Institute I
The of Paper Chemistry
Appleton, Wisconsin
Doctor's Dissertation
The Effect of the Aglycon and Hydroxyl Orientation on Alkali-Oxygen Degradations of Methyl Glycosides
David 0. Hearne
June, 1978
II__
TABLE OF CONTENTS Page SUMMARY INTRODUCTION Perspective Literature Review Oxygen-Alkali Reducing End Reactions Chain Cleavage Reaction Mechanism Thesis Objectives RESULTS Glycoside Degradations General Oxygen Solubility Reaction Exotherms Kinetics Background Methyl Glycoside Kinetic Orders Products Methanol Peroxides General Hydrogen Peroxide Organic Peroxides Acidic Products General Methoxyacetic Acid Lactic, Glycolic, and Glyceric Acids Methyl C-Carboxyfuranosides Oxygen Deficient Degradations 1 3 3 3 4 5 8 10 12 12 12 15 16 17 17 19 21 21 24 24 26 26 29 29 30 35 36 40
iii Page DISCUSSION Introduction Mechanism Mechanism Proposed by Millard (1,2) Autoinhibition Mechanism Proposed by Millard (1,2) Autoinhibition in Methyl Glycoside Degradations General Stable Intermediate Alkyl Radical Effect of Oxygen Pressure Hydroxyl Orientation Rate Effect General Factors Affecting Carbohydrate Acidity Methyl Riboside and Methyl Xyloside Acidity Stabilization of Methyl Glycoside Oxyanions Interrelationship of Oxyanion and Mechanism Fate of the Glycosidic Bond General Implications of Methanol Formation Stereoselective Formation of Methyl C-Carboxyfuranosides Alternative Chain Cleavage Mechanism CONCLUSIONS EXPERIMENTAL General Analytical Procedures Solutions and Reagents Titanium Sulfate Reagent (42) Purified Sodium Hydroxide Sodium Thiosulfate Solution GLC Internal Standards 46 46 46 46 51 52 52 53 57 59 59 60 61 63 65 66 66 67 68 72 74 76 76 77 77 78 79 79
iv Page n-Propyl 8-D-Xylopyranoside Ethanol Model Compounds Methyl ~-D-Xylopyranoside (MBX) Methyl B-D-Ribopyranoside (MBR) Method I Method II Reaction Analysis Conditioning of Reactor and Glassware Preparation of Reaction Solution Glycoside and Methanol Analysis General Glycoside Analysis Methanol Analysis Peroxide Analysis Peroxide Sampling Peroxide Analysis Product Analysis Product Sampling and Derivatization Mass Spectral Analysis Time Distribution of Products Preparative GLC-PMR Analysis Oxygen Solubility ACKNOWLEDGEMENTS LITERATURE CITED APPENDIX I. APPENDIX II. GAS-LIQUID CHROMATOGRAPHY EXPERIMENTAL DATA 79 79 79 79 80 80 81 82 82 82 83 83 83 84 85 85 85 86 86 88 88 89 89 91 92 96 99
v
Page APPENDIX III. APPENDIX IV. MASS SPECTROMETRY ANALYSIS REACTOR SYSTEM 106 116
SUMMARY
Methyl $-D-ribopyranoside (MBR) and methyl B-D-xylopyranoside (MBX) were degraded by molecular oxygen (0.682 MPa partial oxygen pressure) in a Teflonlined reactor at 120C and 1.25M sodium hydroxide. The degradations were
followed by quantitative gas-liquid chromatography and were analyzed for both degradation rate and product appearance. The glycosidic bond was found to
have little effect on the overall rate or pattern of pyranoid ring degradation in oxygen-alkali. Consistent with previous workers (1-3), however, ring MBR degraded
hydroxyl orientation dramatically affected the reaction rate.
at an initial rate approximately six times the initial degradation rate of MBX. It is postulated that the cis 1,2-diol configuration in MBR favors
stabilization via intramolecular hydrogen bonding of the oxyanion formed in the first step of the proposed mechanism of oxygen-alkali degradation of carbohydrates. This stabilization effectively increases the oxyanion concentra-
tion in the MBR system leading to an increase in the frequency of initiation reactions and the degradation rate. Consistent with this hypothesis, MBR
produced more hydrogen peroxide in oxygen-alkali than did MBX.
Kinetic analysis showed the time dependent orders of MBR and MBX degradation to be 2.75 and 3.5, respectively. Due to similarities in degradation
rates and reaction pathways of MBR and MBX, the mechanism for the degradation of the 1,5-anhydroalditols was assumed to be applicable (1-2). predicts the degradations to be second order in carbohydrate. This mechanism The kinetic
orders of MBR and MBX were interpreted to imply autoinhibition, i.e., an intermediate in the reaction serves to abnormally slow it down. This auto-
inhibition is proposed to be the result of a termination reaction between an alpha-hydroxyhydroperoxyl radical and the frequently proposed C-l radical resulting from anomeric hydrogen atom abstraction. Formation of stable
organic peroxides in both methyl glycoside systems supported this hypothesis.
-2-
The major acidic reaction products were identified as their per-0-trimethylsilyl derivatives by gas-liquid chromatography-mass spectrometry. The
formation of lactic, glycolic, and glyceric acids was qualitatively the same as previously reported in other carbohydrate systems (_-3). A methyl 3-C-
carboxy-3-D-tetrafuranoside and isomeric methyl 2-C-carboxy-B-D-tetrafuranosides were identified as major "bound methanol" products. The relative ratios
of these methyl C-carboxyfuranosides varied between the MBR and MBX systems thus supporting the previously observed stereoselective formation of these furanoid acids (1). Methoxyacetic acid was observed for the first time as a The formation
product of the oxygen-alkali.degradation of methyl glycosides.
of this product suggests an alternative polysaccharide chain cleavage mechanism to 8-alkoxy elimination.
Varied C-3 hydroxyl stereochemistry did not affect glycosidic bond cleavage as determined by a gas-liquid chromatographic methanol analysis. The
yields of methanol were identical from both MBR and MBX even though MBR degraded at a much faster rate. The reaction step determining the relative rate
of reaction between MBR and MBX apparently occurs prior to the methanolliberating reaction, a result consistent with the above hypothesis concerning the oxyanion. Although the C-l radical was postulated to play an important
role in autoinhibition, the degradation rate, product, and methanol data implied that C-l radical formation and decomposition did not represent a major degradative pathway for the methyl glycosides.
-3-
INTRODUCTION
PERSPECTIVE
The polluting nature of pulping and bleaching processes utilizing chlorine and sulfur has been widely publicized. Likewise, the potential
advantages of delignification with molecular oxygen and alkali are well established. Hindering the full development of oxygen-alkali processes,
however, is the- harmful degradation of wood polysaccharides in oxygenated, alkaline solutions. This carbohydrate degradation results in a loss of pulp
viscosity and mechanical strength.
A potential solution to the carbohydrate degradation problem involves developing an understanding of.the degradative reactions. Once the mechanism
is understood then the possibility exists to devise methods of intercepting the carbohydrate-degrading reactions. This investigation was designed to namely, the effect of
examine certain aspects of the degrading mechanism:
the glycosidic linkage on carbohydrate reactivity to oxygen and alkali, and how glycosidic bond cleavage by oxygen-alkali changes with varied ring hydroxyl stereochemistry.
LITERATURE REVIEW
Conventionally the reactions of cellulose and other wood polysaccharides in alkali are divided into two categories - peeling and chain cleavage. The
peeling reactions occur at the reducing end of the polymer chain and consist of the stepwise removal of reducing end groups. Chain cleavage involves the Similarly,.
breaking of the polymer chain between two inner polymeric units.
the reactions of wood polysaccharides in oxygen and alkali can be categorized into these two groups - reactions at the reducing end group and chain cleavage reactions. This review will deal only briefly with the reducing end reactions
-4-
as they are generally considered to be unimportant in the oxygen-alkali treatment of wood pulps (3-7). Since oxygen-alkali treatment of wood and
wood pulps causes significant viscosity loss without adverse yield loss (7) the chain cleavage reactions are the more important ones to understand. review will present some of the major studies on chain cleavage reactions with emphasis being placed on model compound studies. Finally, an overview The
of the mechanism of oxygen-alkali degradation of polysaccharides will be presented.
OXYGEN-ALKALI REDUCING END REACTIONS
The apparent stability of oxygen-alkali pulps to alkaline peeling has been attributed to the formation of aldonic acid end groups (4-6,8-11). These end groups prevent the B-alkoxy elimination reaction characteristic of the peeling mechanism. Samuelson and coworkers (4,5) investigated the
aldonic acid end groups in cellulose after oxygen bleaching, detecting arabinonic, mannonic, and erythronic acids as the major end groups. Metasacchar-
inic acids, which are commonly associated with peeling, were found only in minor amounts. Rowell (6) studied the degradation of cellobiose by alkali In the absence of oxygen, alkali degraded cellobiose
and oxygen-alkali.
completely to monomer units.. But when oxygen was present glucosylaldonic. acids were produced in up to 27% yields.
Glycosuloses have been postulated as intermediates in the formation of aldonic acid end groups in wood polysaccharides treated by oxygen-alkali (8-11). By comparing the distribution of aldonic acids formed from glucosone
and xylosone with the distribution of acid end groups produced respectively from cello-oligosaccharides and xylo-oligosaccharides, Malinen (8) concluded that glycosuloses were likely intermediates in the formation of the aldonic acid end groups.
-5-
In summary, some stabilizing effect to peeling reactions seems operative in oxygen-alkali treatments of wood polysaccharides. However, as noted by
Malinen (8) some care must be observed in the reaction conditions employed. In his study, hydrocellulose did not always show stability to peeling and the distribution of aldonic acid end groups in these cases differed from the distribution found from glucosone.
CHAIN CLEAVAGE
Chain cleavage and ring opening reactions of polysaccharides have primarily been investigated by using mono- and disaccharide model compounds, most notably methyl B-D-glucopyranoside (MBG). Parameters which have been
investigated include number and position of hydroxyls (12,23), the effect of additives (13,14), the importance of peroxidic intermediates (1,2,12-17), and ring hydroxyl stereochemistry (1,2).
Brooks and Thompson (18) noted that oxygen increased the rate of MBG degradation many times over the rate of degradation in the absence of oxygen. The difference was postulated to arise from a lower activation energy for the oxidative reaction than for alkaline hydrolysis. It was proposed that in the
presence of oxygen a carbonyl group is introduced to the ring leading to a subsequent 8-alkoxy elimination of the C-l methoxyl group.
McCloskey, et al. (12) examined the oxygen-alkali degradation of MBG and a variety of its methyl ether derivatives. The importance of a free
hydroxyl group was emphasized by the fact that methyl 2,3,4,6-tetra-O-methyl B-D-glucopyranoside did not react in oxygen and alkali. Later Millard (1)
and Millard, et al. (2) determined that the stereochemistry of the ring hydroxyls was also important to the degradation reaction. Millard examined
the oxygen-alkali degradation of 1,5-anhydroxylitol (AX) and 1,5-anhydro-
-6-
ribitol (AR).
All neighboring hydroxyl orientations in AX are trans while The initial degradation rate of AR was ca. 7 times that
they are cis in AR. of AX.
Malinen and Sjostrom (3) have noted a similar relationship in reac-
tivity with varied hydroxyl orientation in their limited study of the oxygenalkali degradation of methyl @-D-glucopyranoside and methyl 8-D-mannopyranoside.
McCloskey, et al. (12) were also the first to note that MBG degradation in oxygen-alkali could occur without cleavage of the glycosidic bond. Their
data showed that the rate of methanol formation was slower than the rate of MBG degradation. Subsequent acid hydrolysis of the product mixture, however, Ericsson, et al. (15) in a later study
yielded the unaccounted for methanol.
of the oxygen-alkali degradation of MBG isolated an acidic product containing the methyl aglycon. This product was.identified as.a methyl 2-C-carboxy-8-D-
pentafuranoside and was proposed to be formed via a diketo intermediate (Fig. 1). The diketo intermediate was postulated to arise from the simultaneous The ring can then contract via a
introduction of keto groups at C-2 and C-3. benzylic acid type arrangement.
Subsequently 2-C- and 3-C-carboxyfuranosides
have been identified in a variety of degradations of methyl glycosides with oxygen-alkali (3).
Millard (1,2) in examining the oxygen-alkali reaction of AX and AR identified as products isomeric 1,4-anhydro-2-C-carboxy tetritols which are analogous to the furanoid acids reported by Ericsson, et al. (15). Depending
on the alditol examined, the ratio of the two isomers would change, suggesting a stereochemical effect to be operating in their formation. The mechanism
of Ericsson (15) could not account for this stereoselectivity and Millard (1,2) offered an alternative mechanism involving an alpha-hydroxyhydroperoxide intermediate.
