176 J. M . de Bruijn et al.

/Alkaline degradation of monosaccharides 111

Recl. Trav. Chim. Pays-Bas 105, 176-183 (1986) 0165-0513/86/06176-07$2.25

Alkaline degradation of monosaccharides 111". Influence of reaction parameters

upon the final product composition

J. M. de Bruijn#, A. P. C. Kieboom and H. van Bekkum

Laboratory of Organic Chemistry, Delft University of Technology, Julianalaan 136, 2628 BL Delft,
The Netherlands
(Received December 19th, 1985)

Abstract. A systematic investigation of the alkaline degradation of monosaccharides is presented.

The influence of several parameters upon the degradation reaction has been determined by analysis
of the reaction products using HPLC and, in some cases, GC. The HO- and the monosaccharide
concentrations markedly influence the final product composition with respect to both the amount of
C, to C, acids and the formation of oligomeric acidic products, the so-called > c,
acids. Maximum
yields of these > C, acids, up to 50 mol-C,-%, are obtained at an HO- concentration of between
M and M when the monosaccharide concentration exceeds lo-' M. The presence of
divalent calcium increases the retro-aldolization of monosaccharides in alkaline medium, as illustrated
by the enhanced production of lactic acid, by complexation with, for instance, D-fructose. Borate partly
protects monosaccharides against alkaline degradation through their borate esters, whilst the amount
of saccharinic acids in the degradation product mixture is doubled. On the other hand, neither chloride
and carbonate nor the reaction temperature influences the final product composition. Alkaline
degradation experiments with pyruvaldehyde, glyceraldehyde and 1,3-dihydroxyacetone, all assumed
to be reaction intermediates, indicate that aldolization of (di)carbonyl compounds causes the formation
of substantial amounts of > C, acids.

Introduction

Reactions of monosaccharides in aqueous alkaline solution
have been studied extensively for almost a century since the
publication of the classical paper' of Lobry de Bruyn and
Alberda van Ekenstein concerning the interconversion of
D-glucose, D-fructose and D-mannose in alkaline medium.
These reactions can be divided into two distinct groups of
transformations :
fi) Initial transformations, in which the sugar moiety
remains intact and undergoes reversible rearrangements
such as ionization, mutarotation2-', enolization and
isomerization6-".
(ii) Degradation reactions6*"-'6, some of which are ir-
reversible and lead ultimately to organic acid products.
The degradation of monosaccharides (Fig. 1) starts with the
elimination of a P-OH group from an enediol anion species,
which is also an intermediate in the isomerization reaction.
The resulting a-dicarbonyl compound can undergo either a
benzilic acid type of rearrangement or a cleavage ofthe C-C
bond between the carbonyl groups (a-dicarbonyl cleavage).
The products formed upon benzilic acid rearrangements are
a-hydroxycarboxylic acids, e.g. lactic acid and saccharinic
acids. Dicarbonyl cleavage gives rise to both a carboxylic
acid, e.g. formic acid, acetic acid, glycolic acid, and an
aldehyde, e.g. formaldehyde, acetaldehyde.
Aldolization and retro-aldolization of carbonyl compounds
present in the degradation mixture also play an important
role in the alkaline degradation pattern of monosaccharides.
These reactions result in elongation and fragmentation of the
received little attention in terms of mechanistic con-
siderations of the alkaline decomposition of carbohydrates,
despite the fact that it may be responsible for the formation
I C . A . Lobry de Bruyn and W . Alberda van Ekenstein, Recl. Trav.
Chim. Pays-Bas 14, 203 (1895).
J. M. Los and L. B. Simpson, Recl. Trav. Chim. Pays-Bas 73,941
( 1954).
J. M . Los and L . B. Simpson, Recl. Trav. Chim. Pays-Bas 76,267
(1957).
W . Pigman and H. S. Isbell, Adv. Carbohydr. Chem. 23, 11
(1968); ibid. 24, 13 (1969).
G . de Wit, A . P. G . Kieboom and H . van Bekkum, Tetrahedron
Lett. 5, 3943 (1975) and Recl. Trav. Chim. Pays-Bas 98, 355
(1979).
J. C. Speck, Jr., Adv. Carbohydr. Chem. 13, 63 (1958).
' H . S. Isbell, H . L. Frush, C. W.R. Wade and C. E. Hunter,
Carbohydr. Res. 9, 163 (1969).
H. S. Isbell, K . Linek and K . E. Hepner, Jr., Carbohydr. Res. 7,
86 (1968).
T. I. Khomenko and 0. V. Krylov, Kinet. Catal. 15, 555 (1974).
l o G. de Wit,A . P. G. Kieboom and H. van Bekkum, Carbohydr. Res.
74, 157 (1979).
I ' J. A . Nef, Ann. 357, 214 (1907); ibid. 376, 1 (1910).
l 2 J. C. Sowden, Adv. Carbohydr. Chem. 12, 35 (1957).
l 3 E. F. L.J. Anet, Adv. Carbohydr. Chem. 19, 81 (1964).

