respect to a
double bond in the 3, )4-position, i .e . ,
4 3 2
1
,
C = C - C . . . .0- ;
{4) in the 1,2-position with respect to an oxygen to carbon linkage in th e
The papers published prior to 1937 are summarized and correlated in a mono graph : (a) Adkins, "Reactions of Hydrogen with Organic Compounds over Cop per-Chromium Oxide and Nickel Catalysts," Univ . of Wisconsin Press, Madi son, Wis ., 1 937 ; (b) Folkers and Adkins, Jour . Amer . Chemical Society, ~ ,
1651 (1932) ; (c) Connor and Adkins, ibid .,
,4678 (1932) ; (d) Wojcik an d
Adkins, ibid., 55, 1293 (1933) ; (e) Zartman and Adkins, ibid., aa, )+559
(1933) ; (f) Wojcik and Adkins, ibid ., 5a, 4939 (1933) ; (g) Burdick an d
Adkins,
)138 {1934) ; (h) Sprague and Adkins, ibid ., .56, 2669
{1934) ; (i) Van Dazee and Adkins, ibid ., 5_1, 1)+7 (1935) ; (j) Musser an d
Adkins, ibid, 60, 664 (1938) ; (k) Mozingo and n :~ s, ibid ., 60, 669 ( 1 9 3g ) .
1
4,5-position, i .e .,
5
3
2
1
-O - C - C - C . . . .0- .
An oxygen to carbon linkage is labile, but not so markedly so as in (a) an d
(b), if it is in the 1,2 position (c) with respect to a 3,4-carbon t o
ox gDn linkag&, i .e .,
4
3
2
1
.
-0-C-C . . .0-- ,
or (d) with respect to a double bond in the 4,5-position, i .e . ,
5 4 3 2
1
C = C - C. tC . . . . 0- .
It is readily seen uC
'f rignin consisted of the units of th e
type IV joined through oxygen, as suggested by Freudenberg many years ago ,
it would give the alcohols and glycols reported above . The linkages a t
which hydrogenolysis might occur are indicated by dotted lines . However ,
carbon to carbon linkages may also be cleaved by hydrogen so that the pre CH 30
c
00H 3
"e
' -C,
HO
IV
dominance of nine carbon units in the products does not necessarily indicat e
that they were joined through oxygen in the lignin . For example, propylen e
glycol is the chief product of the hydrogenolysis of sugars . It is there fore possible that lignin consists of long carbon chains containin g
benzenoid nuclei . The products isolated as described above show ver y
definitely that units containing at least as many as eighteen carbon atom s
do occur in lignin . The presence of such units in lignin has bee n
postulated by Freudenberg and by Hibbert, 3 but as far as we know no on e
has over isolated from lignin a compound containing more than a nine carbo n
unit . Incidentally, it may be pointed out that the very high yield o f
methanol suggests that -it t id got come entirely from methoxyl groups, but
$. . f
4i .
cleaves e . Methanol is so produced in th e
lot
' '.
r .
~••~,•
, r_ . 1 ,
;or recent references to lignin see Freudenberg, Meister, and Flickinger ,
Ber ., 70,500 (1937) ; Leger and Hibbert, Jour . Amer . Chemical Society, 60 ,
565 ( 1938) ; Freudenberg, Papier-Fabr ., 36, (Tech .Tl .) 3 .1 (1938) ; C . A . ,
2909 ( 1 93$) ; Hibbert, Can . J . Research, 16, 71 (193$) .
~,
.
.•
The analysis of the type of lignin submitted to hydrogenatio n
corresponds to the formula C50H59_ O1g . Eight of the carbons and oxygen s
are present in methoxyl groups, while there are two hydroxyl groups capabl e
of methylation• therefore the formula may be written a s
One oxygen is present as a carbonyl group .'
(Ci.2H3308 (OCH3 ) 8 (OH) 2 ) x.
It is apparent that there are more oxygens in lignin than can be accounte d
for in hydroxyl and methoxyl groups and in joining nine carbon atom unit s
together . The presence of even larger carbon units makes it necessary t o
postulate the presence of cyclic ethers presumably of the furan or pyra n
type . It is possible that the oxygens are in carbonyl,(or the corresponding enols) rather than in cyclic ethers, although the ether, for various
reasons, seems the more probable linkage . The proportion of carbon t o
oxygen in lignin, disregarding the hydroxyl, methoxyl, and carbonyl oxygen ,
is six to one . This is the same proportion of carbon to oxygen found i n
the product boiling above 130 0 (1 mm .) .
