T H E SYNTHESIS OF SYRINGALDEHYDE FROM VANILLIN'

ABSTRACT
Springaldehyde has been synthesized in high yields frorn vanillin. T h e process
coiisists of the iociination of vanillin, followed by the interaction of the resultant
5-iodo\~anillinwith sodiuin nlethoxide in anhydrous methaclol a t temperatures
of 130 f 4" C. for one hour in the presence of a copper catalyst. Along with the
syringaldehyde, s ~ n a l lamounts of unchanged 5-iodovanillin and vanillin were
Can. J. Chem. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF YORK on 03/12/13

always found in the reaction r n i s t ~ ~ r eAnalysis

. of the final product was rnade
by a n initial separation of the coinponents by downward paper ~ h r o r n a t o g r a ~ h ~
~isinga mixture of petroleum ether (b.p. 100-120" C.), di-n-butyl ether, and water
( 1 0 : 1 : 1 ) a s the developing agent for a period of 10 hr. T h e separated conlpounds
were extracted from the paper and their concentrations in alcoholic alkaline
solutions determined spectrophotometrically.
Under co~iditionsby which 5-iodovanillin was converted to syringaldehyde
in better than a 95y0yield, 5-brornovanillin gave only a 61% yield and 5-chloro-
vanilliu gave no detectable anlounts of syringaldehyde.

In connection with an investigation into the fundainental nature of isolated

lignins and their degradation products from angiosperms (poplar xiiood and
cereal straws), it was important to have available many reference c o ~ n p o u n d ~
of known structure. For such work the con~poundsrequired are, for the 111ost
part, of two types: those containing the guaiacyl (4-hydroxy-3-methoxyphenyl)
nucleus and those containing the syringyl (4-hydrox)~-3,5-di1nethox)r~hen~1)
nucleus. Many derivatives having the former structure have been synthesized
from vanillin (4-hydr0xy-3-n~eth0xybenzaldehyde) which is readily available
For personal use only.

a s an oxidation product of g)rInnosperin lignin or its derivatives. Much less

work has been reported on the chemistry of compounds having the syringyl
nucleus. A logicai starting material for their syntheses would be the corre-
sponding aldehyde, syringaldehyde (4-hydrox~~-3,5-diinethoxybenzaldehyde).
Even though the oxidation of lignified angiosperin material gives rise to a
111ixtui-e of both vanillin and syringaldehyde, the separation and recovery of
each i s not readily accomplished, SO another source of this latter aldehyde is
Pi-eferred.
Many syntheses, as outlined by Pearl (9), have been reported, and a labora-
tory pl-ocedure has appeared in Org-ailic Syntheses (1.). The main starting-
material in these procedures is either 1,3-dimetl~~~lp~~rogallol or gallic acid,
the conversion of either of which to the aldehyde involves many separate steps
and therefore low over-all yields. A recent publication (10) describes a new
s!-nthesis involving the conversion of vanillin to 5-hydrox~~vanillin; complete
methylation followed by selective dernethylation of the 4-methoxy group to
qive s>-ringaldeh~-de in good !.ield. These aut1101-salso report that their attenlpts
a t tile conversion of 5-halovanillin to syringaldehyde were universally un-
sLlccessful.Some time ago an investigation was begun in these laboratories to
this same synthesis of syringaldehyde from vailillin by the inti-oduction
1 !\{a?~z~scriptreceived Decer~zber23, 1952.
Co?ltribz~tio?rfrom l l ~ eDepnrtnrent of Chenlistry, Ut~iversityof Saskatchewan, Saskatoon.,
s a s k . T h i s paper constit.14te.r part of a thesis sz~btrjilted by J . A. 11facDonald .i?l partial fzrlfillrt~e77.t
of the reqlrirenzents for the degree of Acaster of Scie?zce i n Che7tristry and was prese?~tedat the S 4 t i ~
i171112rn/.Conferencr of the Chemzcal I71stitzrte of Canada, IYir~nipeg,J z ~ n e1951.
 PEPPER A N D J l o c D O N A L D : SYRIhl(;ALDERYDE 477