carbon chain, respectively, until one of the above-mentioned I4 W . Pigman and E. F. L. J. Anet in "The Carbohydrates IA", W.

termination reactions occurs. The aldolization reaction has
* References 30 and 18 are considered to be parts I and 11,
Pigman and D . Horton, eds., Acad. Press, 2nd edition, New York,
1972, p. 165.
l 5 G. de Wit, "Gedrag van glucose, fructose en verwante suikers in
alkalisch milieu", PbD. Thesis, Delft University of Technology,
Delft, 1979.
l6 J. M. Macleod, "Anaerobic alkaline degradation of D-glucose,

respectively. cellobiose, and derivatives", Ph.D. Thesis, Lawrence University,

# Present address: Central Laboratory, CSM Suiker B.V., Val- Appleton, 1975; J. M. Macleod and L . R. Schroeder, J. Wood
veeken 6,4815 HL Breda, The Netherlands. Chem. Techn. 2, 187 (1982).
Recueil des Travaux Chimisues des Pays-Bas, 10516,june 1986
H-C-OQ
sugar anion
II
C-OH
I
P -elimination
HO-C-H
I 1
enediol anion
rearrangement
Fig. 1 . Part of the reaction network of monosaccharides in alkaline medium.
of substantial amounts (up to 50%) of oligomeric acidic
p r o d u ~ t s ~ the
~ ~ ”so-called
, > C, acids.
The present-day mechanistic insights into the reactions of
monosaccharides in aqueous alkaline solution have been
recently reviewedI8. It has been shown that the mechanism
of the initial transformations of monosaccharides in alkaline
medium are well understood. The course of the subsequent
complex alkaline degradation, on the other hand, has only
been partly elucidated. Despite all the efforts in this area, a
major challenge remained, namely the complete quantitative
analysis of the reaction mixtures. D u e to the lack of a
convenient quantitative analysis technique, a systematic
investigation was not possible and oligomeric product for-
mation has not been taken into consideration in mass balance
calculations. Although some investigator^^^^^^^^^-^^ were
aware of the presence of such aldolization products, the
amount, the molecular structure and the mode of formation
could not be determined. A recently developed routine
HPLC method3’ presented us with the opportunity to investi-
gate the influence of several reaction parameters upon the
degradation pattern by quantitative analysis of the final
product mixtures. This paper describes the results of this
research which may be considered as a convenient basis for
further elucidation of the alkaline degradation mechanism of
monosaccharides.
Experimental
Materiab
Pyruvaldehyde was obtained from Janssen Chimica as a 40 wt. %
aqueous solution which was freshly distilled prior to use. D-Psicose
was prepared” according to the literature. All other chemicals were
obtained from Merck (analytical grade).
Apparatus
The alkaline degradation experiments were carried out in a 150-ml
vessel equipped with a thermostat jacket, a pH electrode connected
to a pH meter, a dosimat filled with 2 M KOH, a magnetic stirrer,
a sampling tube, gas-inlet and -outlet tubes and a revolving tubular
device for adding reactants. The gas inlet was connected to a gas
burette containing nitrogen. Experiments were performed at atmos-
pheric nitrogen pressure and constant temperature. The pH was
kept constant with an impulsomat.
Procedure
Since oxygen may induce undesired oxidation reactions, the
reaction vessel, containing an aqueous solution ofthe desired HO-
concentration, was evacuated and N, was admitted. This procedure
177
H -C=O
- I
c=o
I
a- dicarbonyl compound
benzilic acid / \dicarbonyl
/ kleavage
/Oe
c=o H-C=O
p”
I
H-C-OH H-C=O
I I
CH2 CH2
I I
was repeated three times. The pH electrode and the tube connected
to the dosimat were then placed in the reaction vessel. The alkaline
solution was kept under a N, atmosphere and brought to the desired
temperature. The degradation reaction was initiated by addition of
a concentrated monosaccharide solution in water (1-2 ml). The
HO- concentration was kept constant by addition of base from the
dosimat to the reaction mixture.
At the end of the alkaline degradation, which was indicated by
termination of the addition of base, the reaction mixture was cooled
to 0-4°C and neutralized with a weak cation exchange resin
(BioRex 70 H). Reaction mixtures with [ H O - ] > 0.1 M were partly
neutralized (to pH 11-12) using conc. HCI, followed by
neutralization with the exchange resin.
l7 J. M . de Bruijn, A . P. G. Kieboom, H . van Bekkum and P. W. van
der Poel, 12th International Carbohydrate Symposium, Utrecht,
1984, Abstracts of Papers, p. 189; 3rd European Symposium on
Carbohydrates, Grenoble, 1985, Abstract of Papers, p. 318.
J. M. de Bruijn, A . P. G . Kieboom, H . van Bekkum and P . W . van
der Poel, Sugar Techn. Rev. 13, 21 (1986).
L . Lowendahl and 0. Samuelson, Acta Chem. Scand. B30, 691
(1976).
2o L.-A. Lindstrom and 0 . Samuelson, Acta Chem. Scand. B31,479
(1977).
M . Fleming, K . J . Parker and J . C. Williams, Sugar J. 33, 21
(1971); Proc. 13th Congr. I.S.S.C.T. 1781 (1968).
22 N . A . Ramaiah and M . B. Kumar, Proc. 13th Congr. I.S.S.C.T.
1768 (1968).
23 E. Brandes, “Untersuchungen zur Kenntnis von Braunungs-
produkten in technischen Zuckersaften”, Ph.D. Thesis, Tech-
nischen Universitat Carola-Wilhelmina, Braunschweig, 1982.
24a V. Prey and H. Andres, Z. Zuckerind. 23, 373 (1973);
bV.Prey and H . Holle, Z. Zuckerind. 24, 113 (1974);
V. Prey and H. Wesner, Z. Zuckerind. 25, 341 (1975);
dV.Prey and H . Andres, Z. Zuckerind. 26, 245 (1976);
V. Prey, H . Andres and H. Gusbeth, Z. Zuckerind. 26, 453, 592,
660 (1976); ibid. 27, 84 (1977);
H . Andres, B. Pichler and V. Prey, Zuckerind. 103, 753 (1978);
8H.Andres, I . Arisan and V. Prey, Zuckerind. 104, 278 (1979).
25 K . J . Parker, Sucr. Belge 89, 168 (1970).
*‘ K . Vukov, Zucker 18, 168 (1965).
27 I . Ziderman, J. Bel-Ayche, A . Basch and M . Lewin, Carbohydr.
Res. 43,255 (1975).
28 N . Martinez and S. G. Kara-Murza, Revista Cenic 7 , 141 (1976).
29 S. G . Kara-Murza, M . D. Zarajova and M . de Los Angeles, Int.
Sugar J. 72, 90 (1970).
30 J. M . de Bruijn, A . P. G . Kieboom, H . van Bekkum and P. W . van
der Poel, Int. Sugar J. 86, 195 (1984).
3’ M. Makkee, A . P. G . Kieboom and H . van Bekkum, Int. Sugar J.
87, 5 5 , 63 (1985).
178 J . M. de Bruijn et al. /Alkaline degradation of monosaccharides III