-7-
CH2 0H H
CH20H 0H
0 OCH3 HOCH3
CH 2 OH
OCH H HOH
HO
OH-
H OH
HO L//
0
H HO
H
HO COOH
0
Figure 1. Mechanism Proposed for Formation of Methyl 2-C-Carboxyfuranosides (15) Hydrogen peroxide has frequently been reported as an intermediate in the degradation of carbohydrates by oxygen and alkali (1,2,12-14,16-18). Sinkey
and Thompson (13), in their work with MBG, provided strong evidence for the identity of hydrogen peroxide as an intermediate. The oxygen-alkali degradation
of MBG exhibited an induction period during which hydrogen peroxide increased in concentration. Additionally, a rough correlation existed between the end of the These results
induction period and the peak in hydrogen peroxide concentration.
paralleled the earlier work of Minor and Sanyer (17) on the oxygen-alkali degradation of glucitol. Sinkey (13) also examined the effects of magnesium, iron, In each case the effect of the
and iodide ion on the degradation of MBG.
additive could be related to the expected behavior of the hydrogen peroxide intermediate in the presence of the additive. Weaver (19) and Ericsson (15),
working independently, examined the alkaline hydrogen peroxide degradation of MBG. Their results demonstrated a commonality in products with the products
from the oxygen-alkali reaction of MBG. Sinkey (13) also reported that organic peroxides were formed in the oxygen-alkali degradation of MBG. These peroxides did not appear to be
reactive intermediates and were hypothesized to be dialkyl peroxides arising from radical chain termination reactions. Weaver (19) and Weaver, et al. (20)
-8-
provided evidence in their study for an intermediate organic peroxide which they proposed to be an alpha-hydroxyhydroperoxide. Millard (1,2) detected
an intermediate organic peroxide in the oxygen-alkali degradation-of AX. Based upon their experimental evidence this intermediate organic peroxide was also hypothesized to be an alpha-hydroxyhydroperoxide.
In addition to the.previously discussed furanoside products, numerous other acids have been reported as products in oxygen-alkali studies, most notably lactic, glycolic, and glyceric acids (1-3,15). Minor concentrations The one char-
of many other acids such as dibasic acids have been reported.
acteristic of all of-these acids is that they are proposed to be formed via a keto intermediate (1-3,12-15,18).
In summary, model compound studies have generally supported the reaction scheme adapted by Kolmodin and Samuelson (5) for the oxygen-alkali degradation of cellulose (Fig. 2). This mechanism was adapted from the mechanism
proposed by Haskins and Hogsed (21) for the depolymerization of cellulose during the aging of alkali cellulose. However, the mechanism of carbonyl
formation in the cellulose and how this carbonyl formation is related to hydrogen peroxide and organic peroxides remains unclear.
REACTION MECHANISM
This section will present a current view of the reaction mechanism of oxygen-alkali degradation of carbohydrates. It should be remembered that
although many of these reactions.are based on experimental observation and analogous autoxidation reactions, they are still largely speculative.
Initiation of oxidative degradation requires the ionization of an hydroxyl to form the carbinolate anion (1,2,12-14,17). The increased elec-
tron density provided by the carbinolate ion facilitates hydrogen abstraction
-9-
CH 2 OH
J o
CH 2 OH
CH 2OH
H
'0 H0
H, OH H + 0 HO H
CH2 OH
0-O 0H , o
D-GLUCOSONE END GROUP
DEGRADATION PRODUCTS
Figure 2.
DEGRADATION PRODUCTS
Mechanism for Cleavage of Cellulose by Oxygen in Alkali Proposed by Haskins and Hogsed (21)
by molecular oxygen (Fig. 3).
This view is consistent with reactions reviewed
by Russell (22) as important in autoxidation.
H
I
-C-
H OH -C- + H 2 0
0-
O OH
*OO
H
-C-
I
-.b I.
+ HOO. %c.
0-
-t Lto .. ? *. I 0 o= I
+ H00-
Figure 3.
Initiation of Carbohydrate Oxidative Degradation Proposed by McCloskey (23)
-10-
The reactions usually show induction periods (1,2,12,13,16,17) during which hydrogen peroxide increases in concentration. Subsequently the carbo-
hydrate degrades rapidly, probably via a free radical mechanism (1,2,12-14). The free radicals are generated by hydrogen abstraction with oxygen (Fig. 3) and metal-catalyzed decomposition of the perhydroxyl anion (l,2,12-14) Inter-
mediate hydroperoxides are hypothesized (1,2,14,17) to arise in a variety of reactions between the carbohydrate and radical species. The hydroperoxides
then decompose to acidic degradation products, possibly through carbonyl-containing intermediates (1-3,12-15,18). The presence of organic peroxides in
oxygen-alkali reactions has been noted (1,2,13,14) although their structures are as yet undetermined. Studies of the alkaline hydrogen peroxide reaction
of MBG suggested the presence of an alpha-hydroxy-hydroperoxide (20,21) and Millard (1,2) has suggested an alpha-hydroxyhydroperoxide as an intermediate in the oxygen-alkali reaction of the 1,5-anhydroalditols.. The peroxides noted by Sinkey (13,14) were believed to be dialkyl peroxides arising from termination reactions.
THESIS OBJECTIVES
As noted in the Literature Review, simple glycosides have frequently been used as model compounds in studies of wood polysaccharide degradation by oxygen-alkali. However, little is known as to how the aglycon affects the Millard (1,2) recently examined the oxygen-
reactivity of these glycosides.
alkali degradation of selected 1,5-anhydroalditols and obtained information on ring cleavage reactions and how these reactions changed with varied hydroxyl orientation. By examining the glycosidic analogs - methyl 0-D-ribo-
pyranoside (MBR) and methyl 8-D-xylopyranoside (MBX) - of the alditols studied by Millard, information on the effect of the glycosidic bond was obtained.
-11-
0.V
OCH 3
/OH
X\
CH 3
/ OH MBX
HO HO OH MBR
HO
Differences in reactivity between the alditols and the methyl glycosides were projected because of the following:
1)
the glycosidic bond enhances the acidity of the C-2 hydroxyl
(24-26),
2) the glycosidic bond enhances radical formation at C-1 through resonance stabilization (27,28), and 3) the methyl aglycon can compete with. the ring oxygen as a leaving group in 6-alkoxy elimination.
By examining glycosidic bond cleavage and the differences in kinetics and products between the glycosides and the alditols, information was obtained concerning:
1) 2)
the impact of the aglycon on pyranoid ring degradation, and the effect of varied hydroxyl orientation on glycosidic bond cleavage.
-12-
RESULTS
GLYCOSIDE DEGRADATIONS
GENERAL
To provide a common basis of comparison, the reaction conditions used were the same as employed by previous workers (1,2,12-14): NaOH, and 0.682 MPa partial oxygen pressure*. at a carbohydrate concentration of 0.1M. 120C, 1.25M
All reactions were conducted
The glycoside analyses were conducted by quantitative gas-liquid chromatography (GLC). The glycosides were analyzed as their per-0-acetylated
derivatives and n-propyl B-D-xylopyranoside was utilized as an internal standard. The NaOH was purified to minimize metal catalysis (29). The
experimental procedures are described in the Experimental section and chromatographic conditions are stated in Appendix I.
Duplicate degradations of both methyl B-D-ribopyranoside (MBR) and methyl 8-D-xylopyranoside (MBX) were conducted at the stated conditions. Excellent agreement was found between the duplicated runs as indicated in Fig. 4 (MBX) and Fig. 5 (MBR). The dashed lines in the two figures repre-
sent the degradation curves reported for the analogous 1,5-anhydroalditols (1,2). The methyl glycosides showed surprising similarity in reactivity to However, minor differences do exist between the degradation
the alditols.
curves of the anhydroalditols and the methyl glycosides which are unexplained by experimental error. Thus the glycosidic bond affects.the rate of de-
struction of the pyranoid ring only slightly and the dramatic effect on rate
*Total reactor pressure at 120C was 1.014 MPa (147.1 psia). This pressure accounted for 0.199 MPa (28.8 psia) water vapor and 0.133 MPa (19.4 psia) nitrogen.
-13-
o
I!
E-4
0
pt0
II
4-
0A
4i
U4~J
0
$4i
o~~~
h0C~
.CI
4.-=
0
I-
rJ0
4
0 0
>Nco
4~i
0 0-
0 Cd
Cd0 4J~Z
am
04
00~0
I.,---PM
QH
I 0
I
I 0
to0
I
I 0
I
I.
cm
I
I
00
*14
0
6
6
ieIpi/salow 'NOI11~N33NO3
6
6
0
6
-14-
0~~~
MLJ~~~~C4
0
(a0
c~o
0
V.-
0~~~~~~~~~~~~~~~~~~~~~
02~~~~~~~~~~~~~C~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~C
.c~~~~c
.~~~~~~~~~~~~~~~
~
~
~
~
D
00
a
ca
Cli
~
I
ePI
4jC~~~~~~~I0 04.
/~~~~~~~~~~~~~0
.4
01-
o
0 0
0
'NOI1~~~~~~~1N33NO3~-HC1 JB$IiiSO~~~~~~~~oW
0 0
0~~~~~~~~~~~~~~~~~~ N 0~~~~~~~~~~~~~~~0z V 0~~~~~~~~~~~~~~~~~~~~~~~~
-15that Millard (1,2) noted for hydroxyl stereochemistry extends to the methyl glycosides. Neither MBX nor MBR showed any degradation after 72 hr at 120C
and 1.25M NaOH in a nitrogen atmosphere (see Experimental Data, Appendix II).
OXYGEN SOLUBILITY
In studies of oxygen-alkali degradation of carbohydrates, a frequently posed question is whether or not the reaction is oxygen diffusion controlled. Previous workers have interpreted their data to exclude this possibility (2,30,31). Recently an apparatus for determining solution oxygen concentraBy employing this apparatus,
tion at elevated pressures was developed (32).
solution oxygen concentration was determined at the described reaction conditions.
The experimentally determined saturated oxygen concentration was 2.9 x 10-3M which is 25% higher than the value of 2.2 x 10-3M obtained by a linear extrapolation of Bruhn, et al. data (33). In Table I oxygen dissolution
rate is illustrated for conditions of no agitation and with agitation in the presence of carbohydrate. In the presence of 0.1M MBR, ca. 80% of saturation
was reached within 6 minutes and saturation was achieved within 30 minutes. As seen in Fig. 5, MBR degrades rapidly and a problem of oxygen diffusion should have shown up as a depressed oxygen concentration during the first 1.5 hr of reaction. Therefore, in agreement with previous studies (2,30,31)
it was concluded that oxygen diffusion did not present a kinetic problem. However, under conditions of no agitation the rate of oxygen dissolution was suppressed. This is important because contrary to previous reports (2,31)
a threshold stirring rate is necessary to insure adequate oxygen dissolution. To illustrate this fact, degradations were conducted under conditions of limited agitation. The results of these degradations are reported in the
section titled Oxygen Deficient Degradationso
-16-
TABLE I OXYGEN SOLUBILITY IN 1.25M NaOH AT 120C AND 0.682 MPa 02 PRESSURE Dissolved 02, x 103 M Run OS3 a 10 41 150 330 948 1554
C Run OS4
Time, min
0.8 0.9 1.4 1.7 1.9 2.1
6 32 60 133 297 450
2.3 3.0 2.9 2.7 3.2 3.0
aNo agitation or carbohydrate. bPressurized with oxygen at time zero. Cwith agitation and 0.1M MBR.
REACTION EXOTHERMS
Reaction exotherms have been noted in oxygen-alkali pulping and bleaching experiments (34) and oxygen-alkali treatment of isolated lignin (35). this work a temperature rise of ca. 1.3C was noted during the MBR degradations. Very slight exotherms (ca. 0.2-0.3C) were noted during the MBX In
degradations, but it was difficult to determine the exact temperature rise as this temperature variation corresponded closely to the control limits of the bath. No temperature rises were noted in the oxygen solubility experiIn the case of MBR, the rise in tempera-
ments when carbohydrate was absent.
ture began 5-10 minutes after oxygen was charged to the reactor and reached a maximum after 0.7-1 hr. The temperature remained at this level for ca. 0.5
-17-
hr and then slowly declined back to 120C.
This is the first report, to the
author's knowledge, of an exotherm during the oxygen-alkali degradation of a carbohydrate.
KINETICS
Background The kinetics of the methyl glycoside degradations were examined to determine the order of the degradations with respect to glycosidic concentration. A comparison with the kinetics of the oxygen-alkali degradation of the 1,5anhydroalditols provided information as to the effect of the methyl aglycon of the degradations.
The two most common methods of kinetic order determination are the method of integration and the differential method (36,37). The method of integration
employs the integrated form of the rate equation and the data are plotted as dictated by the integrated form after first assuming a reaction order. Be-
cause the analysis is conducted on a single reaction plot, the resulting order is with respect to time. The differential method involves plotting the
logarithm of the experimentally determined reaction rate against the logarithm of concentration. This yields a line which has a slope equivalent to order.
If several reactions are conducted, the initial rates and concentrations can be used to determine the true order. If a single reaction plot is utilized
the order determined is said to be with respect to time.
The method of integration allows for a quick check for integral orders, but suffers from sensitivity to slight deviations from integral order (37). Additionally, only in simple reactions do true order and time dependent order correspond. The differential method offers the advantage of not
having to assume the reaction order, thereby eliminating the trial and error approach of the method of integration. The disadvantage of the differential
-18-
method is that rates must be determined graphically from reaction plots. While errors in rate determination are not significant when applying the differential method to initial rates determined at a wide range of concentrations, errors in rate determination from a single reaction plot can be considerable and cause a loss in sensitivity. A convenient cross-check of order
determined by application of the differential method to a single reaction plot is to apply the order thus determined to the method of integration.
In this work the rate of carbohydrate degradation is assumed to be a function.. of carbohydrate, oxygen, and alkali concentrations as expressed in Equation (1).
-
d[ dt
= k[A]n[OH-]m[O
(1)
where
[A] = glycoside concentration - d[A]/dt = rate of glycoside disappearance k
=
rate constant
[OH ] = alkaliconcentration [02]
=
oxygen concentration
n,m,l = reaction orders
By examining the reactions under conditions of excess oxygen and alkali concentrations such that they are in apparent constant concentration through the course of the reaction, then Equation (2) may be written (36,37):
- diA] = k'[A]n dt where k' = k[OH-]m[02] termed psuedo orders.