of a methoxy grouping. One value of such a method would be the possibility

of thereby converting a mixture of the aldehydes obtained b>r the oxidation
of hardwoods into pure syringaldehyde should a valuable use be found for
this chemical.
In a previous colnmunication (8) it was reported that preliminary experi-
ments on the two-step synthesis: vanillin -+ 5-bromovanillin + syringalde-
hyde, appeared pron~ising.By treatment of the 5-bro~novanillinwith sodiun~
methoxide in anhydrous methanol in the presence of copper turnings a t
temperatures of 160-200" C., small yields of syringaldehyde were obtained
along with other reaction products. Because of the well known reactivity of
Can. J. Chem. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF YORK on 03/12/13

halogen substituted aromatic compounds increasing froin chloro- through

bromo- to iodo-derivatives in nucleophilic substitution reactions (3), it was
thought that the use of 5-iodovanillin in a similar synthesis would be Inore
suitable. Excellent yields of syringaldehyde (11) have now been obtained from
vanillin (I) by this method according to the following equations:

C HXO ---- CHaO--

I? NaOCH3 CH~Op
Ho-0-CHO --+ HO- / \ \
/-C H O / \ CHO
I- CH,O-- /-

Preliminary runs were carried out in which 5-iodovanillin, ailhydrous

methanol, sodium methoxide, and a copper catalyst were heated together in
For personal use only.

a rocking autoclave a t temperatures ranging from 72" C. to 180" C. and for

reaction times of from one to eight hours. After dilution of the orange to dark
red reaction mixture with water and filtration to remove the copper, the
products were extracted into chloroform and then treated as follows. First
they were separated into bicarbonate-soluble and bicarbonate-insoluble
fractions, and then each of these was subdivided into sodium bisulphite-
soluble and sodium bisulphite-insol~~ble portions. I t was often noticed that
complete acidification of the original alkaline reaction mixture caused any
unchanged 5-iodovanillin to precipitate. This insolubility of 5-iodovanillin in
acidic aqueous methanol permitted its ready separation from the other conl-
ponents of the reaction mixture. Support was found in these experin~entsfor
an observation reported earlier (11) t h a t , frorn "neutral" lignin fractions,
compounds having a syringyl nucleus were isolated. In this case the inconlplete
liberation of phenolic compounds b y the exhaustive carbonation of a strongly
alkaline solution was borne out by finding phenolic aldehydic products in both
the bicarbonate-soluble and bicarbonate-insoluble portions.
T h e aqueous residues from the various runs were analyzed for iodide ion
by a standard method (12). These results along with the anrount of unchanged
5-iodovanillin recovered are given in Table I , where iodide ion is expressed as
the corresponding weight of 5-iodovanillin. The values reported for the iodide
ion determinations are averages of duplicate ailalyses which agreed within
one per cent and are mi~limumvalues since, prior to the analyses, the aqueous
solutions had been acidified and extracted several tinles with chloroforn~.I t
 TABLE I
EXPERIMENTS USING 5-IODOVANILLIN~
PRELIMINARY
--
I I I I I
Syring- Percentage of original 5-iodovanillin
aldehyde

I
Average Reaction Nonaldehydic containing
R u n temp., time, fractionb, fractionc, Recovered
No. " C. hr. gm. gm. ur~chai~ged
Nil
N,il
82.4
Nil
36.9
Can. J. Chem. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF YORK on 03/12/13

19.2
25.4
Nil
Nil
13.0

a
For each r u n , 5-iodovanillin (13.0 gm.), anlrydrorrs ~?zethanol(250 wrl.), sodium (10.0 gm.)

'
(with the exception of R Z I1~i n which 5.0 gm. toere used), and copper catalyst A (12.8 gm.) were
re cted
Bicarbonate-insol~~ble, sodium biszrlphite-insolzrble.
Combined bicarbonate-insoluble, sodizrnt bistrlphite-solzrble, and bicarbonate-soltrble fractions.
R u n carried out at atn2osplteric pressure under rtfEux.

seems reasonable to conclude that, under the conditions employed, all or

nearly all of the 5-iodovanillin which reacts loses iodine.
Varj-ing small amounts of syringaldehyde were isolated by crystallization
and sublimation from many of the fractions obtained in these preliminary
For personal use only.