Analysis method 100

HPLC analyses of the carboxylic acids in the neutralized final
reaction mixtures were performed as described e 1 s e ~ h e r -(column
e~~
HPX 87 H from Bio-Rad, 300 mm length, 7.8 mm i.d., column
temperature 60°C, aqueous 0.005 M H,SO, or 0.01 M trifluoro-
acetic acid as the eluent, flow 0.6 ml/min, RI-detection). G C ana-
lyses" were carried out after freeze-drying and trimethylsilylation of
the samples (capillary CP Sil 5 column, 25 m length, 0.23 mm i.d., 00
temperature programme: 5 min at 75"C, increase at a rate of
8"C/min to 280°C, 5 min at 280"C, flame ionization detector).
The amounts of acidic products have been expressed as percentages
mol-C, produced from the starting C,-monosaccharide according to
the following equation:

n 60
mol C,, = - * mol-C,
6

i
Qo
in which C, represents an organic acid containing n carbon atoms.
On the basis of this definition, the amount of a degradation product
is expressed as the fraction of the C, monosaccharide which is
converted into that particular product. For example, if 100 mol 40
D-fructose produces, by alkaline degradation, 6 mol formic acid, I
the yield of formic acid is 1 mol-C,-% ;6 mol acetic acid corresponds
with 2 mol-C,-%, etc.
The total amount of > C, acids has been defined as the deficit on
the mass balance after summation of the quantitatively analyzed
GC, acids. The >C, acids did not produce separated peaks in
HPLC but resulted in a raising of the base-line at the first part of 20
the chromatogram due to size exclusion of these oligomeric products
by the stationary phase (sulfonated styrene-divinylbenzene copo-
lymer, H form). There was no interference of the > C, acids with
+
"P
the other acidic reactions products.