.
(2)
Kinetic orders determined in this manner are
While solution oxygen concentration was not, strictly
speaking, in excess relative to carbohydrate concentration, it was continuously introduced into solution by agitation. It has already been shown (see
-19-
previous discussion - Oxygen Solubility) that the rate of oxygen dissolution did not affect the degradation rate and therefore the assumption of constant oxygen concentration was valid. Initial alkali concentration was 1.25M NaOH
which was an order of magnitude higher than the initial carbohydrate concentration. At long reaction times sufficient concentrations of acidic products
are produced to depress the alkali concentration and thus limit the validity of assuming a constant alkali concentration. Millard (1,2), however, showed
that this ca. 10-15% drop in alkali concentration did not apparently affect the kinetics.
To apply the differential method, the logarithms of both sides of Equation (2) are taken as shown in Equation (3). log {- d[A]/dt} = log k' + n log [A] (3)
Thus, a plot of log rate versus log concentration should yield a line of slope n (order). Application of the method of integration is made by employing the
integrated form of Equation (2) which may be written for all n except unity as in Equation (4). k' t (n-l)
=
[A] (n-l)
[A] (n-l)
(4)
where [A
]
= initial concentration.
If a plot of time versus 1/A(n)
is
linear, the assumed order is correct.
Methyl Glycoside Kinetic Orders Application of the differential method to the methyl glycosides of this study yielded orders of 2.75 for MBR and 3.47 for MBX (Fig. 6). In contrast,
Millard determined orders of 2.0 for 1,5-anhydroribitol (AR) and 3.46 for 1,5-anhydroxylitol (AX) under the same reaction conditions. The orders derived
for MBR and MBX by the differential method were supported by applying these
-20-
-4.5
m
-5.0
x
m
x
X
0
I
-.
Slope = 2.74
X
/
o/
0
. 0
-6.0 1-5.5I E 5 -6.0 o
0 -j
.B
X
0
0 0
/
-6.5
I Im
0
Slope = 3.47
-7.0
I
i(
-2.0
A
I
/
/
I I I.0 -I.0
-1.8 -1.6 -1.4 -1.2 LOG CONCENTRATION, moles/liter
Figure 6. Degradation of 0.1M Glycosides (X - 6MBR,O- 8MBX) in 1.25M NaOH at 120C and 0.682 MPa 02. Data Plotted According to the Differential Method
-21-
orders to the integrated rate equation.
Linear relationships were obtained
for both MBR (Fig. 7) and MBX (Fig. 8) although at long reaction times the MBX relationship breaks away from linearity. order of 3.5 is also seen in Fig. 6. This divergence from an MBX
A similar but less dramatic break from
the derived order was exhibited by MBR.
In comparing the kinetics of the anhydroalditols versus methyl glycosides subtle differences were noted. First, although both MBX and AX showed a
kinetic order of ca. 3.5 through the first 50% of reaction, MBX exhibited a lower reaction order at long reaction times. reaction order. AX showed no such change in
This divergence in reaction order occurs at the same point Secondly,
that the reaction plots of MBX and AX begin to diverge (Fig. 4).
AX showed a tendency to higher reaction orders at the beginning of the reaction while MBX did not. Likewise, the subtle differences in the reaction plots of It was
MBR and AR (Fig. 5) are manifested in differences in kinetic order.
primarily the difference in kinetic order between MBR (2.75) and AR (2.0) and the difference in organic peroxides, which are reported later, that led to the interpretation of what effect the methyl aglycon has on the degradations. This interpretation is presented in the Mechanism section.
PRODUCTS
METHANOL
To determine the extent of glycosidic bond cleavage and how this cleavage might vary between MBR and MBX, the formation of methanol with time was examined. The methanol was analyzed by quantitative GLC (38) employing ethanol The method is described in the Experimental section
as an internal standard.
and the GLC conditions are tabulated in Appendix I.
-22-
6000
5000
O0 a
H
/
3000X0
N
.
/
/ /
/
0
1000
3000 0O
20
30
40
5
1000 0
10
20 TIME, HR, .TIME, HRE
30
40
50
Figure 7. Kinetic Plot of Reaction 6MBR According to the Method of Integration 7. Assuming an Order of 2.75 O 2000
-23-
0
0 c-
v.
.~0
0
o
o
0
\C0
0v60
0
\\140
\oo
0
O OO \0 \tOc
o
.'4-
A\O
\'0 -H C1
0oo0 CUO
C
\*
t
r l
e\o "
r l
N
5 5 'oo
z[Q
(
ao
o.0(a80M/s9TOm
0 ROI~~laKD9ES0O
0M
NO~uIaaNaONOO awiSOOLT)/L
-24-
The yield of methanol from both MBR and MBX was identical.
This is
illustrated by a plot of methanol yield versus percent reacted glycoside (Fig. 9). For both MBR and MBX, glycosidic bond cleavage can account for ca. 60% In other words, the percentage of glycosidic bond
of the total degradation.
cleavage is unaffected by the reactivity difference between MBR and MBX and therefore dependent on a common factor. This percentage of glycosidic bond
cleavage compares favorably with the data of McCloskey (23) on the formation of methanol from methyl B-D-glucopyranoside. However, Malinen (3) in his
study of the oxygen-alkali degradation of several methyl glycosides reported much higher methanol yields (80-95%) and the yields varied with the methyl glycoside degraded. The discrepancy between this work and Malinen's work
(3) could be a function of differences in trace metal contamination as suggested by Weaver's data (19). Alternatively, the similarity in methanol
yields in this work and McCloskey's work (23) could be fortuitous.
PEROXIDES
General Hydrogen peroxide and organic peroxides were shown to be different in the oxygen-alkali degradation of AX and AR and it was these differences which allowed Millard (1,2) to explain the differences in kinetics that AR and AX exhibited. Similarily, peroxide formation from the methyl glycosides was
monitored to assist in the interpretation of the kinetic and rate data.
Hydrogen peroxide and organic peroxide concentrations were determined by the acidic titanium(IV) sulfate method used by previous investigators (1,2,12-14,39-41). Under strongly acidic conditions hydrogen peroxide will In
form a complex with titanium(IV) which can be measured colorimetrically.
the presence of strong acid (pH 0-1) organic peroxides will hydrolyze to yield
-25-
70
O
O
on
060
C
X
0
( 40
0
D- 5MBR 0O- EMBX
'30
20
3 100
I I , I , , ! f ' I
0
20
40
60
80
100
PERCENT REACTED GLYCOSIDE Figure 9. Yield of Methanol as Function of Percent Reaction During Degradation of 0.1M Methyl Glycosides in 1.25M NaOH at 120C and 0.682 MPa 02 hydrogen peroxide which can be complexed by titanium(IV).
Hydrogen peroxide com-
plexes immediately with the titaniumCrV) and therefore the initial absorbance reading is taken as a measure of hydrogen peroxide concentration. Any increase
in absorbance with time is taken to be proportional to organic peroxide concentration. Because the organic peroxides can degrade by other paths that do not
yield hydrogen peroxide, it has generally been assumed that this method gives a low estimate of organic peroxide concentration (1,2,19). Peroxide concentra-
tions were determined on duplicate runs and excellent agreement was noted (see Appendix II, Experimental Data).
-26-
Hydrogen Peroxide The formation of hydrogen peroxide and organic peroxide in the oxygenalkali degradation of MBR and MBX is shown in Fig. 10. Hydrogen peroxide
from the degradation of MBR is seen to have peaked at a concentration ca. 2.5 times the peak in hydrogen peroxide concentration formed from MBX. The con-
centration levels of hydrogen peroxide produced and the ratio between hydrogen peroxide from MBR and that from MBX parallels hydrogen peroxide formation from the analogous 1,5-anhydroalditols (1,2). The early peak in hydrogen
peroxide concentration relative to total reaction time is consistent with the data of previous investigators (1,2,12-14,17). It is also interesting
to note that the hydrogen peroxide curves correlate well with the previously discussed reaction exotherms.
Organic Peroxides As can be seen from Fig. 10, both methyl glycosides formed organic peroxides which continually increased in concentration. This pattern of
organic peroxide formation is in sharp contrast with the trends in organic peroxide formation from the anhydroalditols (1,2). Organic peroxides were
only occasionally noted with AR and then only at detection limit levels. Intermediate organic peroxides were formed from AX and their formation and decomposition mirrored hydrogen peroxide formation. Only at longer reaction
times (5 hr) were any other organic peroxides detected from AX and then the concentrations were an order of magnitude less than the organic peroxide concentrations noted in this work. This variation in organic peroxides is
the most marked difference between the anhydroalditols and the methyl glycosides when they are reacted in oxygen-alkali.
Many attempts were made to detect an early, intermediate organic peroxide in either methyl glycoside degradation. The peroxide tests were
-27-
I0
51
8
61-
0
X
w
a
410
U
w 2 a
-
W
I
0
2
I ---
I 1__
4 TIME, hr
6
-----------
I
- - --- - - - -
I
1
8
10
I
o0
51
0
x
0
\o / 0 H 20 2
A Orgalnic Peroxides
A
w 04
x
w 2
NS
a.
0
TIME, hr
Figure 10. Peroxide Formation During the Degradation of 0.1M Methyl Glycosides in 1.25M NaOH at 120C and 0.682 MPa 02
-28-
conducted at two pH levels [pH 0 which corresponded to Sinkey's work (13) and pH 1 which corresponded to Millard's work (1,2)]. The AX intermediate
peroxide was postulated to be an alpha-hydroxyhydroperoxide, a peroxide which should readily hydrolyze at pH 1. Organic peroxides formed later in
oxygen-alkali-carbohydrate degradations are generally considered to be dialkyl peroxides which hydrolyze with much more difficulty (42). In this
work, color development at pH 1 always lagged color development at pH O. Therefore, if an alpha-hydroxyhydroperoxide was produced early in the oxygenalkali degradation of MBR and MBX, the ratio of absorbance at partial color development at pH 1 to absorbance at full color development at pH 0 might be different early in the reaction than the ratio exhibited at later reaction times. times. No difference in this ratio was noted between early and late reaction The velocity of color development was also examined for variation Time to full color development was This
between early and late reaction times.
approximately the same for all samples, being ca. 24-36 hr at pH 0.
time for full color development compares favorably with the time reported by Sinkey (42) for the hydrolysis of organic peroxides produced in the oxygenalkali degradation of methyl B-D-glucopyranoside, but differs.markedly from the time reported by Millard (1). Millard reported full color development
in 20-28 hr at pH 1, suggesting organic peroxides which are much less stable than the organic peroxides detected in this work. Finally to examine the
possibility that an alpha-hydroxyhydroperoxide was formed but was hydrolyzed immediately and was being measured as hydrogen peroxide, the peroxides produced from MBX were examined at a test pH of ca. 2. The hydrogen peroxide
values measured (see Appendix II, Experimental Data) were comparable to other runs but no organic peroxides were detected after 24 hr of acid hydrolysis. The organic peroxides measured in this work are, therefore,
-29-
concluded to be dialkyl peroxides.
This does not preclude the existence of
an intermediate alpha-hydroxyhydroperoxide in the reaction mechanism, but it does suggest that the methyl aglycon alters the mechanism such that the dialkyl peroxides predominate in the oxygen-alkali degradation of methyl glycosides.
ACIDIC PRODUCTS
General As explained in the Introduction, differences in reactivity and reaction pathways had been expected between the anhydroalditols and the methyl glycosides. The major acidic degradation products from the oxygen-alkali degrada-
tion of MBR and MBX were examined to determine if the methyl aglycon caused either a shift in the type of acidic products formed or a shift in the distribution of the products.
The acidic degradation products were determined as their per-O-trimethylsilyl (TMS) derivatives. These derivatives were identified by GLC (relative The sample workSodium bisulfite
retention time) and GLC interfaced with mass spectrometry. up procedure was adapted from previous workers (1,2,19,43).
was employed in the work-up procedure in an effort to detect an intermediate similar to the one reported by Weaver (19,20).
The acidic products identified in the oxygen-alkali degradation of MBR and MBX were identical for both glycosides and are reported in Table II. sample chromatogram is shown in Fig. 11. A
With the exception of methoxyacetic
acid, these products (or analogous products) have been identified in previous investigations of the alkaline oxygen degradation of ring carbohydrates (1-3, 15). Methoxyacetic acid was identified as a product of the alkaline hydrogen Although the products of
degradation of methyl g-D-glucopyranoside (19).
-30-
MBX and MBR were the same, the relative ratios of the "bound methanol" products (products with the methyl aglycon intact) varied between the two methyl glycosides. The methyl C-carboxyfuranosides were examined with time and, as evidence for a stereoselective reaction
with the 1,5-anhydroalditols (1), was found.
The formation of methoxyacetic acid was also found to vary beNo attempt was made to quantify lactic, glycolic,
tween the methyl glycosides.
or glyceric acids other than to note that their distribution was qualitatively the same as reported by other investigators (1-3,15). Many other minor products
were evident in the GLC chromatogram, but they were not identified.
TABLE II METHYL GLYCOSIDE DEGRADATION PRODUCTS Identification Methoda GLC, MS GLC, MS GLC, MS GLC, MS MS, PMR MS, PMR
Product 1 2 3 4 5 6,7
ification Product Ident: Methoxyacetic acid Lactic acid Glycolic acid Glyceric acid Methyl 3-C-ca:rboxy-a-Dtetrafurano!side Methyl 2-C-car rboxy-8-Dtetrafurano!side
GLC
by relative GLC retention time as compared to knowns MS - identified by mass spectrometry PMR - identified by Fourier transform-proton magnetic resonance spectrometry
-identified
Methoxyacetic Acid @* Methoxyacetic acid has been previously reported as a product of the alkaline hydrogen peroxide degradation of methyl a-D-glycopyranoside (19), but never as a product of the oxygen-alkali degradation of a methyl glycoside.