experiments, but the amounts given in Table I represent only the weight of
the crude products shown to contain syringaldehyde. I t appears that, with
such a reaction mixture, increasing temperatures lead t o an increase in the
syringaldehyde-containing fractions until a temperature of 135" C. is reached
and that above this temperature the yield of the nonaldehydic fraction be-
comes more and more appreciable. The chemical nature of this derivative,
referred to as the "Unknown" in Fig. 1, has not as yet been 'determined but
cjualitative tests show that it is phenolic and contains no iodine.
In an analysis of the reaction product of Run 10 by a chromatographic
method, an appreciable quantity of vanillin (28y0, based on 5-iodovanillin)
along with syringaldehyde was detected. The presence of vanillin, characterized
by the comparison of its ultraviolet absorption spectrum and chroinatographic
behavior with an authentic sample, is not surprising. The reductive cleavage
of aryl halides is not uncommon under similar alkaline conditions (3, 10).
The lack of a single fraction containing all of the syringaldehyde, and the
detection of vanillin in the reaction product, indicated that a new method of
determining yields was required. A method of analysis similar to that used by
Stone and Blundell (13) proved satisfactory. The use of a 10: 1: 1 mixture of
petroleum ether (b.p. 100-120" C.), di-n-butyl ether, and water as the de-
veloping solvent gave an adequate separation of syringaldehyde, vanillin,
5-iodovanillin, and the unidentified reaction product on a paper strip (Fig. I ) .
The positions of the various phenolic con~poundswere determined by
 PEPPER A N D MacDONALD: S Y R I N G A L D E H Y D E
Can. J. Chem. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF YORK on 03/12/13
For personal use only.

FIG.1. Chro~natographicseparation of syringaldehyde, vanillin, 5-iodot~anillin,and a n

unidentified phenolic reaction product. Chro~natogramdeveloped for 10 hr. using petroleulli
ether (b.p. 100-120" C.), butyl ether, and water (10.1.1).
 spraying test strips of the chromatogram lvith a ferric chloride - potassium
ferricyanide mixture (2). The separated bands were then extracted with
ethanol and the concentrations of the extracts determined spectrophoto-
metrically. Such analyses of linolvn mixtures showed that an average of 7.6Yo
of the syringaldehyde, 18.0y0 of the vanillin, and a negligible amount of
5-iodovanillin were lost during the analysis. The results of subsequent analj.ses
were corrected for these losses.
A second series of experiments was then carried out and the products analyzed
by this preferred method. The results are given in Table 11. The results of
Runs 11 and 12 indicate that a copper catalyst is necessary in this reaction.
Can. J. Chem. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF YORK on 03/12/13

I t was thought that the extreme variations in yields of products of Runs 11,
13, 14, and 15 were not due primarily to the slight temperature differences
recorded but rather to differences in the catalyst, a fresh batch of which had

TABLE I1

1 1 Analysis of reaction mixture , .

No.
R un I Average
1
Copper temp., CII componentd Duplicate malgses, g m
catalystb 1 mean
yield, % 1 Total
yield, %
For personal use only.

a For each run 5-iodovnnillin (13.0 gm.), sodizrm (10 giiz.), anhydrozls methawol (250 i121.),
and copper catalyst (16.8 gm.*) were zrsed; reactiotz time one hour at elevated tefnperatzlre.
Catalyst A prepared by precipitating copper froin copper szllphate solution with si?tc;catalyst B
was British Drug Hozrses' precipitated copper powder, reagcvzt grade.
J/la.zimum temperature variation, f4" C.
Abbreviations: S A for syringaldehyde, V for vanillin, afld 5 I V for 5-iodovanilJi~z.
* Some later work has indicated that under similar cofzditions this amou~rtof copper nray be
reduced to 2.0 grn. witkotrt seriozrs decrease i n percentage conversio~zsand also that yields of nrozrnd
5 0 % of syringaldekyde m a y be obtained if 0.5 gm. iodine are used i n place of the metallic copper
catalyst.
 PEPPER AND MacDOh'ALD: SYRINGALDEHYDB 481

been prepared for each run. Using catalyst portions taken from the same
source (Catalyst B), reproducible results were obtained (Runs 16 and 17).
Furthermore, the high yields of syringaldehyde obtained indicated that the
British Drug Houses precipitated copper powder was the preferred catalyst
for this reaction. Subsequently, many runs similar to Runs 16 have been made
and pure syringaldehyde readily recovered and recrystallized as light yellow
needles from Skellysolve "C", m.p. 109-110" C. in yields of 85y0 or better.
The possibility of converting either of the cheaper 5-chloro- or 5-bromo-
vanillin to syringaldehyde under similar conditions was investigated. Both
these 5-halovanillins moved further than vanillin or syringaldehyde under
Can. J. Chem. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF YORK on 03/12/13