Results and discussion 0

-5 -4 -3 -2 -1 0
Preliminary experiments log [HO-]

D-Glucose, D-fructose, D-mannose and D-psicose were
treated with alkali under comparable conditions in order to
verify that these monosaccharides are degraded via common
anions as intermediates. The almost identical final product
compositions (Table I) prove that alkaline degradation of
interconvertible monosaccharides indeed proceeds via the
same 1,2- and 2,3-enediol anion species.
.,
Fig. 2. Formation of carboxylic acids as a function of the
H O - concentration (0.025 M D-glucose, D-fructose or D-man-
nose, K O H , H,O, 78"C, N , , 100% conversion): grey region,
> C, acids; @, 6 C, acids: a, saccharinic acids; A , glycolic
acid; 0 , lactic acid; V , formic acid; 0 , acetic acid;
2.4-dihydroxybutyric acid.

For [HO-] > M, these results are in agreement with

Effect of hydroxyl ion concentration
In order to determine the influence of the HO- concen-
tration upon the final degradation product composition,
monosaccharides were reacted over a wide HO- concen-
tration range, i.e. from lo-' M to 2.5 M. The results given in
Fig. 2 show that in particular the amounts of lactic acid and
the >C, acids are strongly dependent upon the HO- con-
centration applied.
the earlier observations of De Wit". Although the selectivity
towards lactic acid at increasing HO - concentration also
confirms the data of earlier o b ~ e r v a t i o n s ~a~good
. ~ ~ , com-
parison is difficult because of the use of buffer systems
32 K. Vukov,Zucker 18, 168 (1965).
33 M.L. A. Verhaart and N . H. M. de Visser, 10th Session CITS,
London 1 (1957).

Table I .final acidic product composition of alkaline degraded monosaccharidesa.

D-psicose D-fructose D-glucose D-mannose

mol-C,-%)

saccharinic acids 20.3 22.5 21.5 23.0

glycolic acid 4.7 4.5 5.0 5.0
lactic acid 6.4 6.8 6.9 6.6
formic acid 3.0 3.6 4.0 3.4
acetic acid 9.8 10.1 10.5 9.5
2,4-dihydroxybutyric acid 8.6 8.0 8.5 7.8

total b C , acids 52.8 55.5 56.4 55.3

total > C , acids 41.2 44.5 43.6 44.7

Reaction conditions: 0.025 M monosaccharide, 0.01 M KOH, H,O, 78"C, N2,7 h, 100% conversion.
Recueil des Travaux Chimiques des Pays-Bas, 10.516. june 1986
containing different types of cations which might affect the
course of the degradation reaction (see below).
In order to determine the saccharinic acid composition, we
carried out GC analyses which further confirmed the data
obtained by HPLC. Apart from trace amounts of saccharinic
and isosaccharinic acid, metasaccharinic acid is the major
saccharinic acid formed in the presence of the monovalent
bases (KOH or NaOH) studied. In addition, minor amounts
-
of 3-deoxypentonic acid ( 2 mo&-%), 2-methylglyceric
and glyceric acid ((0.5 mol-C,-%) were present according
to GC.
The remarkably high production of >C, acids at HO-
concentrations between M and lo-' M indicates that
the formation of these (higher molecular weight) products
involves aldolization reactions between neutral and ionized
(di)carbonyl compounds. As a first approximation, the > C,
acid curve reflects the overall pH dependence of the kinetic
expression:
rate = constant * [anionic species] * [neutral species].
Efect of monosaccharide concentration
De Wit15 has investigated the influence of the D-glucose
concentration upon the degradation pattern in aqueous
0.39 M NaOH at 50°C. No change in the product composi-
tion was observed upon increasing the D-glucose concen-
tration from 0.015 M to 0.15 M. As shown above, aldoli-
zation reactions will be of minor importance at this high
HO - concentration. For this reason we have reinvestigated
the influence of the monosaccharide concentration at an
HO- concentration of lo-, M where maximum > C, acid
100
80
60
I
10-1 10'2 10-3
Fig. 3. Influence of D-glucose concentration upon carboxylic
acid formation (0.01 M KOH, H20,7 8 ° C . N2,7 h, 100%
conversion): grey region, > c 6 acids; 8 , < c 6 acids: e,
saccharinic acids; A , glycolic acid: 0 ,lactic acid; 8,formic
acid; ,acetic acid; m,
2,4-dihydroxybutyric acid.
179
production occurs. Variation of the monosaccharide concen-
tration between 2 * M and l o - ' M now shows a strong
dependency of the >C, acid content in the final product
upon the starting monosaccharide concentration (Fig. 3).
Decreasing the monosaccharide concentration below
lo-, M causes a sharp decrease in the formation of > C ,
acids. Concomitantly, the content of saccharinic acids and
lactic acid markedly increases. Since cr-dicarbonyl
compounds are well-established precursors for saccharinic
and lactic acid formation, the shift of these C, and C3 acids
towards > (26 acids upon increasing D-glucose concentra-
tions indicates the importance of u-dicarbonyls in the
oligomerization reaction.
The effect of the sugar concentration upon the product
formation has been verified at two other HO- concentra-
tions, 9 * M and 4.4 * lo-, M, respectively, which are
also expected to create a situation for maximum > C, acid
production. As illustrated in Fig. 4, the dependence of the
amount of >C, acids produced upon the monosaccharide
concentration is comparable for the three HO- concentra-
tions investigated.
.i I
10-1 10-
[g~ucose], (MI
Fig. 4.' Amount of > C, acids produced versus the starting
monosaccharide concentration. Reaction conditions: 0,
4.4 * l o p 2M K O H ; x , M KOH; @,9 * M KOH;
H,O, 78 C, N,, 7 h, 100 % conversion.
Effect of calcium(II)
As shown previ~usly'~,the nature of the base used, in
particular the valence of the cation, may markedly influence
the final product composition of the alkaline degradation of
monosaccharides. This effect is of direct importance for the
juice purification stage in sugar manufacture which involves
the alkaline degradation of invert sugar by calcium hy-
droxide.
Under standard reaction conditions (0.025 M mono-
'
saccharide, 0.01 M KOH, H20, 78"C, N2), the influence of
the Ca(I1) concentration has been studied (up to 0.04 M) by
the addition of CaCI,, while the HO- concentration was
kept constant during the reaction by the addition of 2 M
KOH. The composition of the final product (HPLC analysis)
as a function of the Ca(I1) concentration is shown in Fig. 5 .
Increasing the molar Ca(II)/monosaccharide ratio increases
lactic acid formation at the cost of glycolic acid, acetic acid,
saccharinic acids and the > c6 acids. The major effect occurs
180 J . M . de Bruijn et al. /Alkaline degradation of monosaccharides III