*The circled numbers in the text refer to the chromatography peaks in Fig. 11.
-31-
I
C\M
N~~~4 0 0
0
44
0
0
0
C-H
$40
.1-I$
C\J.~~~~
~ OSW
~
~
~~~~~
H
CUj $ 00 0
CM
0
$4 C.,
CD
b,
LAW
-32-
By examining the behavior of known methoxyacetic acid subjected to the sample work-up procedure, the identification of product @ as methoxyacetic acid
was substantiated and clues provided as to why it had not previously been identified in oxygen-alkali reactions.
The mass spectrum of product
0
is shown in Fig. 12 and for comparison Principal fragments
a spectrum of known methoxyacetic acid is also presented.
are ascribed to m/e 162 (M), m/e 147 (M-CH 3), m/e 117 (M-CH3 OCH 2), and m/e 89 (M-TMS). As noted in Fig. 12, these spectra were taken on samples derivaIn Fig. 13, spectra of product
tized by alternative methods.
Q
and known
methoxyacetic acid are also presented.
These spectra were made from samples
which had been carried through the normal work-up procedure employing sodium bisulfite. Note that although strong peaks beyond m/e 162 are exhibited,
the same peaks are present in both samples, i.e., m/e 177, m/e 191, and m/e 206.
Although much-difficulty was encountered in analyzing product
, the
belief that it is methoxyacetic acid is supported by the fact that known methoxyacetic acid consistently exhibited identical behavior to product A summary of this behavior follows: .
1.
Methoxyacetic acid and product @ times.
had identical GLC retention
2.
Although methoxyacetic acid is easily trimethylsilylated by itself, in an artificial product mixture a response for methoxyacetic acid was obtained only when NaHSO 3 was used in the sample work-up. product O Similarily, reaction product samples exhibited The use of
only when NaHS03 treatment was used.
NaHS03 in an artificial product mixture not containing methoxyacetic acid did not give any response resembling product .
-33-
0
u
Li
*14-)
a)
0 Cd
0 a.) H -H
CO) Cl
0
U') x<
~
~
a)
4
Ea) 0
91 ~
- L,
*r4I
E91-
A
w w
~a)
Cd 0 W, -H
I
bOH
4-)4 P4
Ha) 04-Ja *H 0
Cl)i1
4JC
~H
a)I Cl a
0 -H 0
E
04) HCd
C, 0~
E
-r
4.i4
0~~~ 0 Cd P
-M
ju 4i u C.)
LD
~~~~0.
cd ur
C4. I I I
0
-
CO
CD
(D
CD
--C\J d
0
6
0 0
'
to
0
0
LD
0
1d
0
a)
AkiISNJiNI
3AI1V138
-34-
ci
4-
C3~~~~C
ci~~~~~~1 C3 i
$4-
0
g9 1
- w
E91
4-S
0c N owa
-.
E
U) 4 04-
o
~~~~P
-a
A4
ci~~~~~~c
ci~~~~~~c
-4
Cs
ci
-
I
ai
I
c I
'If
I ci
I
ci eN~
ci ID
6
-
i
0
I
0
I
0
I
C)
ci ID
-'4
AiISN31NIl
3AI 1V-13 U
-35-
3.
As described above, mass spectra of samples employing NaHSO
3
treatment gave identical fragments for both methoxyacetic acid and product Q (Fig. 13).
It is concluded that methoxyacetic acid was identified as a degradation product due to the fortuitous use of NaHSO3. For the same reason, previous
workers have not identified it as a product in the oxygen-alkali degradation of methyl glycosides. Finally, as noted in Fig. 12, methoxyacetic acid (as
the free acid) can also be derivatized in a product mixture if a strong silylating reagent is used and thus obviate the awkward use of NaHSO 3.
Methoxyacetic acid concentration was examined in the final product mixture by two methods - a methanol balance and quantitative GLC. shows very good agreement between the two methods. Table III
The methanol balance was
done by summing the concentrations of unreacted glycoside, methanol, and the methyl C-carboxyfuranosides and subtracting this figure from initial glycoside concentration. the MBX system. Methoxyacetic acid formation seemed to be favored in
TABLE III METHOXYACETIC ACID IN FINAL REACTION MIXTURE Methoxyacetic Acid by Methanol Balance, % yield 14 25 Methoxyacetic Acid by GLC, % yield 14 26
Glycoside MBR MBX
% Reacted Glycoside 90 73
Lactic, Glycolic, and Glyceric Acids; Lactic @, glycolic
Q
, (
, @ acids have been reported as
Q
, and glyceric
major products in all product studies of oxygen-alkali-carbohydrate systems (4,10,15,16). Therefore, a discussion of their mass spectra will not be made.
-36-
The mass spectra are reported in Appendix III and these spectra are observed to have the same major fragments as the spectra reported in the literature for these acids (44). Qualitatively, the contribution of these acids to the
product distribution in this work agreed with the product distributions reported in previous investigations (1-3,15).
Methyl C-Carboxyfuranosides;
(
,
,
O
Three methyl C-carboxyfuranosides were identified in this work as major products from the oxygen-alkali degradation of MBR and MBX. Methyl C-carboxy-
furanosides or analogous products have been identified several times in oxygen-alkali reactions of ring carbohydrates (1-3,15). The mass spectra of
the three furanosides were essentially identical and a representative spectrum is presented in Fig. 14. The important diagnostic peaks occur at m/e 394 A
(M), m/e 379 (M-CH3), m/e 334 (M-COCH30-CH 3 ), and m/e 291 (M-CH20TMS).
fragmentation pattern for the major ions is given in Appendix III, Fig. 31o
Following the convention of Ericsson (15) and Malinen (3), products and O were identified as isomeric methyl 2-C-carboxy-S-D-tetrafuranosides
and product
(
was identified as a methyl 3-C-carboxy-B-D-tetrafuranoside.
However, neither Ericsson nor Malinen provided definitive data for the assignment of the 3-C-carboxyfuranoside. Furthermore, although Ericsson (15) formed from the oxygen-
adequately characterized the 2-C-carboxyfuranosides
alkali degradation of methyl S-D-glucopyranoside, Malinen (3) in his study of other methyl glycosides simply assumed that the same relative retention
times were applicable.
and
To eliminate this ambiguity, products
(
I
,
,
were collected by preparative GLC (see Experimental section) and
Fourier transform-proton magnetic resonance (PMR) spectra were taken (Fig. 15 and 16). correct. These spectra demonstrate that the above assignments are
The PMR spectrum of product
shows the anomeric proton as a
-37-
0
H4.)
Ca I
o~~~~t
CS~~~~~~~~~~~
0u 0 0 u~~~~~~~~~~~~~~~~,~
.
A1 I S NEUN -. 3 I 3 A I IV
-38-
-o
E 0
I
a)2 10
.4J
cmJ
C)
0 0
01
co -I 4j U,
0
4~i
4i4
4-,I
-4
ow '0Q
-I ea
. 11
o(JId
r4I
. C-
-39-
-o
0 0
e4
0 00
I~
m *N~
0
44 .0
0)
oi
c'4
4I
0
~4
-U
C4-'
-4 'a) C) ."o
U
-I 04 0
- r-
54
I'dI
-40-
doublet centered at 4.74 ppm whereas the PMR spectrum of product the anomeric proton as a singlet at 4.95 ppm. proton in product (
)
shows
The splitting of the anomeric Product
is due to the ring proton at C-2.
(
does Under
not have a ring proton at C-2 and therefore no splitting was observed. the conditions employed, resolution of products @ and O
was difficult,
but a sample was collected which should have consisted mostly of product This sample gave the same spectrum as shown in Fig. 16.
Q.
The distribution with time of the methyl C-carboxyfuranosides was examined by GLC for both the MBR and MBX degradations. approximation in that molar response factors for Q to n-propyl 0-D-xylopyranoside were estimated at 0.9. upon calculated response factors for MBX and MBR. The analysis is an , and O relative
, Q
This estimate was based
The trends, therefore, in
concentration are accurate but the absolute concentration may not beo
Plots of products
Q,
,
and Qagainst percent reacted glycoside are was the most abundant one formed from the formation of Q and (was en-
shown in Fig. 17 and 18. both glycosides.
Product @
However, relative to.
hanced in the methyl riboside degradations. products show degradation, with product at the reaction conditions.
At long reaction times all three evidencing the most instability
This difference in carboxyfuranoside formation
between MBR and MBX parallels the observed formation of anhydrotetritol acids from the 1,5-anhydroalditols (1).
OXYGEN-DEFICIENT DEGRADATIONS
Due to problems with the air-driven magnetic stirrer, some degradations were conducted at a condition of very limited agitation. reactions were diffusion controlled. As a result these
Because oxygen diffusion is a frequently
cited problem in oxygen-alkali pulpingand bleaching, the diffusion controlled
-41-
IOr0 Product
.. Product X Product
2-C- Carboxyfuranoside 3- C- Carboxyfuranoside
81
51 0 ml
-o
60^
0
~
0
z
0 0/0
0: 4-
z
0
o
U
2H
P'
.,.X- I
0
20 PERCENT
I. I 40 60 REACTION - METHYL
I
I
80 XYLOSIDE
100
Figure 17.
Formation of Methyl C-Carboxyfuranosides in the Degradation of 0.1M Methyl Xyloside in 1.25M NaOH at 120C and 0.682 MPa 02
-42-
Product D Product X Product (5
0
o) 2-C- Carboxyfuronoside
3-C- Carboxyfuronoside
0
10
-
8
0
l21 to 6
'
z
0 P-
z z
w
U
4
w
0 U
2
I -
0
I
I
I
I
0
20 PERCENT
40 60 80 REACTION-METHYL RIBOSIDE
I 100
Figure 18.
Formation of Methyl C-Carboxyfuranosides in the Degradation of 0.1M Methyl Riboside in 1.25M NaOH at 120C and 0.682 MPa 02
-43-
reactions were analyzed to examine the possibility of shifts in the reaction mechanism.
Figure 19 shows that the degradation rate of both glycosides under these conditions* was much slower than the kinetic runs (Fig. 4 and 5). The relaVery
tive rates of degradation between MBR and MBX also shifts significantly.
long induction periods were noted, while hydrogen peroxide levels were rarely significant. Organic peroxides in significant concentrations were measured;
but the values tended to be spurious, probably due to stratification from poor stirring.
As noted for the kinetic runs, methanol formation followed the same pattern as the degradation curves (Fig. 19). Likewise, glycosidic bond cleav-
age accounted for ca. 60% of the degraded glycoside (see Appendix II, Experimental Data). This result corresponds to the percentage of glycosidic bond Finally, the formation of
cleavage found in the well stirred, kinetic runs.
the methyl C-carboxyfuranosides was stereoselective, similar to the kinetic runs. In Fig. 20, chromatograms of the-methyl C-carboxyfuranosides from the
oxygen deficient runs and the kinetic runs are compared at approximately the same percent reaction. Note that the formation of products Mand
Q
is en-
hanced in both the MBR kinetic run and in the MBR oxygen-deficient run. Similarly; the other major products were qualitatively the same in relative Therefore, it is concluded that an oxygen-deficient condition
concentration.
has no effect on the reaction mechanism other than the expected slowing down of the degradation rate.
*Due to an error these runs were conducted at a slightly lower oxygen partial pressure (0.618 MPa) than the kinetic runs.
-44-.
0
co.
0
/
4~~j
0
0e'J
c co
~00
cd0
fr1~ 4
x
1%
o~~~~~~c
x
7
ij
x) /
/
I
s-.
o
~00)
C, 0 H- 0"J
04J0
H00J
Ocd
I L
I
'0 0
~I
I~
0t 0
I
I
I___
0
60~~ 0
c
~6
6'
jet I/ Sa low
N 0
0
a,
0) -H
6;
'NOIIVH1N33NOD
~44
-45-
-,
%O 0 0
0~
0
C:
wnwwm r-
0 4-I .00
%'0 i ---
00
4-4
N.
0C
0
Q toc
'I
"NM=ffmm=
tr"
0-r 0 C.)Q
l m
X -,
10 0;
%O u-i
0N
-46-
DISCUSSION
INTRODUCTION
The underlying foundation for this work was the study by Millard (1,2) of the oxygen-alkali degradation of 1,5-anhydroxylitol (AX) and 1,5-anhydroribitol (AR). By examining the oxygen-alkali reaction of methyl g-D-xylo-
pyranoside (MBX) and methyl B-D-ribopyranoside (MBR) information was gained as to the impact of the glycosidic bond on the degradation of a pyranoid ring in oxygen-alkali. As a general conclusion, the glycosidic bond was
determined to have only a minor effect on the degradation rate and products of model ring carbohydrates having the "ribo" and "xylo" hydroxyl configuration. Therefore, the mechanism proposed by Millard (1,2) for 1,5-anhydro-
alditol degradation in oxygen-alkali was applied to the methyl glycosides of this study. The minor differences between the 1,5-anhydroalditols and methyl
glycosides and their kinetics arise from the methyl aglycon and the effect it has on the termination reactions detailed in Millard's mechanism. effect will be discussed in this section. Also discussed are 1) the This
dramatic effect that C-3 hydroxyl orientation has on degradation rate and the relationship between this effect and glycoside acidity, and 2) the distribution of the methyl aglycon among the major degradation products and the implications of this distribution.