The capacity of lignin to take up hydrogen is perhaps best in dicated by the fact that a given weight of lignin will take IT mor e
hydrogen than the same weight of ethylene or benzene, It lignin consiste d
of nine carbon atom units such as IV, and contained the methoxyl, hydroxyl ,
carbonyl, and ether oxygens indicated by analysis, it would not take up a s
much hydrogen in the process of conversion to the alcohols, glycols, an d
ethers, as has been found to be true . This fact should be considered i n
formulating a structure for lignin .
The proportion of the lignin converted to the nine carbo n
alcohol and glycols, as compared with the proportion converted to product s
boiling above 1300 (1 mm .), is not necessarily an indication of the pro portion in lignin of nine carbon units as compared with larger units . For
if hydrogenation of unsaturated linkages precedes hydrogenolysis of carbo n
to carbon or oxygen to carbon linkages, then cleavage (hydrogenolysis) may
not ensue, because the unsaturated groups which facilitate cleavage are n o
longer present . If hydrogenolysis takes place first, then hydrogenatio n
will ensue . Thus the proportion of the products may be a function of th e
relative rates of hydrogenation and hydrogenolysis and may not be due t o
differences in structure . The relative rates of hydrogenation and hylro genolysis are often quite sensitive to small changes in experimen .al conditions . In fact, in some hydrogenations of lignin we have obtained a
larger proportion of the higher boiling products .
•
-E . E . Harris, Jour . Amer . Chemical Society,
g 94 (1936) ; also a paper
given before the Cellulose Division at the American Chemical Societ y
meeting in Pittsburgh, 1936 .
R1176
_4_
Experimental Par t
Preparation ofLgaixl .--Freshly out ;speen (PopuLls
oides )
wood was made into coa se s•a i ust mnd while still green extra- c,on-tine=sly with cold ethyl alcohol for 6 days, and with an etiaene],=bo f i '
mix .re (1 :2) for 2 days in a Lloyd extractor . The ektracted wood 050
g .) containing 22 percent lignin was then air dried and refluxed fdt 3
days with 4+ .5 liters of methanol (99 percent) containing 3 percent hy4s e
chloride by weight, The methanol solution was removed and concentrated t o
about two volumes and then poured into 10 liters of cold mater . l'igeetigni n
(6o-65 g .) thus thrown out was filtered from the solution . The lip*t
after air drying in the dark for 1 day, was again .dissolved in 750_AAA'
methanol and precipitated by pouring into 10 liters of water, a .s'.~l 6 :ed,,
and then dissolved in 300 inl . of glacial acetic acid . The inool <e frac tions were discarded . . The glacial acetic acid s :olutiron was TOkfecl int o
20 Liters of water, and after filtering was washed until free of ateCa e
acid . It was first air dried and dried at 35 0 (5 mm .) overnight.. flee
yield was 59 to 60 g . of a very light tan colored powder . The yic~~ c1 . ,
be increased to about 90 g . if the sawdust . is subjected to rep e,aA:ed 'tee
ments .
Reaction of Hydrogen with Lignin .--The most favo--r'eble
for the reaction of hydrogen with ligni
. nexprimtdscbenhtxabove Eighty
solved and made up to a volume of 800 ml, with d
was submitted to hydrogenation in six leteltiOMA..• •
with 7 g . of copper-chromium oxide in a
6
270 ml.
Hydrogen was admitted to a pressur e
vessel then heated to 260° within 50 minutes .
about 400 atm . The pressure dropped ab'ou ! 4
during the course of 4 or 5 hours and t
'
the next 10 or 12 hours . Aite#= q b
products was ep Q .Ied until after t1
this involved a total reaction time of 18 t o
eemal
1
con s
■
l-r
Lignin obtained from wood by the soda process or the sulfuri c
acid method has also been hydrogenated with somewhat similar results .