up to a 1/1 molar Ca(II)/monosaccharide ratio. Of the
various possible Ca(I1)-monosaccharide complexes, both
the P-D-fructofuranose-Ca(I1) and P-D-fructopyranose-
Ca(I1) complexes will occur predominantly because of their
tridentate mode of coordination. As shown in Fig. 6, these
complexes may explain the promoted C3 - C4 bond fission by
retro-aldolization.
. obtain more detailed information about the effect of Ca(I1)
upon the composition of the reaction products, both HPLC
and GC techniques were used (Table 11). The Ca(I1)-induced
formation of 2-methylglyceric, 2-methyltetronic, 3-deoxy-2-
-hydroxymethyltetronic, saccharinic and isosaccharinic acid
originates from benzilic acid rearrangements of 2,3-di-
carbonyl compounds. Apparently, 2,3-dicarbonyls are
formed in higher amounts in the presence of Ca(II), whilst the
dicarbonyl cleavage reaction becomes less important as
indicated by the decrease in glycolic and acetic acid.

60
1 Table 11 Eflect of cation upon the final product composition of the
alkaline degradation of Dlfnrctosea.

Carboxylic acid
Ca(I1)
(mol-C,-%)
40 formic acid 4.0 3.8
acetic acid 9.5 3.9
glycolic acid 5.0 2.4
glyceric acid 0.5 0.9
lactic acid 12.5 27.3
2,4-dihydroxybutyricacid 8.0 7.1
2-methylglyceric acid 0.1 2.1
20' 3-deoxypentonic acid 2.1 2.9
2methyltetronic acid <0.1 1.2
3-deoxy-2-(hydroxymethyl)tetronic acid <0.1 0.5
I metasaccharinic acid 22.5 9.1
1 \ 8 , r r r I
saccharinic acid <0.1 3.8
I isosaccharinic acid <0.1 2.5
0 total QC, acids 64.8 68.1
0 0-02 0.04 total >C6 acids 35.2 31.9
Ca(I1) ( M I