MECHANISM
MECHANISM PROPOSED BY MILLARD (1,2)
Because no major differences were encountered between the degradation rates and reaction pathways (excluding organic peroxide formation) of the 1,5-anhydroalditols and the methyl glycosides, the mechanism proposed by
-47-
Millard (1,2) for the oxygen-alkali degradation of the 1,5-anhydroalditols was concluded to be equally applicable to the oxygen-alkali degradation of the methyl glycosides. A summary of this mechanism is presented to provide
background for the explanation as to why the methyl glycosides exhibited different kinetic orders and organic peroxide formation trends from the 1,5-anhydroalditols.
Equations (5)-(27) represent the reactions used in Millard's derivation of a rate expression for the oxygen-alkali degradation of AR. Several
of the reactions are depicted more specifically in Fig. 21, along with the symbols used in Equations (5)-(27). K A + OH %
v
A
+ H20
(5)
k
A- + 02
6k7
__%
A
+ HOO'
(6)
HOO0 + OH
027 + H 20
(7)
k_7 k8 02 + H2 0 -) HO. + HOO (8)
+ M+ n+ l HOO HOO-M^^k9V +
HOO. + M+
(9)
k AX + H20 10 kA k_10 A + OH(10)
A + HO.
k11 k
A. + HO
(11)
(12)
A
+ HO- -----A 02 + + 02 = -
A
+ H20
A
O13 rlA + HOO0 A
(13)
-48-
k14
A- + 02 k 15
AO2
(14)
A' + 02 AO2 ' + OH
k
16
AO2' AO * + HO
(15) (16)
A AO'
M+n
k17 -17 ' k H-M -18
AO A02
+ + n+ l + M+nM
(17)
A00 +M AO2 + H20
AOOH + OH
(18)
AOAO 2 2
k19 -19
HOZ + HOO>
(19) (20)
k AOOH + OH
k20 k_20 k21 -
AOOH
+ H20
AOOH + OH
\
AOOH + H20
(21)
k21
k22 2 k k
22
AOOH-
' Z + HOO
(22)
AOOH --
2> k24 24
products
(23)
AOOH z ---
products products
(24) (25) (26)
k
k26
02
+
M
l k27
2 + M+ OH- + M+ +
HO. +M+n
where M = catalytic meta 1 ion.
(27)
-49K5
OH H20
s>
A
kll HO' 1
Oz 02 -HOO
k-io
/-HOO
OH
H20
_
HO
Ak
OH
klo
Akl4 o2
05|2
k 16
OHO
H2 0
k-16
AO,
M+ n+ I M+ n
AO-
*f
k-21 H2 0
< OH
+ HOO
,0 OH -AOOH
k 23
k21
z
k25
ACIDIC PRODUCTS AOOHFigure 21.
" ACIDIC
PRODUCTS
Proposed Mechanism of Degradation of 1,5-Anhydroribitol by Molecular Oxygen in Alkali (10)
-50-
This mechanism is unique in that it incorporates a hydroperoxide as an intermediate in the degradation of a carbohydrate by molecular oxygen in alkali. Furthermore, the mechanism postulates the relationship between the
hydroperoxide and the keto derivative proposed by many investigators as the precursor to acidic products Q(-5,15,18,21). Specifically, the carbonyl is
formed from an ionized alpha-hydroxyhydroperoxide [Equations (19) and (22)]. Acidic degradation products can also be produced directly from the ionized alpha-hydroxyhydroperoxide as proposed by Isbell (45) (Fig. 22).
HO-O-C-OH r, _-C kJ <1 -C / ~,
+ +
-C, -C
O
+ OH H
O
"-H-C-O
OH
HOO-C-OH H-C-OH
I
+OH
I
I
HO
C-O
H -C-0-H
, -Ce0
+
-C
H
+ H20
Figure 22.
Acidic Degradation Products Formed Directly from Ionized Alpha-hydroxyhydroperoxide as Proposed by Isbell (45)
By applying the steady-state approximation that all radical species are very reactive and therefore present in very small concentrations and that metal ion concentration is also small, Millard (Q,2) arrived at the following kinetic expression:
-d[A]/dt = K5 k6 [A][OH-][02] + K52k6k3[A]2 [OH-][O2]/k26[tMn1] + K5 k6 ks H20][A] [OH] [02/k26[M+n+l 8 The third term involves reactions associated with the hydroxyl radical. By (28)
ignoring the reactions involving the hydroxyl radical, a kinetic expression
-51-
was derived which contained the same first two terms but the third term was absent. Millard (1,2) concluded that the hydroxyl radical does not play a This conclusion
major role in the mechanism under the conditions studied.
was consistent with the postulate that in minimum-metal-catalyzed systems the perhydroxyl (HOO') and the superoxide (027) radicals are more important than the hydroxyl radical (46,47). nate at the reaction conditions. The second term was concluded to domiThe appearance of metal concentration in
the denominator causes this term to be large at small metal concentrations. The derived expression breaks down at zero metal concentration and alternative reactions would probably have to be considered. The assumption that
the second term predominates (and thus yielding second-order kinetics) was consistent with the data obtained from the oxygen-alkali degradation of AR (1,2). For a more detailed discussion of this mechanism and its derivation
the reader is referred to Millard's thesis (1).
AUTOINHIBITION MECHANISM PROPOSED BY MILLARD (1,2)
In the above mechanism, Equations (17), Equation (9) are chain termination reactions.
(26), (27) and the reverse of In addition to these reactions, These
Equations (29) and (30) can be written as chain termination reactions.
reactions are directly analogous to reactions frequently proposed in hydrocarbon autoxidations (22).
AO0 + R 2 2A0 2
--
AOOR Z + 02 + AOOH
(29) (30)
where Z, AOz , and AOOH are specified in Fig. 21. 2
Millard proposed these two
reactions as probable chain termination reactions in the oxygen-alkali degradation of AX. system. It was also argued that Reaction (30) predominated in the AX
-52-
In the degradation of AX, Millard detected an intermediate organic peroxide which he deduced to be an alpha-hydroxyhydroperoxide (see AOOH in Fig. 21). He further postulated that in the AX system this alpha-hydroxyThis
hydroperoxide could be stabilized via intramolecular hydrogen bonding.
effectively increased the concentration of the radical A0 2 and the significance of termination Reactions (29) and (30). These termination reactions
produce only nonradical species resulting in a decrease in carbohydrate degradation. Thus, stabilized A0 2 was a postulated intermediate which Reactions
serves to inhibit the radical degradation of AX in oxygen-alkali;
that are slowed down by a mechanism-interfering intermediate are termed autoinhibited (37). A characteristic of autoinhibited reactions is that Millard's analy-
they exhibit time dependent orders higher than true order.
sis indicated this to be likely in the oxygen-alkali degradation of AX.
AUTOINHIBITION IN METHYL GLYCOSIDE DEGRADATIONS
GENERAL
The time dependent orders for MBR and MBX were determined to be ca. 2.75 and 3.5, respectively. If, as previously stated, it is assumed that
MBR and MBX degrade via the mechanism proposed by Millard (1,2) then these kinetic orders suggest that the oxygen-alkali degradation of methyl glycosides is autoinhibited. The oxygen-alkali degradation of ring carbohydrates
has generally been reported to be second order in carbohydrate concentration (1,2,23,31,42), although subsequent reanalysis of the data indicated higher time-dependent orders (1,2). Therefore, kinetic orders higher than two
might indicate the reactions to be autoinhibited.
In comparing the oxygen-alkali degradation of the 1,5-anhydroalditols with the oxygen-alkalidegradation of the methyl glycosides, organic peroxide
-53-
formation is seen to differ markedly in the two systems.
Stable organic per-
oxides were produced from the methyl glycosides, suggesting the formation of dialkyl peroxides (12-14). In the degradation of AX, the intermediate organic
peroxide was concluded to be an alpha-hydroxyhydroperoxide and stable organic peroxides were detected only in minor concentrations. From this it is concluded
that autoinhibition in the methyl glycoside degradations results from an increased probability of Reaction C29) occurring.
If the stable organic peroxides in this work are indeed the result of autoinhibition, then the kinetic orders predict that autoinhibition is more predominant in the MBX system. Such a situation should lead to a higher This is appar-
concentration of organic peroxides from the MBX degradation. ently the case.
If the yield of organic peroxides is plotted against percent
reaction (Fig. 23), MBX is seen to produce a much higher concentration of organic peroxides at equivalent percent reaction than does MBR.
In autoxidations, Equation (29) is generally considered to be important only at low oxygen pressures and when the intermediate alkyl radical (in this case R') is stable (22). The following discussion will be addressed to these
two points in relation to the oxygen-alkali degradation of methyl glycosides.
STABLE INTERMEDIATE ALKYL RADICAL
A frequently proposed radical in the oxidation of glycosides is the C-1 radical resulting from anomeric hydrogen abstraction (19,27,48). The stability
+OCH 3
CHCH
HO O HO OH
HO HO OH
HO_ HO OH
-54-
D -5MBR o-7MBX 2112
a'z0 0
a 0/
/ /
8-0
0
0
/
DO
0
,4II
I
I I
60
20 40 REACTION, %
Figure 23.
Organic Peroxide Concentration as Function of Percent Reaction in the Degradation of 0.1M Methyl Glycosides in 1.25M NaOH at 120C and 0.682 MPa 02
of this radical is enhanced by the adjacent oxygens and their ability to contribute to resonance. Although it will later be argued that formation and decompo-
sition of this radical does not appear to be a major pathway in the degradation of the methyl glycosides (see Fate of Glycosidic Bond), anomeric hydrogen abstraction does provide the necessary stabilized radical for Equation (29). Thus, the
C-l radical is believed to be produced in sufficient concentrations to provide for autoinhibition of the methyl glycoside degradations. as shown in Fig. 24 (28). highly reactive radical. The C-l radical can decompose
Decomposition (Path A) can yield the methyl radical, a It is doubtful that the methyl radical enters into terFurthermore, no products have
mination reactions owing to its extreme reactivity.
-55-
ever been identified from the oxygen-alkali degradation of glycosides which would correspond to C-1 radical decomposition via Path A (3,15,48). The primary radical
formed from C-1 radical decomposition via Path B would probably react quickly with molecular oxygen to form a hydroperoxyl radical (-CH200'). Similarly, the C-l
radical could react with oxygen to form a tertiary hydroperoxyl radical (R(R')COO.).
-
O
A-CH3
HO 0 B
/
0 CH3. + 110 HO Figure 24.
9_HO
\
Gi 0 00/CH2OCH3 =
OH
HO
OH
Possible C-1 Radical Decomposition Pathways (28)
Reaction (30) is commonly proposed to proceed through a tetroxide intermediate as illustrated for an analogous reaction between the alpha-hydroxyhydroperoxide radical (A02 ') and the C-1 tertiary hydroperoxyl radical (22). This
reaction yields oxygen, a keto glycoside (Z), and a tertiary hydroperoxyorthoester.
J0t*kO
Ox
OH OH
xb,
/OA
WH /
-O
0
V
C
uVn
CH 3
+J?
0 - CI\
(Z)
N\
a.,I"u
OH OH
-56-
The same reaction with the above primary hydroperoxy radical (-CH200') yields a primary hydroperoxide instead. Reaction (29) between the C-l radical and A02 ' The formation of peroxyorthoesters would
would form a dialkyl peroxyorthoester.
CH3 (
correspond to the stability of the detected organic peroxide in alkaline solution. The question is whether or not peroxyorthoesters would exhibit the same resistance to acid hydrolysis that the detected organic peroxides did. et al. (49) suggests that they might have this stability. The work of Seyfarth, The UV irradiated,
oxygen reaction of 2-methyl 1,3-dioxolane yielded 1,3-dioxolan-2-yl hydroperoxide. This hydroperoxide required rather severe conditions (iN H 2 S04, 20C, 2 hr) to
[I
RI
uv/0
2
----
0
yield hydrogen peroxide.
OOH
The reaction of this hydroperoxide with hydrogen
peroxide yielded bis-(1,3-dioxolan-2-yl) peroxide which was stable to vacuum distillation at 80C. Thus, it is postulated that autoinhibition in the methyl
-OR
+ H02 2 -
-(\OOH
ro/vRX LO\R OJ
-57-
glycosides occurs due to the C-1 radical entering into a termination reaction with AO2' to yield peroxydiorthoester.
reactions of AO2
The formation of peroxyorthoesters in
with R(R')COO. and of hydroperoxides in reactions of A02' These peroxyorthoesters and
with -CH200
can be postulated as shown above.
primary hydroperoxides, however, would not be expected to be as difficult to hydrolyze as the detected organic peroxides were.
So far the methyl glycosides have been unique among the carbohydrates in their.ability to form stable organic peroxides during oxygen-alkali degradation. Organic peroxides similar to the ones formed from MBR and MBX were
formed from the oxygen-alkali degradation of methyl S-D-glucopyranoside and methyl a-D-glucopyranoside (13,50) while glucitol (17) and mannitol (51) yield hydrogen peroxide but no organic peroxide in oxygen-alkali. The 1,5-
anhydroalditols generated hydrogen peroxide and, in the case of AX, an intermediate organic peroxide but not significant concentrations of stable organic peroxides when reacted with oxygen in an alkaline medium (1,2). Cellobiitol,
a glycoside, generated hydrogen peroxide readily in oxygen-alkali but does not yield any stable organic peroxides. Presumably, in the case of cellobi-
itol two factors operate against the envisioned formation of dialkyl peroxides. First, the glucitol aglycon offers more sites for hydrogen abstraction and thus decreases the frequency of anomeric hydrogen abstraction. Secondly, the
glucitol aglycon would provide steric hindrance to radical combination.