Separation ofProducts .--The, catalyst was removed by centrifugin g
and the products of the 6 hydrogenations combined . The methanol-dioxane water fraction was then removed through a Widmer column . The material s
boiling above 100° (7 4+0 mm .) (59 g,) were then distilled without a column ,
the lower portion (35 g .) at 65-130° (1 mm,) and the higher (18 g .) at
130-270° (1 mm .) ; an electrically heated oil-bath was used for the forme r
and a Wood's metal bath for the latter . The fraction, b .p . 65-150° (1 mm .) ,
411
5- The catalyst was prepared as described on p . 13 of reference 2a and th e
reaction vessel was the one described as number 11 on p . 31 . The metho d
of agitation by rocking is shown on p . 38 . The Journal reference fo r
catalyst preparation is Connor, Folkers, and Adkins, Jour . Amer . Chemical
Society, 51F, 1138 (1932) .
81176
-5-
was carefully fractionated and refraotionF te& t} u a "modifiet1
column" using the column -and technic real ly described . 5
LL
•
The methanol-dioxane-water fraction, boiling below 1000 at 740
mm ., was dried over oven-dried anhydrous potassium carbonate far se days, then decanted and fractionated through a Widmer odium having a
spiral 30 cm . in length. The part boiling below 75° at X4O mm. wtor ti adand refractionated ; 22 g . boiled between 64 and 660 (740 mm .) ,
d.20) 0 .803, n 25 D 1 .33 g .
Chaaracteriz tiot_of Products .--A samp3. of 4..n=prap. yll .y lk
anol-1, I, was prepared by the following series of re:a6t .or eo P:itttr,Ol peo:. a
pinnate was rearranged to p-hydrox propieft-enen .e, m . p . 150°, b .p .:-1$-50
(8 mm .) by the method given in "Organic Synthesize for hyd:roxyso•etO*e'Dgmes .
.
Thirty grams of the ketone in 90 ml . of ethanol was hydrogenated om ,:p
y
nickel within 30 nutes between 100-200° . The re-salting 4-n-psap „
cyclohexanol-1,
1 .4617, b .p . 92-93° (7 mm ._), was obtained in. g1.
t
n`-'D
quantitative yield
. The a-naphthyluretharn was prepared and f'©ugd r t,e lm44t
at 136° (from petroleum ether, b .p . 90-100°) while the phenyl'et Aa • .om
the same solvent melted at 131° .
The almo,ho ., n2 5- D ],, -615, from lignin had the s-ame boiling,ra .e.
1-1, i
e- an a-naph0l1vanBth L rprp . 1k- :°D : aud a
4
III, was prepared by rea r
piophenone, g m.p . 145°, b.r)
desired compound as in the c
ethatr -.
alcohol) whil e
° . *le
i '_-L
-
. of
From
rim. -Mtby the fOllm*ing px6c•ed re : Th e
containing ].S g-e of sodium dichroma,,
water. -The
er-ature rose thro .
contiritzed 'til the lt+emperature d dop e
and ex rasetett Attt 25 z , tf 10 percen t
-Smith and .041'4 ., Ja°ur:,40er . Chemical S'bd i sty, 60, 6,7 C193 g )
LMiller and Hart t g- "Q.g Q sy . .,ee , V=xII, John Miley & Sons,,no . ,
New York City, 4193, p . 90 . .
Rosenmund, Ber ., 61, 2601 (1928) .
I+
•
111176
6-
then acidified with hydrochloric acid, extracted with ether, dried ove r
sodium sulfate and distilled . There was obtained 0.35 g . of a solid acid
having a neutral equivalent of 169 as compared to a calculated value o f
170 for C 9H140 3 .
The glycol (4.7 g .) was also oxidized by shaking at room temperatur e
for 11 hours with a mixture of chromic oxide (12 g .), acetic acid (100 g .) ,
water (50 ml .) and benzene (100 ml .) . The layers were separated, and the
water layer washed with two 20 ml . portions of benzene . The combine d
benzene layers were extracted with three 25-ml . portions of 10 percen t
sodium hydroxide . The alkaline solution was acidified with hydrochlori c
acid, saturated with sodium sulfate and extracted six times with benzene .
The benzene solution was dried over sodium sulfate . After evaporation o f
the benzene there was 1 .1 g . of solid keto acid . This was distilled at 1 mm .
and the product had a m .p . of 55-60 c and a neutral equivalent of 171 . Th e
boiling point by a micro method was apparently about 250°
(740 mm .) .