.,
Fig. 5 . Formation of carboxylic acids as a function of Ca(II)
concentration (0.025 M D-fructose, 0.01 M KOH, H20, Ca(II)
supplied as CaCl,, 7 8 ° C N , , 5 h, 100% conversion); grey
region, > C, acids: 8,< C , acids: @, saccharinic acids: A ,
glycolic acid: 0 ,lactic acid: V ,formic acid: m, acetic acid:
2.4-dihydroxybutyric acid.
In addition, complexation of Ca(I1) with reaction inter-
mediates such as enediol anion species and a-dicarbonyl
compounds, as weil as with product a-hydroxy acids, will
occur. This is indicated by the increase of the benzilic acid
rearrangement/dicarbonyl cleavage ratio of a-dicarbonyl
intermediates upon increasing the Ca(I1) concentration. The
slight decrease in >C6 acids shows that Ca(I1) has only a
small influence upon the oligomerization reacti0n.h order to
, H@ ..Ca (11)
a Reaction conditions: 0.025 M D-fructose; 0.03 M (OH and
0.08 M Ca(OH), (-0.03 M HO-),respectively; H,O, 78"C, N,,
4 h, 100% conversion.
The influence of the HO- concentration upon the alkaline
degradation of monosaccharides in the presence of Ca(II),
see Fig. 7, resembles that found in the presence of mono-
valent cations (Fig. 2).
EIfect of borate
The possible influence of inorganic anions upon the alkaline
degradation of monosaccharides has been investigated for
chloride, carbonate and borate; chloride, since it is present
in the reaction mixtures used to determine the influence of
Ca(I1) upon the degradation pattern, carbonate, because of
the carbonation of the alkaline juice in the industrial juice

ti0 7
ti 0 y;$;-;.
H

..
HO/ll
(OH

0
+
cOe
HO----&II]

/ Oh
.--,Ca( I I )

/(0
H
Fig. 6. Promoted retro-aldolization of D-fructose by complexation with calcium(II).
Recueil des Travaux Chimiques des Pays-Bas, 10516,june 1986 181

The differences between borate ester formation and cal-

cium(I1) complexation of monosaccharides are ( a ) the
greater degree of covalent coordination of borate with the
monosaccharides and ( b )the proton removal of both
hydroxyl groups involved. Consequently, formation of these
borate ester moieties has a stabilizing influence upon mono-
saccharides in alkaline medium, thus protecting them against
fast degradation3,; borate ester formation at the anomeric
site of the monosaccharide is expected to inhibit the
enolization, while the second esterification ofborate with two

40t other hydroxyl groups of the monosaccharide may inhibit the

retro-aldolization. In addition, borate esters of subsequent
reaction intermediates apparently favour the benzilic acid
rearrangement at the cost of the a-dicarbonyl cleavage.

HO HO OH
\-/
?-0

0 k
0 -3 -2 -1
log [HOj
~~

HO'
0-
'OH
A
qp
B-
HO' 'OH
B
Fig. 7. Formation of carboxylic acids in the presence of Ca(II)

.,
as a function of the H O - concentration (0.025 M D-fructose, Fig. 8. P-D-Fructopyranosel12;4,5-diborate(A) and P-Dfruc-
KOH. 0.015M CaC12, H,O, 78°C. N,, 100% conversion): topyranose 2.3 ;4,5-diborate (B).
grey region, > c6 acids: @ <I,c, acids: ,saccharinic acids; Effect of temperature
A , glycolic acid and formic acid; 0 ,lactic acid: 0 ,acetic acid;
2,4-dihydroxybutyric acid. The influence of the reaction temperature upon the final
product composition has been studied at 50-90°C under
purification and borate, since it is to shift the standard reaction conditions (0.025 M D-glucose, 0.01 M
isomerization equilibrium of monosaccharides by borate KOH, H,O, N,, 100% conversion). As shown in Fig. 9, the
ester formation. For instance, an increase in the maximum degradation pattern is hardly influenced by the reaction
D-fructose concentration from 30% to SO% has been temperature which is in accordance with the data ofDe Wit''.
r e ~ o r t e d , ~on' ~adding
~
-
boronates to the alkaline isomeri-
zation of D-glucose, while 50% D-fructose is obtained in
the presence of borate36.
Apparently, the differences in activation energy for the for-
mation of the various degradation products are negligible.