EFFECT OF OXYGEN PRESSURE
As previously mentioned, Reaction (29) is generally considered to be important only at low oxygen pressures (22). Although this point was not
examined in this work, some data of Thompson, et al. illustrate the effect that oxygen pressure has on organic peroxide formation from the oxygenalkali degradation of methyl B-D-glucopyranoside (14). Figure 25 illustrates
-58-
2.5
A[0-
0.500 MPa 02 0.682 MPa 02 0.864 MPa 02
2.0
1.5
0
0 z 0
0.5
0 48 TIME,
Figure 25.
10
HR
Formation of Organic Peroxides in the Degradation of 0.03M Methyl B-D-Glucopyranoside in 1.25M NaOH at 120C. Effect of Varied Oxygen Pressure (14)
-59-
peroxide formation at three oxygen pressures.
In increasing oxygen pressure
from 0.5 MPa to 0.864 MPa, organic peroxide formation passes through a maximum. Although these oxygen pressures may not necessarily be considered low,
the oxygen pressure only serves to determine the solution oxygen concentration. It is the solution oxygen concentration which ultimately determines Therefore, it is con-
whether or not R' is efficiently scavenged by oxygen.
cluded that the solution oxygen concentrations were low enough to allow Reaction (29).
HYDROXYL ORIENTATION RATE EFFECT
GENERAL
Varied C-3 hydroxyl orientation can dramatically affect the rate of degradation of pyranoid rings in oxygen-alkali. This observation has now
been made for both the 1,5-anhydroalditols (1,2) and the methyl glycosides. The C-2 epimeric methyl glycosides have also been reported to vary in degradation rate in oxygen-alkali (3). Malinen (3) has suggested that methyl
glycosides having cis 1,2-diol configurations can more easily form the frequently proposed alpha-dicarbonyl intermediates. Millard (1,2) did not
offer an explanation for the degradation rate differences between AX and AR. Caution must be observed in interpreting the autoinhibition noted by Millard for the AX degradation. Although the AX degradation may have been slower.
than it would have been in the absence of autoinhibition, this doesnot preclude different initial rate constants for the respective degradations of AX and AR. The hypothesis presented here is that the degradation rate
differences arising in the C-3 epimers of the 1,5-anhydroalditols and the methyl glycosides result from an increase in hydroxyl ionization when neighboring hydroxyls are in a cis configuration. The increase in ioniza-
tion is believed to be a function of stabilization via intramolecular hydrogen
-60-
bonding.
The rate degradation increases with increasing oxyanion concentra-
tion due to the facilitated hydrogen abstraction to form the oxyanion radical (1,2,23,42).
FACTORS AFFECTING CARBOHYDRATE ACIDITY
The ionization constants for carbohydrates generally are in the range of 10- 12 to 10 - 14 and are larger than the ionization constants for monohydric aliphatic alcohols (ca. 10 16) (52). The reducing sugars tend to be the most
acidic while straight chain alditols, cyclitols, and glycosides of similar molecular weight and hydroxyl content have approximately the same acidity (52). Among similar carbohydrate molecules, however, acidities can vary For example, methyl
appreciably with slight structural changes (53,54).
a- and B-D-glucopyranoside can be separated on a strongly basic ion-exchange resin (53). In the case of the g-anomer, the dipole moments of the ring and
glycosidic oxygens presumably reinforce each other and activate the C-2 hydroxyl to ionization. The a-anomer, however, has the glycosidic oxygen
oriented such that it partially cancels the dipole of the ring oxygen thereby diminishing the activation of the C-2 hydroxyl. This example also demon-
strates that within a polyhydric molecule the acidity of a specific hydroxyl varies with its environment. Equatorial hydroxyls are considered to be more Upon ionization, the resulting equatorial
acidic than axial hydroxyls (55).
anion, being less hindered than the axial epimer, is more easily solvated and is therefore more strongly basic. Intramolecular hydrogen bonding can If the resultant oxy-
also affect the acidity of hydroxyls (52,54,56,57).
anion is stabilized by hydrogen bonding, then ionization of the hydroxyl group becomes more favorable. The acidity of the hydrogen-donating hydroxyl
is probably lowered due to its bonding with anionic oxygen (52).
-61-
METHYL RIBOSIDE AND METHYL XYLOSIDE ACIDITY
The C-2 hydroxyl ionizes most easily in the methyl glycosides due to ring oxygen activation (24-26). The acidity of the C-2 hydroxyl can vary
between a- and 0-anomers as mentioned above (53), but because MBR and MBX are both a-anomers the effect of ring and glycosidic oxygens should be equivalent in the two glycosides. Evidence in this work, however, has been
obtained which indicates that MBR is more acidic than MBX.
In determining a GLC response factor for MBR, the response factor was found to decrease at low MBR concentrations. A series of experiments
examining various steps in the sample work-up procedure indicated that MBR was being selectively absorbed on the ion-exchange resin used in the workup*. The ability of strongly basic resins to ionize and retain methyl When the MBX
glycosides has been demonstrated by Neuberger and Wilson (53).
response factor was determined, the response factor was constant over the entire range of concentrations examined.
Further evidence supporting the hypothesis that MBR is more acidic than MBX was obtained via
13
C-NMR.
Spectra were taken of MBR and MBX in both D 20 It was postulated that if MBR is more acidic
and D 2 0/NaOD (ca. 0.5M NaOD).
than MBX, then MBR should show greater 13C peak shifts in alkali than does MBX. Examination of the data in Tables IV and V demonstrates that MBR in
alkali exhibits significant shifts in its l3 C spectra, while MBX shows only very slight shifts. The assignment of the peaks corresponding to C-l, C-5,
and C-6 was made from the off resonance spectra and the expected splitting from the protons that these spectra exhibited. The C-2 peak of MBR and MBX
*The resin used was Amberlite MB-3, consisting of a mixed bed of IR-120 cation resin and IRA-410, a strongly basic anion exchange resin.
-62-
TABLE IV
13
C PEAK SHIFTS OF METHYL XYLOSIDE IN ALKALIa
Carbon C-1 C-3 C-2 C-4 C-5 C-6
MBX/D2 0, ppmb 106.9 78.8 76,0 72.3 68.2 60.3
MBX/D20ONaOD,
Shift ppm ppm 0.1 0.1 0.1 0.0 0.1 0.0
107.0 78.9 76.1 72.3 68.3 60.3
Carbohydrate concentrations = 100 mg/0.5 ml D 2O; alkali concentration = 0.59M NaODo bRelative to 3-(trimethylsilyl)propane sulfonic acid. CPositive shift means downfield shift.
TABLE V
1
3C PEAK SHIFTS OF METHYL RIBOSIDE IN ALKALI a
Carbon C-1 C-2 C-3 C-4 C-5 C-6
MBR/D2 O, ppm b 104.4 73.2 70.8 70.7 66.1 59.1
MBR/D2O/NaOD, ppmb 104.8 73.6 71.3 70.8 66.5 58.9
Shift, ppm b 0.4 0.4 0.5 0.1 0.4 -0.2
Carbohydrate concentration = 100 mg/0.5 ml D 2O; alkali concentration = 0.53M NaOD. bRelative to 3-(trimethylsilyl)propane sulfonic acid. Positive shift means downfield shift,
-63-
assignment was based upon selective decoupling experiments.
All of the above To ex-
assignments are based on those reported by Bock and Pederson (58).
clude the possibility of the observed 13C shifts being a function of a salt effect, a
13
C spectrum of MBR in 0.5M NaCl/D20 was obtained.
This spectrum
had no significant differences between it and the spectrum of MBR in just D 20.
STABILIZATION OF METHYL GLYCOSIDE OXYANIONS
As indicated in the above discussion, MBR is apparently more acidic than is MBX. In view of the factors affecting carbohydrate acidity, intramolecular
hydrogen bonding would seem to be the most plausible contributor to this difference in acidity. The ring oxygen and the glycosidic oxygen have equivalent
effects upon the C-2 hydroxyl in both glycosides.
In order to discuss the difference in acidity between MBR and MBX, it is first necessary to develop an understanding of the conformational equilibria of these two glycosides in solution. An estimate of the percentage of each
conformer present in solutions of MBR and MBX can be made from the data of Angyal (59) and application of this data to Equation (31).
A GO.= RT ln (Ni/N 2) A G = difference in conformational free energies of 4C1 and 1 C4 conformers R T = gas constant = absolute temperature
4
(31)
where
N1,N 2 = mole fraction of
CI and 'C conformers, respectively 4
It was assumed that A G would be approximately the same at 120C as it is at 25 . With this assumption MBX was found to have a conformational equilib4
rium of 0.95:0.05;
C: 'C
4
at 120C.
At this same temperature MBR was deterLiterature
mined to have a conformational equilibrium of 0.68:0.32; 4C 1 : C4 .
-64data indicate that the
1
C4 MBR conformer may be even more predominant than Therefore, in considering the solution However, MBX can be
this calculation indicates (58).
reactivity of MBR, both conformers must be examined. assumed to exist mostly in the interconversions
4
C1 conformer.
Moreover, the frequency of
between the two chair conformers must be greater for MBR.
A principal factor determining hydrogen bond energy is the oxygen-oxygen (0---0) distance (60). The optimum 0---0 distance varies somewhat depending is 2.75 0.13 A
on the molecular environment but the average for O-H---O (60).
An approximation of the various 0---0 distances in MBR and MBX was These distances are
obtained by measurement from Dreiding molecular models. presented in Table VI.
Also presented in the table are 0---0 distances
(indicated in parentheses) obtained from x-ray crystallographic data on MBR (61) and MBX (62). The distances were calculated through the use of the No clear distinction between the favorability
computer program CARTSET (63).
of hydrogen bond formation in MBR versus MBX can be made from the data presented in Table VI.
TABLE VI OXYGEN-OXYGEN DISTANCES IN METHYL RIBOSIDE (MBR) AND METHYL XYLOSIDE (MBX) (0-2)---(0-3), A MBX 2.9 a'b (2.81)a (0-3)---(0-4), A 2.9 a b ' (2.91) ad 2. 7 a 'b 'c d (2.87) c ' (0-2)---(0-4), A 2.7b c
'd
MBR
' 2.7ab, d (2.87) c '
2.7 b c (2.77)c 'd
Distance determined for
4
C 1 conformer.
Distance determined by measurement from Dreiding molecular models. CDistance determined for
1
C 4 conformer.
dDistance determined by CARTSET program (63) from x-ray crystallographic data on MBR (61) and MBX (62).
-65-
Another factor affecting the probability of hydrogen bond formation between neighboring hydroxyls of MBR and MBX involves the spatial orientation of the hydroxyls to the plane of the ring. The dihedral angle between Yet in
neighboring hydroxyls of MBR and MBX are approximately the same.
MBX neighboring hydroxyls are located trans to the ring plane formed by C-2, C-3, C-5 and the ring oxygen. cis to the ring plane. In MBR the neighboring hydroxyls are oriented
In forming a hydrogen bond the O-H---O system shows The result of this
appreciable contraction of interatomic distances (60).
contraction is to reduce the dihedral angle between the neighboring hydroxyls. With MBX this contraction causes increased puckering of the ring and an increase in the diaxial interaction between hydrogens. In the case of MBR,
the ring tends to flatten and no increase in diaxial interactions is observed. An example of this effect was reported by Kuhn (64) in a study of intramolecular hydrogen bonding of 1,2-diols. cis-Cyclohexane-1,2-diol was
observed to have more intramolecular hydrogen bonding in solution than did trans-cyclohexane-l,2-diol and Kuhn applied the above discussion as an interpretation of this result. The larger frequency of conversions between hydrogen bonds.
MBR chair conformers also assists the formation of O-H---O
The conversion between the 4C1 and 1C4 conformers involves a number of possible boat, twist-boat, and half-chair intermediate conformers (65). In several
of these intermediate conformers of MBR, the dihedral angle between neighboring hydroxyls is reduced. Thus, the contraction in hydrogen bond formation
described above is partially accomodated by the conformational equilibrium of MBR.
INTERRELATIONSHIP OF OXYANION AND MECHANISM
To understand the relationship between increased oxyanion concentration and the increased degradation rate, the reader is referred to the previously
-66-
discussed mechanism.
The proposed mechanism of degradation initiation is via
hydrogen abstraction by molecular oxygen from the oxyanion [Equation (6)]. By increasing the concentration of oxyanions, the frequency of initiation reactions is increased. Hydrogen peroxide is generally assumed to be a prodExperimentally, MBR was observed to form Additionally, the in-
uct of this initiation (1,2,12-14).
much more hydrogen peroxide than did MBX (Fig. 11).
crease in the number of initiation reactions increases the concentration of perhydroxyl radicals. In alkali the perhydroxyl radical ionizes to the In the proposed mechanism (1,2), the superoxide Thus by
superoxide radical (66).
radical also abstracts hydrogen from the oxyanion [Equation (13)].
increasing the superoxide radical concentration the degradation is accelerated by an increase in radical chain propagating reactions.
FATE OF THE GLYCOSIDIC BOND
GENERAL
As noted in the Literature Review and shown in the Results section, the oxygen-alkali degradation of methyl glycosides can occur without cleavage of the methyl aglycon by forming so-called "bound methanol" products. Glycosidic
bond cleavage has been proposed to occur via $-alkoxy elimination from a C-3 keto intermediate (4,7,14,15,17). The formation of the "bound methanol"
products has generally' been proposed to occur via a dicarbonyl intermediate (3,15), although the recent data of Millard (1) suggested alternative intermediates. This section will attempt to place in perspective the various reac-
tion products containing the methoxyl group and the information that these products yield concerning the mechanism of oxygen-alkali degradation of methyl glycosides.