A 2,14-dinitrophenylhydrazone, (from alcohol) m .p . 50°, of the ket o
acid was prepared by the method of Allen . .
A 3,5-dinitrobenzoate, m .p . 130 1440, of the glycol, was also prepared . The amounts of these compounds avail able did not permit a rigorous purification .
Eleven grams of a fraction from lignin, b .p . 150-260° (1 mm .), was
heated under hydrogen for 4 hours at 140 00 with 1 g . of alumina (ex aluminum
isopropoxide) distributed over 5 g . of Italian pumice . The liquid produc t
was then hydrogenated over Raney nickel at 200° for 3 hours under 300 atm .
of hydrogen . Upon fractionation of the product 6 g . of material boiling
over the range from 70° (1 mm .) to 1 1400 (1 mm .) was obtained . Analysis o f
these fractions showed the compounds to be hydrocarbons .
A similar residue (15 g .) from the hydrogenation of lignin bu t
which contained 5 g . of material boiling above 260° (1 mm .) was put through
the same process . About 5 g. of hydrocarbons was obtained . The analysi s
for a portion of this product boiling 100-115° (1 mm .) is given in th e
table . The analysis for the mixture from the hydrogenation of lignin boil ing above 130° (1 mm .), which was submitted to dehydration, is also give n
in the table .
The authors are indebted to E . E . Bowden for the syntheses o f
p-hydroxypropiophenone and 3,4-dihydroxypropiophenone needed for th e
syntheses of )k-n-propylcyclohexanol-l and )+-n-proprlcyclohexanediol-1,2 ,
respectively, and to H . A.Campbell for the analyses of various compound s
and products .
•
Allen, Jour, Amer . Chemical Society, 5_2, 2957 (1930) .
R1176
-7-
SUMMARY
Lignin, extracted from aspen and dissolved in dioxane, re acted with hydrogen so that one mole was taken up by 25 g . of lignin .
Four products, i .e ., methanol, 4-n-pro ylcyclohexanol-l ,
u-n-propyleyclohexanediol-1,2, and 3-('-hydroxycyclohexyl )-propanol 1 have been separated and identified . The weight of methanol isolate d
and the combined weight of the three 9-carbon atom compounds identi fied were about 28 and 44 percent, respectively, of the weight o f
the lignin submitted to hydrogenation, The highest yields, previousl y
reported, from a hardwood lignin by any treatment were 4 percent fo r
methanol and 9 .5 percent of aromatic compounds isolated after methyla tion . It has been shown that there are units of fifteen or mor e
carbon atoms in the higher boiling products . These products appea r
to be hydroxy saturated ethers .
81176
-8-
Table l .--Analysesofvariouscompound s
B .p . or m .p . :
Formula
Name
------------------------ :
:
92-93
4-Pr-cyclohexanol-1 (I) :
ct-Naphthylurethan of
o-Phenylurethan of I
136
131
;M .p .
3-(4-Hydroxycyclohexyl)- :
125-127
propanol-1 (II )
3,5-Dinitrobenzoate of 'M .p . 130-144
II
Keto acid from II
7
M .p .
1
55-60
3,5-Dinitrophenylhydra- M.p . 90-93 •
zone of the keto aci d
4-Pr-cyclohexanediol-1,2 :
107-110 • 1
(III)
bis-a-Naphthylurethan
of III
Residue from hydrogenation of lignin
Hydrocarbons from
residue above
81176
:M .p . 2 1 g -219
Above
Analyses, percen t
Calcd
Foun d
-
.-
09H1 g0
:C, 76 .0
H, 12 .7
C20B25N02 :N, 4 .50
0i6H23NO2 :N, 5 .40
C9Hl go2
C, 6 g .3
H, 11 .4
C23H22N40 '0, 50 .6
:H, 4 .o
: N, 10 .4
C9Hi403
0, 63 .5
H, 8 .2
C15H18N406 N, 16 .0
C9H1g02
:C, 68 .3
H, 11 .4
C31H32N204 :N,
5 .65
130
1
: C,
100-115
1
;C,
H, . .
75 . 9
12 . 6
4 .6 5
5 .36
68 . 5
11 . 5
50 . 7
4.2
9 .9
63 . 2
8.3
15 . 1
68 . 3
11 . 3
5 .56
72 . 9
10 . 8
g6 . 6
12 .5