The effect of chloride, carbonate and borate upon the final Degradation of some reaction intermediates
product composition of the alkaline degradation is given in
Table 111. Both carbonate and chloride have only a small Experiments have been carried out with some model reaction
effect upon the degradation pattern. This indicates that - as intermediates. Pyruvaldehyde (2-oxopropanal) was chosen
expected - no strong interaction between these anions and as a representative of a-dicarbonyl compounds, while
the reactive species occurs. Borate, on the other hand, glyceraldehyde and 1,3-dihydroxyacetone have been used as
markedly influences the final product composition in favour the intermediates formed by retro-aldolization of mono-
of the saccharinic acids. Under the conditions used, saccharides. Reaction ofthese compounds (0.05 M) has been
D-fructose is almost completely present as its borate esters, performed under standard conditions (0.01 M KOH, H,O,
of which the P-D-fructopyranose diborate esters A and B are 7S°C, N2). Since the degradation of monosaccharides con-
the main species (Fig. S),'. tinuously generates a-dicarbonyl compounds in very low
concentration, a degradation reaction was performed in
Table IZI Influence of inorganic anions upon carboxylic acid forma- which pyruvaldehyde was added in 50 small portions at
tion intervals of 2 min. The results are summarized in Table IV.
The composition of the reaction mixture after the alkaline
Carboxylic acid -b 0.03 M 0.03 M degradation of both glyceraldehyde and 1,3-dihydroxyaceton
(mol-C,-% ) c1- co: - resembles that of the monosaccharides. The formation of
saccharinic acids shows that substantial aldolization of C,
saccharinic acids 21.5 21.4 20.0 42' fragments occurs prior to the alkaline degradation of these
glycolic acid 5.0 4.8 5.2 2.2 intermediates. Since both C, and C, saccharides result in
lactic acid 8.5 6.2 8.7 7.0 comparable product formation, there appears to exist a
formic acid 4.0 4.0 4.5 1.5
acetic acid 10.5 8.3 10.1 3.7
2,4-dihydroxy-
butyric acid 8.0 7.4 8.7 6.0

total <C6 acids 57.5 52.1 57.2 62.4

total >C6 acids 42.5 47.9 42.8 37.6 34 J . F. Mendicino, J. Am. Chern. SOC.82, 4975 (1960).
35 S. A . Barker, B. W. Hatt and P. J. Somers, Carbohydr. Res. 26,
a Reaction conditions: 0.025 M D-fructose, 0.01 M KOH, H,O, 41 (1973).
78"C, N,, 100% conversion; anions added as KC1, K&O, and 36 J. M. de Bruijn, A . P. G. Kieboom and H. van Bekhm, to be
Na,B,O,. 1OH,O. Standard degradation without additives. published.
Analyzed using GC because of the coincidence of the borate and 37 M . Makkee, A . P . G . Kieboom and H . van Bekkum, Recl. Trav.
the saccharinic acid signals in the liquid chromatogram. Chim. Pays-Bas 104, 230 (1985).
182 J . M . de Bruijn et al. /Alkaline degradation of monosaccharides 111

Table I Y Alkaline degradation of some reaction intermediates".

Carboxylic acid Pyruvaldehydeb Pyruvaldehyde Glyceraldehyde Dihydroxyacetone D-Glucose

(mol-C,- %) 0.001 M 0.05 M 0.05 M 0.05 M 0.025 M
saccharinic acids <0.1 t0.1 16.8 18.9 21.5
glycolic acid <0.1 <0.1 5.2 3.7 5.0
lactic acid 51.7 15.5 7.6 6.1 6.9
formic acid 5.9 5.9 5.5 4.8 4.0
acetic acid 7.3 9.5 9.8 7.9 10.5
2,4-dihydroxybutyricacid 3.3 7.0 8.3 5.5 8.5
total <C6 acids 68.2 37.9 53.2 46.9 56.4
total >C, acids 31.8 62.1 46.8 53.1 43.6

a Reaction conditions: 0.01 M KOH, H,O, 78"C, N,, - 5 h, 100% conversion. Added in 50 portions over 100 min.

relatively fast equilibrium between C, and C, saccharides

I
-x-x
X involving aldolization and retro-aldolization reactions.
The experiments with pyruvaldehyde (which is also an
important intermediate in the alkaline degradation of
glyceraldehyde and ly3-dihydroxyacetone) show that dicar-
bonyls are important intermediates for the formation of > C,
acids. The formation of > C, acids decreases upon lowering

401
,. the dicarbonyl concentration, as would be expected for a
bimolecular aldolization reaction. This, favours the uni-
molecular benzilic acid rearrangement of pyruvaldehyde into

* O t * - w lactic acid. It may be noted that, during the alkaline

degradation of monosaccharides (including glyceraldehyde
and 1,3-dihydroxyacetone), a-dicarbonyl compounds will be
responsible for the formation of substantial amounts of > C,
acids by mutual aldolization as well as by aldolization with
other carbonyl compounds present in the reaction mixtures.
01 1 I I 1 1
50 60 70 80 90
Conclusions
T ("C)
Fig. 9 . Formation of carboxylic acids as a function of the The present results allow the modelling of the alkaline trans-
reaction temperature (0.025 M D-glucose, 0.01 M K O H , H20. formation of monosaccharides by means of the overall
N, , 100 % conversion); x , > C, acids; 0 ,saccharinic acids: reaction scheme, as presented in Fig. 10.
A , glycolic acid; 0 lactic acid; V ,formic acid; ,acetic acid;
my 2,4-dihydroxybutyric acid.