-67-
IMPLICATIONS OF METHANOL FORMATION
The single most abundant product from the oxygen-alkali degradation of MBR and MBX was methanol. In analyzing the methanol data it is concluded This con-
that a C-3 keto is the intermediate to glycosidic bond cleavage.
clusion is derived from the fact that both MBR and MBX produce methanol in approximately 60% yield, yet MBR degrades at an initial rate six times the initial rate of MBX. Since MBR and MBX are C-3 epimers, the C-3 keto deriva-
tive of these methyl glycosides would be equivalent and should yield equivalent amounts of methanol upon S-alkoxy elimination. However, a qualifica-
tion must be made to this discussion as the C-3 keto derivative is not necessarily formed in equal yields from MBR and MBX. Some methanol liberation Yet the linear
probably arises from the degradation of secondary products.
relationship between percent reaction and methanol yield indicates that the majority of the methanol liberation is probably via g-alkoxy elimination from a keto derivative of the original glycoside.
A second implication of the methanol data involves the previously discussed C-l radical. It had been originally hypothesized that differences in
the degradation patterns of the 1,5-anhydroalditols and the methyl glycosides could arise from the increased probability of C-l radical formation from the methyl glycosides. However, the overall degradation rates and the major
reaction pathways of the methyl glycosides were generally equivalent or analogous to the degradation rates and pathways of the corresponding 1,5-anhydroalditols. Additionally, no major reaction products were identified that This observation
would be indicative of C-l radical decomposition (Fig. 24).
was also made by Kano, et al. in a study of the oxygen-alkali degradation of cellobiitol (48). Furthermore, although C-l radical formation could lead to
equivalent methanol yields, MBR would not necessarily be expected to degrade
-68-
faster than MBX.
Hayday and McKelvey (67) have reported that photochemically
induced hydrogen abstraction from 2-methoxy-4-methyltetrahydropyran is preferential to abstraction of an axial "anomeric" hydrogen. The previously discuss-
ed conformational analysis indicated that MBX exists predominantly in the conformer (4 C 1) having the anomeric hydrogen axially oriented, whereas MBR exists in this conformer to a lesser degree. Assuming the effect noted by
McKelvey and Hayday to be applicable, then MBX would be predicted to react faster than MBR if C-1 radical decomposition represented a major pathway. In
summary, the C-l radical is not believed to play a major role in methyl glycoside degradation but, as previously discussed (Autoinhibition in Methyl Glycosides), it is believed to be produced in sufficient concentration to cause autoinhibition in the degradations.
STEREOSELECTIVE FORMATION OF METHYL C-CARBOXYFURANOSIDES
Three methyl C-carboxyfuranosides were identified as major products of the oxygen-alkali degradation of MBR and MBX. Two of these acids (products
Q
and
O
) were determined to be isomeric methyl 2-C-carboxy-3-D-tetrafurano) was identified as a methyl 3-C-carboxy-
sides and the third acid (product O (-D-tetrafuranoside.
Importantly, the relative ratio of formation of these This result is in agreement
acids varied between the MBR and the MBX systems.
with the previously reported formation of the analogous 1,4-anhydro-2-Ccarboxytetritols from AX and AR (1,2). This variation in methyl C-carboxy-
furanoside formation from MBR and MBX reemphasizes the observation of Millard (1) that an alpha-dicarbonyl intermediate (3,15) does not adequately account for the observed formation of these furanoid acids. The alpha-dicarbonyl
intermediates from MBR and MBX would be identical and the relative ratios of the methyl C-carboxyfuranosides would be expected to be the same from both'glycosides.
-69-
A mechanism which accounts for this stereoselective formation has been proposed for the oxygen-alkali degradation of the 1,5-anhydroalditols (1,68) and is adapted to the oxygen-alkali degradation of the methyl glycosides in Fig. 26 and 27. These reactions indicate carbonyl formation at C-2 and lead O and 0. Carbonyl formation at C-4
to the methyl 2-C-carboxyfuranosides
in MBR can account for the formation of methyl 3-C-carboxyfuranoside 0. Formation of a carbonyl at C-3 yields identical intermediates from MBR and MBX thus eliminating any possibility of stereoselective formation of the furanoid acids.
The first step in the proposed mechanism (1,68) involves the reversible addition of hydroxide ion to the carbonyl with the resulting oxyanion being stabilized via hydrogen bonding. The hydroxyl ion adds in a manner such This allows for more
that the oxyanion is cis to the neighboring hydroxyl.
favorable hydrogen bond formation as discussed previously (see Hydroxyl Orientation Rate Effect). Any of various radical species may then abstract This radical is
the C-3 hydrogen forming the hydroalkyl radical at C-3.
stabilized to inversion via hydrogen bonding resulting in the stereoselective addition of oxygen to the radical to form the alpha-hydroxyhydroperoxyl radical. Subsequent hydrogen abstraction by this radical yields an alphaMethyl C-carboxyfuranoside formation can occur by a
hydroxyhydroperoxide.
semibenzilic type mechanism (69,70) from either the above alpha-hydroxyhydroperoxide or its conjugate base. The formation of the carbonyl moiety This
of the carboxylic acid group is concerted with ring contraction.
rearrangement must be from conformations in which the (C-1)-(C-2) bond is antiperiplanar with the carbon-oxygen bond of the hydroperoxide group. The conformations allowing for this are
4
C1 for MBR and 1C4 for MBX.
These
conformations also allow for the bulky hydroperoxide group to be equatorial.
-70-
0)
0 -0
P
'44
0
-4J 5.1
0
098 2: 0)
0 0 2:
54 I 'i 0 2: 0 0 ,0
-H
4-i
I
I
I w0)
I
01 2:
0
0
0
14-
0,
x :
0 0
0).
0us0
0
boo
4-H
0 'O.
0
.i
P.4
0 0 x 0'
0)-6 2: 0
-71-
10
~
00
.44
2:
S
0 2:~~~~~~~~~~~~~~~~~~~~~~C
I~~~~~~~~~~~~~~~~~~~~~
s-I
0
0~~~~~~~~~~~~~~
0~~~~~~~~~~~~~~~~~~~
4-1
2: I I I
0~~~~~~~~~~ 2:~~~~~~~~~~~~ I
4-)~~~~~~~~~~~~~~~~~~
0
-
0
0 02: 0 0~~~~~~~~~~~~~ 0 0
~~~~~~U0$020~~~~~~~~~0$ 0 0)~~~~~~~~~3 001 En-
o
2: 0
0
~~~~~~~~~~~~~~
o
C-
00
0~~~~~~~~~~~~~~~~6
00 00 0,~~~~~~~~~~~~~.
C3
.
Ca 0
~~~~~c
0~~~~~~~~~ 2:
0 ~~~~~~
2:~~~~
i-
01
I-
2:
1
-72-
Specific assignment of isomeric structure of these acids could not be made based upon the mass spectral or PMR data. of Millard (1), product O furanoside and product Similarily, product side. However, following the logic
is believed to be methyl 2-C-carboxy-8-D-erythroto be methyl 3-C-carboxy-8-D-erythrofuranoside.
(
is believed to be methyl 2-C-carboxy-g-D-threofurano-
These assignments are consistent with the observation in this work
and by previous workers (1-3) that ring carbohydrates having cis 1,2-diol configurations degrade faster in oxygen-alkali than do trans 1,2-diols. Products O and ( exhibited less stability at the reaction conditions
than did product
(Fig. 16 and 17)
ALTERNATIVE CHAIN CLEAVAGE MECHANISM
Chain cleavage of polysaccharides such as cellulose is usually considered to involve the breakage of bonds at the glycosyl oxygen, i.e., elimination of either the C-l or C-4 substituent. 8-alkoxy
However, the detection
of methoxyacetic acid in this work indicates that chain cleavage in alkaline oxidative reactions can occur via ring scission. Methoxyacetic acid was
previously reported as a product of the alkaline hydrogen peroxide degradation of methyl 8-D-glucopyranoside (20), but this is the first report of methoxyacetic acid as a product of the oxygen-alkali degradation of a methyl glycoside. The proposed mechanism of methoxyacetic acid formation (20) (Fig. 28) After
involves the ring oxygen as a leaving group in 8-alkoxy elimination.
the elimination reaction, the resulting a,B-unsaturated molecule can ketonize to an alpha-dicarbonyl. Subsequent oxidative cleavage of this dicarbonyl
yields methoxyacetic acid.
Although the above mechanism accounts for methoxyacetic acid formation, it would predict that equivalent amounts of methoxyacetic acid are formed from MBR and MBX. Yet, apparently more methoxyacetic acid is produced from
-73-
/
OCH
00CH g
3
CH 2 O HOOCH3 H20
CH 2 0H
CH 2 0H
CH 2 0CH 3
CH^OCH3< ---- II
II 0
m-
+H
COOH
O
Figure 28.
Proposed Formation of Methoxyacetic Acid According to Weaver (9)
MBX than from MBR.
Moreover, this difference in methoxyacetic acid formation
accounts for the difference in total methyl C-carboxyfuranosides formed in the two systems, while methanol yields from the two glycosides remain, the same. This suggests a commonality in intermediates between the formation of methoxyacetic acid and the formation of the methyl C-carboxyfuranosides. commonality is not as yet understood. However, this
-74-
CONCLUSIONS
The information gained from the degradation of methyl B-D-ribopyranoside (MBR) and methyl O-D-xylopyranoside (MBX) by oxygen in an alkaline medium indicates that the methyl aglycon has little effect on the overall degradation rate and product formation pattern of a pyranoid ring. The configura-
tion of ring hydroxyls, however, is shown to affect dramatically the degradation rate, thus confirming the observations of earlier workers in other systems (1-3). This effect was postulated to be the result of oxyanion stabili-
zation via intramolecular hydrogen bonding with a neighboring hydroxyl in molecules having a cis 1,2-diol configuration, i.e., MBR. evidence was obtained which supported this hypothesis. Experimental
Because of similarities in degradation rates and reaction pathways between the 1,5-anhydroalditols and the methyl glycosides, the mechanism proposed for the oxygen-alkali degradation of the 1,5-anhydroalditols adapted to the methyl glycosides. kinetic orders of 2.0. (1,2) was
This mechanism predicts carbohydrate
The time dependent kinetic orders of MBR and MBX
were 2.75 and 3.5, respectively, indicating the reactions were autoinhibited. The C-l radical was postulated to cause the observed autoinhibition through a radical termination reaction with an alpha-hydroxyhydroperoxyl radical (AO2 '). In addition to the kinetic orders, autoinhibition was indicated
by the formation of stable organic peroxides as would be predicted by a termination reaction between C-1 radical and AO2'o Additionally, the higher
kinetic order for MBX (suggestive of..more.autoinhibition in the MBX system than in the MBR system) corresponded to greater yields of organic peroxides from the MBX degradations.
-75-
The varied hydroxyl stereochemistry did not alter the type of products identified from the degradation of MBR and MBX in oxygen-alkali, although the distribution of the "bound methanol" products was affected by hydroxyl stereochemistry. The relative ratio of the identified methyl C-carboxyfurano-
sides varied between MBR and MBX lending support to the stereoselective mechanism proposed for the 1,5-anhydroalditols (1,68). Methoxyacetic acid
was identified as a product of the oxygen-alkali reaction of MBR and MBX. This product has important implications as to an alternative chain cleavage mechanism of polysaccharides in oxidative, alkaline systems.
The orientation of the C-3 hydroxyl did not affect glycosidic bond cleavage as MBR and MBX yielded equivalent amounts of methanol even though their degradation rates differed markedly. Although the C-l radical was
postulated to play an important role in autoinhibition, the degradation rate, product, and methanol data implied that C-l radical formation and decomposition did not represent a major degradative pathway for the methyl glycosides.
-76-
EXPERIMENTAL
GENERAL ANALYTICAL PROCEDURES
Melting points were determined on a Thomas Hoover capillary melting point apparatus. The melting points were corrected against calibration
based upon known compounds.
Optical rotations were determined on a Perkin-Elmer Model 141 MC recording polarimeter.
Quantitative visible spectroscopy was-done with a Gary Model 15 recording spectrophotometer. source. A tungsten lamp was utilized as the light
Thin-layer chromatography (TLC) was done on microscope slides coated with Silica gel G (Brinkman Instruments, Inc.). as follows: Developing solvents were
solvent 1 - chloroform:ethyl acetate, 19:1, v/v; solvent 2 Detection was accomplished with methanolic sulfuric acid
ethyl acetate.
spray (20%, v/v) followed by charring.
Quantitative gas-liquid chromatography (GLC) was done on a Varian Aerograph 1200 gas chromatograph equipped with a hydrogen flame ionization detector and a Honeywell Electronic 16 recorder with a Disc integrator. Prepurified nitrogen (Matheson Gas Products) was used as the carrier gas. Column descriptions and operating conditions are given in Appendix I.
Mass spectral analyses were conducted on a Du Pont 21-491 mass spectrometer interfaced via a jet separator to a Varian Aerograph 1400 gas chromatograph equipped with a hydrogen flame ionization detector and a Hewlett-Packard 7128A recorder. Helium (UHP helium, minimum purity 99.999%;
-77-
Matheson Gas Products) was used as the carrier gas.
Mass spectra were Chromatographic
recorded with a Century GPO 460 oscillographic recorder.
conditions and mass spectrometer control settings are reported in Appendices I and III.