the previously used chromatographic conditions and hence did not interfere
with the analyses. Using 5-bromovanillin, runs carried out a t 130" C. and
140" C. gave rise in each case to products representing a 61% conversion to
syringaldehyde. Under similar conditions using 5-chlorovanillin a t any of
130" C., 140" C., or 175" C., no syringaldehyde was detected. This decreased
conversion of the 5-halovanillins to syringaldehyde from 5-iodo- through
5-bromo- to 5-chloro- is in agreement with the known reactivities of halogen
substituted aromatic compounds.
This niethod for the ,preparation of large amounts of syringa1deh~-dehas
since been used very successfully in our laboratories.
EXPERIMENTAL*
Reagents.-The 5-halovanillins were prepared according to the following
previously reported procedures: 5-iodovanillin (n1.p. 178.5-179.5' C.) by the
For personal use only.

niethod of Erdtman (6)**; 5-bromovanillin (m.p. 164" C.) by McIvor and

Pepper (8), and 5-chlorovanillin (m.p. 160.5-162" C.) according to Hopliins
and Chisholm (7).
Copper Catalyst A.-The early experiments were carried out using a copper
preparation prepared by the addition of zinc to copper sulphate solution
according to the method of Brewster and Groening (4). T h e catalyst was pre-
pared immediately prior to each run and washed with 10 portions of absolute
methanol before use.
Copper Cutulyst B.-British Drug Houses' precipitated copper powder,'
Reagent Grade, was used in the later experiments.
Appurutzis.-All the reactions were carried out in a stainless steel liner,
capacity 1080 ml., of an Alninco high pressure hydrogenator, Model No. 406-01
DA. This apparatus served as a rocking autoclave. A Brown Indicating
Controller (0-600" C.) coupled with an Aminco variable voltage transformer
inserted into the automatic heating circuit permitted temperature control to
within f 4 " C. A Beckmann Model DU spectroplzotometer was used to obtain
the ultraviolet absorption data.
Procedz~reEmployed i n Runs 11-17
These runs (Table 11) were maintained a t the stated reaction temperatures
for one hour, after \vhich the bomb was cooled slowly. In each case the product
* A l l melting points are corrected.
** I t appears that the 0.2 N sodium hydroxide reported should read 3.0 N sodizcitz I~ydroside.
 482 CANADIAN JOUKXAL OF CI-IEMSTRY. VOL. 31

was diluted with water (500 ml.) and filtered to remove the copper. The
filtrate, which varied in color from yellow to orange red, was analyzed by the
chromatographic-spectrophotometric method.
T h e chromatography was carried out according to the method of Stone and
Blundell (13). Strips of Whatman No. 1 filter paper, 22 in. X 6 in., were
spotted along a base line with the alkaline reaction mixture. The volunles of
the drops were measured by a 1 rill. capacity Emil Greiner ultramicroburette
which was graduated in 0.001 mI. divisions. After acidification of the material
on the paper using acetic acid vapor, the development was conducted in a
descending manner in an apparatus similar to that described by Consden et
Can. J. Chem. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF YORK on 03/12/13

al. (5). Water was placed in the bottom of a tall glass cylinder and the organic
phase from a 10: 1: 1 mixture of petroleuln ether (b.p. 100-120' C.), di-?z-butyl
ether, and water was placed in the small trough near the top and also in a
small beaker placed on the bottonl of the tightly closed vessel which was
allowed to stand for one-half hour prior to putting the paper in place. After
development for 10 hr., the paper strips were removed, and the test strip cut
off and sprayed with a 1% solution of ferric chloride followed by a 1% solution
of potassium ferricyanide, which indicated, by pronounced blue spots, the
position of the phenolic con~pounds(2). The main chromatogran1 was then
cut into horizontal strips, each of which contained only one compound which
was separately extracted with ethanol in Soxhlet extractors for two hours.
Each extract was transferred to a 50 ml. volumetric flask containing 4 1111. 01
0.2 N potassium hydroxide in ethanol, and made up to 50 ml. with ethanol.
For personal use only.