1
mannose
triose ( 2 x ) - 2
pyruvaldehyde
3
+lactic
acid

glucose & 1.2-enediol fructose

2,3-enediol
4
C , a-dicarbonyl dicarbonyl
cleavage
6 products

aldolization
products
saccharinic
acids

oligomeric
acids

Fig. 10. Overall reaction network of monosaccharides in alkaline medium involving the following transformations:
1 enolizationlisomerization,
2 p-elimination,
3 benzilic acid rearrangement,
4 a-dicarbonyl cleavage,
5 (retro-)aldolization,
6 extended (retro-)aldolization of (di)carbonyl compounds.
Recueil des Travaux Chimiques des Pays-Bas, 10516,june 1986 183

The main characteristics are:
(i) Enediol anion species are involved in both isomerization
and degradation reactions as shown by the formation of
similar product mixtures upon alkaline degradation of
various interconvertible monosaccharides.
(ii) A fast equilibrium occurs between C, and C, sac-
charides, involving aldolization and retro-aldolization
reactions, prior to the subsequent alkaline degradation
reactions.
(iii) Dicarbonyl compounds, the P-elimination products of
enediol anion species, are key-intermediates which, via
benzilic acid rearrangement, a-dicarbonyl cleavage and
aldolization, lead to a variety of acid products. It may be
noted that aldehydes formed by the a-dicarbonyl
cleavage reaction will also be involved in aldolization
reactions. The elongation of the carbon chain by aldoli-
zation of (di)carbonyls is terminated by benzilic acid
rearrangement or a-dicarbonyl cleavage of oligomeric
dicarbonyl intermediates.
(iv) Oligomeric acidic products, the so-called > C, acids, are
formed via extensive aldolization of intermediate
(di)carbonyl compounds. An intermediate H O - concen-
tration (10-3-10-2 M ) is responsible for the occurrence
of substantial amounts of both neutral and ionized
(di)carbonyl compounds, as required for optimal aldoli-
zation, while a high monosaccharide concentration
(> lo-* M) generally favours the oligomerization
reaction.
(v) Retro-aldolization of the ketoses is a major reaction
pathway leading to lactic acid formation. Enhanced ioni-
zation of the ketose (high HO- concentration) or the
formation of cation-ketose complexes (Ca(I1) addition)
increases the retro-aldolization.
Addition of borate stabilizes the monosaccharides
through borate ester formation.
Acknowledgements
This investigation was carried out with support from CSM
Suiker B.V. We wish to thank Dr. P. W. van der Poel,
N . H . M.de Visser and M . A . M.de Schutter of Central Labo-
ratory, CSM Suiker B.V., Breda, The Netherlands, for help-
ful discussions.

Red. Trav. Chim. Pays-Bas 105, 183-185 (1986) 0165-0513/86/06183-03$1.25

1,3-Halogen shift of 2-halogeno-4,5-dimethylphenolsfollowing oxidation with

Frenzy's salt

A. W. C. M. Jacobs* and L. Maat

Laboratory of Organic Chemistry, Technische Hogeschool Delft, Julianalaan 136, 2628 B L Delft,
The Netherlands
(Received November 20th. 1985)

Abstract. Oxidation of 2-bromo-6-deuterio-4J-dirnethylphenolwith Fremy's salt, followed by in situ

reduction, yielded mainly 3-bromo-4,5-dimethylpyrocatecho1**,which affords evidence for a
1,3-bromine shift during the oxidation. The same treatment of 2-chloro-4,5-dimethylphenolyielded
4,5-dimethylpyrocatechol with a minor amount of 3-chloro-4,5-dimethylpyrocatechol.The 2-iodo
compound gave 4,5-dimethylpyrocatechol as the only reaction product

Introduction

Earlier investigations in this laboratory have shown that the

oxidation of 2-bromo-3-hydroxy-N-methylmorphinan-6-one * Present address: Naarden International, P.O. Box 2, 1400 CA
with Fremy's salt (potassium nitrosodisulfonate), followed by Bussum, The Netherlands.
in situ reduction, is accompanied by a migration of the ** Pyrocatechol is 1,2-benzenediol.
bromine atom from the 2- to the 4-position, yielding, un-
X OH

CH3 CH3 CH3 CH3

la X = Br, Y = H 2a X = Br 3a Y = H 4a Y = H
Ib X = C1, Y = H 2b X = C1 3b Y = D 4b Y = D
IC X=I,Y=H
Id X = Br, Y = D

Scheme 1 . Fremy's salt oxidation of 2-halogeno-4,5-dimethylphenols(la, l b , lc, l d ) followed by reduction.