FIG.2. Concelltratioll vs. optical density curves in alkali-ethanol for (1) vanillill, (2)-syring-
aldehyde, and (3) 5-iodovanillin.
 PEPPER A N D A-facDOiVALD: S Y K I N G A L D E H Y D E -Isa

The concentration of the extract was then deterinined by the measurcinent of

its optical density and by inakiirg use of previously determined optical densit?.
versus concentration curves (Fig. 2). The wave lengths ~lsed were those
determined by an exami~rationof the ~ ~ l t r a v i o l eabsorptioil
t spectra of the
compounds in alcoholic potassium hydroxide and were 368, 352, and 356 1 1 1 ~
for syringaldehyde, vanillin, and 5-iodovanillin, respectively.
Syringa1delzyde.-Run 16 serves to illustrate a typical preparation of syring-
aldehyde by this method. Sodium (10.0 gin.) was dissolved in anhydrous
inethanol (250 1111.). This solution together with 5-iodovanillin (13.0 gm.) and
British Drug Houses' precipitated copper powder catalyst (12.8 gm.) was
Can. J. Chem. Downloaded from www.nrcresearchpress.com by UNIVERSITY OF YORK on 03/12/13

heated in the rocking autoclave a t 124-128' C. for one hour. After cooling,
the reaction product was diluted with water (500 ml.), filtered to remove the
copper, and acidified with hydrochloric acid. A small amount of unchanged
5-iodovanillin precipitated and was removed by filtration. The filtrate was
extracted with chloroforin (5 X 200 1111.) and the chloroform extract back
extracted first with a few milliliters of sodium thiosulphate, to remove any free
iodine, and then with water (2 X 10 ml.). After drying over anhydrous sodium
sulphate the chloroform solution was concentrated to a small volume by
distillation. The residue was crystallized from Sltellysolve "C", yielding light
yellow needles (7.28 gin.) (85.5y0), 1m.p. 109-110' C. A inixed melting point
with authentic syringaldehyde prepared according to Pearl (9) showed 110
depression. A similarly obtained product, recrystallized several times from
Skellysolve "C" and twice from water, melted a t 109.8-110.2' C. Calc. for
C9HI004:C , 59.32; 11, 5.53y0. Found*: C, 59.33, 59.20; H , 5.59, 5.55%.
For personal use only.

ACKNOWLEDGMENTS
The authors wish to thank the Saskatchewan Research Council for the
award of two Graduate Research Fellowships t o one of them (J.A.M.) and
for other summer financial assistance. Grateful ackirowledgment is made t o the
Howard Smith Cheinicals Limited for their kind donation of the vanillin used
in this research.
REFERENCES
1. ALLEN,C. F. H. and LEUBNER, G. W. Organic syntheses, Vol. 31. John Wiley & Sons,
Inc., New York. 1951. p. 92.
2. BARTON, G., EVANS,R. S., and GARDNER, J . A. F. Nature, 170: 4319. 1952.
3. BUNNETT, J . F. and ZAHLER,R. E. Chem. Revs. 49: 273. 1951.
4. BREWSTER, R. Q. and GROENING, T. Organic syntheses. Coll. Vol. 2. John Wiley
& Sons, Inc., New York. 1944. p. 446.
5. CONSDEN, R., GORDEN, A. H., and MARTIN,A. J . P. Biochem. J. 38: 224. 1944.
6. ERDTMAN, H. Svensk. Kern. Tid. 47: 223. 1935.
7. HOPKINS,C. Y. and CHISHOLM, M. J. Can. J . Research, B, 24: 208. 1946.
8. RiIcIvo~,R. A. and PEPPER,J. M. Can. J . Chem. 31: 298. 1953.
9. PEARL,I. A. J. Am. Chem. Soc. 70: 1746. 1948.
10. PEARL,I. A. and BEYER,D. L. J. Am. Che~n.Soc. 74: 4262. 1952.
11. PEPPER,J. M., BROUNSTEIN, C. J., and SHEARER, D. A. J. Aln. Ctlern. SOC.73: 3316-
1951.
12. SCOTT,W. W. Standard methods of chemical analysis. Vol. I. 5 t h ed. D. Van
Nostrand Company, Inc., New York. 1939. p. 456.
13. STONE,J. E. and BLUNDELL, M. J . Anal. Chern. 23: 771. 1951.

* Aftalyscs by Dr. R . U. Lerniez~x,Natiortal Research Council Laboratory, Saskatoorz, Snsk.