Nuclear magnetic resonance (PMR and
13
C-NMR) spectra were taken with
a Joel FX-100 pulse Fourier transform nuclear magnetic resonance spectrometer equipped with a Texas Instrument 980B computer.
Preparative GLC was conducted on a Varian Aerograph 200 gas chromatograph equipped with a thermal conductivity detector and a Cole-Palmer 261 recorder. Prepurified helium (Matheson Gas Products) was used as the Column description and operating conditions are given in
carrier gas. Appendix I.
SOLUTIONS AND REAGENTS
Absolute methanol was distilled from magnesium methoxide (71).
Dry
pyridine was prepared by fractional distillation from sodium hydroxide (72). The triply distilled water was.prepared by a method to minimize trace organic contaminants (73).
TITANIUM SULFATE REAGENT (42).
Titanium sulfate (50 g, TiS0 4 H 2 4 S0 *8H 2 0, Fisher Scientific Co.) was dissolved in a concentrated sulfuric acid (50 ml) and diluted to 500 ml with triply distilled water. After standing one day, the solution was filtered
through a medium porosity, sintered glass filter and stored in a glass bottle.
-78-
PURIFIED SODIUM HYDROXIDE
A stock sodium hydroxide solution was purified by a method adapted from Reiner and Poe (9). contamination. This was done in an effort to minimize trace metal ion
All glassware used in the purification procedure was cleaned
with 35% nitric acid followed by several rinses with triply distilled water.
Reagent grade sodium hydroxide (650 g) and freshly boiled distilled water (1.7 liters) were mixed in a 3 liter round bottom flask. Phenyl 2-
pyridyl ketoxime (1.2 g) dissolved in a minimum amount of hot ethanol was added to the sodium hydroxide solution with 10% palladium on charcoal catalyst (0.2 g). The mixture was heated to ca. 100C and stirred with a Hydrogen gas was bubbled through the solution
Teflon coated magnetic bar.
for 8 hr, the solution was then stoppered and allowed to cool overnight. The sodium hydroxide was filtered through a medium porosity, sintered glass filter and each liter then extracted with ethanol:isopentyl alcohol; 25:75 (3 x 50 ml). The colored complex easily dissolved into the alcohols.
The sodium hydroxide was returned to the 3 liter round bottom flask with additional phenyl 2-pyridyl ketoxime (1.2 g). The solution was heated After 12 hr the The solution
as before, but nitrogen was bubbled through the solution.
solution was again stoppered and allowed to cool overnight.
was then extracted as before, followed by extraction with chloroform (3 x 150 ml) (42). The last traces of alcohol and chloroform were removed by The cooled
boiling for 6 hr while bubbling nitrogen through the solution.
sodium hydroxide solution was stored in a paraffin-lined, Teflon-stoppered glass jar under a nitrogen atmosphere. Analysis by GLC showed no ethanol
remaining in the caustic and a UV spectrum of the extracted caustic was identical to UV spectra of unextracted caustic thus assuring the removal of the phenyl 2-pyridyl ketoxime.
-79-
SODIUM THIOSULFATE SOLUTION
Air dry sodium thiosulfate (16.22 g) and sodium carbonate (Ool g) were dissolved in a liter of distilled water. This solution was titrated against
potassium iodate as a primary standard (71).
GLC INTERNAL STANDARDS
n-PROPYL 8-D-XYLOPYRANOSIDE
n-Propyl tri-O-acetyl-B-D-xylopyranoside (39.5 g), kindly provided by Dr. L. R. Schroeder, was dissolved in hot methanol (350 ml). 5 ml of 3% sodium methoxide was added. While still hot,
The reaction was monitored by TLC The reaction solution was deion-
(solvent 2) and was complete after 0.5 hr.
ized with Amberlite IR-120 cation exchange resin and evaporated in vacuo to a sirup (26 g). The sirup was thinned with acetone (25 ml) and placed in a The n-propyl B-D-
refrigerator where crystallization occurred spontaneously.
xylopyranoside was recrystallized from acetone:diethyl ether, 1:3, v/v to give 22.1 g (93%) yield: m.p. 93-94.5C; [a]29 = -61.6o (c 2.0, CH 3OH) (74). (c 1.0, CH30H). Literature:
m.p. 92-93C; [a] 2 0 = -58.6o
ETHANOL
Reagent grade ethanol (95%) was used as an internal standard in the methanol analysis.
MODEL COMPOUNDS
METHYL B-D-XYLOPYRANOSIDE (MBX)
Methyl 8-D-xylopyranoside (MBX) was purchased from Pfhanstiehl Laboratories, Inc. (Waukegan, Illinois). MBX was recrystallized from absolute
-80-
ethanol (2X), pulverized, and dried in a vacuum oven: -65.8 .( 9.0, H 20). 1.134, HO2). Literature (75):
m.p. 156-7.5C;
[a]20 =
m.p. 156-7C; [a] 2 0 = -65.8 (c
METHYL 8-D-RIBOPYRANOSIDE (MBR)
The methyl B-D-ribopyranoside used in this work was synthesized via the tri-O-benzoyl-ribopyranosyl bromide intermediate (Method I)o An alternate
method (Method II, Fisher glycosidation) was also used and is presented here because it was a better method.
Method I In a beaker 1,2-dichloroethane (380 ml), dry pyridine (200 ml) and benzoyl chloride (190 ml) were chilled to -20C by an acetone-dry ice bath (76). D-Ribose (50 g; Pfhanstiehl Laboratories, Waukegan, Illinois) was
added slowly and the temperature allowed to rise to 0C and kept there for 1 hr. The reaction mixture was refrigerated overnight and then allowed to The reaction mixture was then poured The organic layer was augmented The organic layer was
stand at room temperature for 6 hr.
into ice water and agitated for 0.5 hr.
with CHC13 (700 ml) and separated from the water.
washed with ice cold 3N H 2 S04 (3 x 360 ml), saturated NaHCO 3 (2 x 360 ml), and distilled water (3 x 360 ml). The organic layer was then dried over
Na 2SO 4, filtered through a carbon bed, and evaporated in vacuo to a clear, pale yellow sirup (230 g) of 1,2,3,4-tetra-O-benzoyl-D-ribopyranose (I).
Sirupy (I) was thinned with 1,2-dichloroethane (400 ml) and allowed to react with 200 ml of 33% (w/w) hydrogen bromide-acetic acid. was monitored by TLC (solvent 1). The reaction
After 4 hr, CHC13 (1000 ml) was added The organic
and the reaction mixture washed with ice water (3 x 2000 ml).
layer was then dried with CaCl2 , filtered, and evaporated in vacuo (35C) to
-81-
a solid.
The 2,3,4-tri-O-benzoyl-B-D-ribopyranosyl bromide (II) was recrysm.p. 147-150C. Literature:
tallized from diethyl ether to give 136 g (78%): m.p. 151-3C (76).
Crystalline (II) (105 g) was refluxed (3-4 min) until dissolved in absolute methanol (1000 ml) (46). After overnight refrigeration, crystalline (III) formed. Concentration of Compound (III) was = -69.70 (c
methyl 2,3,4-tri-0-benzoyl-B-D-ribopyranoside
the mother liquor yielded an additional crop of crystals. recrystallized from hot ethanol (82 g, 86%): 1.134, CHC13). Literature: m.p. 109-10C;
m.p. 108-9C; [a]l2
[a]21 = -69.5 (c 0.820, CHC13) (76).
Compound (III) (88 g) was suspended in absolute methanol (1000 ml) and CHC13 (350 ml). (30 ml). Debenzoylation occurred upon addition of 3% sodium methoxide The reaction solu-
The reaction was monitored by TLC (solvent 2).
tion was extracted with triply distilled water (1 x 500 ml, 2 x 300 ml) until TLC did not show any product in the CHC13 layer*. The water layer was
deionized with Amberlite IR-120 (H+ ) resin (30 g) and concentrated in vacuo (40C) to a sirup (31 g). The sirup was thinned with ethyl acetate (40 ml) The MBR was
and placed in a refrigerator where crystallization occurred. recrystallized twice from ethyl acetate (22 g, 73%): -106.2 (c 1.024, H 20). (c 0.47, H 20) (78). Literature:
m.p. 81-2C; [a]21
m.p. 82-3 (77); [a] 2l = -105.0o
Method II Absolute methanol (100 ml) was cooled in an ice bath and acetyl chloride (2 ml) slowly added. D-Ribose (10 g) was then dissolved in the acidified After the mixture had cooled, the yellow
methanol and refluxed for 6 hr (78)..
*The methanol caused difficulty in separation as it partitioned between the two layers. It is recommended that the procedure be modified by first evaporating the reaction solution to sirup before doing the extraction.
-82-
solution was decolorized with carbon (3X) and then the pale yellow solution was passed through Amberlite MB-3 (H , OH ) resin. evaporated in vacuo to a sirup (8 g). The neutral solutionwas
A portion of this sirup (4 g) was
eluted through a Silica gel column (145 g, 60-200 mesh) with ethyl acetate. Fractions (10 ml) were collected and MBR located in Fractions 26-50. Evapora-
tion in vacuo of these fractions yielded 1.5 g of sirup which readily gave crystalline MBRo If this alternative method is followed then MBR can theoretAfter the sirup has been column chromato-
ically be made in high yields.
graphed and the MBR collected, the column can be stripped and the resulting sirup refluxed again in methanol hydrogen chloride. As shown.by Cooper and
Bishop (79) an equilibrium will be reestablished with MBR as the predominant glycoside. This mixture can then be put back through the above work-up.
REACTION ANALYSIS
CONDITIONING OF REACTOR AND GLASSWARE
Prior to each reaction run all Teflon surfaces inside the reactor and all glassware used in preparing the reaction solution were washed with Alconox, acetone, distilled water, 35% nitric acid (v/v), and triply distilled water. The reactor and glassware were then allowed to air dry before use.
PREPARATION OF REACTION SOLUTION
Triply distilled water (ca. 400 ml) was boiled for 10 min while covered with a watch glass and then placed immediately in a nitrogen atmosphere inside a glove bag. Stock sodium hydroxide solution (ca. 33.6 g) was weighed
into a volumetric flask (250 ml), the flask purged with nitrogen and placed in the glove bag. After cooling, the triply distilled water was poured into The caustic solution
the volumetric flask to mark and thoroughly mixed.
was titrated against potassium acid phthalate to a phenophthalein end point.
-83-
If necessary, the alkalinity was adjusted until a 1.25 0.005M solution was achieved as determined by triplicate titrations. The reactor was placed in By weighing
the glove bag and alkaline solution poured into it (ca. 200 ml).
the volumetric flask before and after adding the solution to the reactor, the volume added was calculated from the density of 1.25M NaOH solution (1.0529 g/ml). Carbohydrate was then added to the reactor to achieve the desired The reactor cover was bolted into place and the solution The reactor was
concentration.
back-flushed through the sample line with nitrogen (1 min). then placed in the oil bath and the bath turned on.
GLYCOSIDE AND METHANOL ANALYSIS
General After the thermocouple output indicated that the temperature was stable (ca. 3.5 hr for a cold bath), the sample lines were purged by drawing two samples through them (for a description of the sampling procedure see Appendix IV, Reactor System). A time zero sample was taken in duplicate All samples were taken in dupli-
and the reactor pressurized with oxygen.
cate into tared 4 ml vials with generally one sample being used for glycoside analysis and one sample for methanol analysis. After each sample, the Prior
sample loop was back-flushed with water and acetone and vacuum dried.
to each sample, the sample lines were purged twice due to the dead volume between the reactor and the sample valve.
Glycoside Analysis Each sample size was determined gravimetrically and then internal standard.(n-propyl xyloside), as an aqueous solution, was added gravimetrically. The sample was then deionized over an Amberlite MB-3 (H , OH ) resin column (5 ml) and eluted with distilled water (3 x 4 ml). The deionized sample was
-84-
then evaporated in vacuo to ca. 1 ml in a pear shaped flask, transferred to an Erlenmeyer flask (10 ml, GGS) and evaporated in vacuo to dryness.
Each sample was dissolved in dry pyridine (0.8 ml) and acetylated at room temperature with acetic anhydride (0.3 ml). After 18 hr of mechanical
shaking, ice water (8 ml) was--added to each sample and the samples shaken for 0.5 hr. The aqueous mixture was extracted with CHC13 (2 x 5 ml) and the
chloroform extracts washed with iN HC1 (15 ml) and distilled water (10 ml). The chloroform extract was then dried over sodium sulfate, decanted, and evaporated in vacuo to dryness. The acetylated glycosides were dissolved in
1 ml of CHC1 3 and analyzed in triplicate by GLC (Appendix I, Conditions A).
Methanol Analysis Each sample size was determined gravimetrically and ethanol internal standard added volumetrically. The samples were then analyzed directly by
GLC (38) in triplicate (Appendix I, Conditions B).
To investigate whether or not the sampling method employed would give an accurate picture of liberated methanol, the reactor was-charged with a known concentration of methanol in 1.25M NaOH. The reactor was then placed in the Four
oil bath and the experiment conducted at normal reaction conditions. samples were taken from the reactor.
Between each sample, the headspace was The
varied in the reactor by emptying approximately 1/4 of the solution. reactor was allowed to equilibrate for 1 hr between each sample.
All of the
samples yielded methanol values (by GLC) that agreed within 1.5% of the initial methanol concentration (1.8 x 10-2 M) as determined gravimetrically.
-85-
PEROXIDE ANALYSIS
Peroxide Sampling The peroxide samples were taken through a line which by-passed the glycoside sample loop (Appendix IV, Reactor System). sample sizes to be obtained. This allowed larger
Prior to the taking of the peroxide samples The peroxide
the lines were purged with ca. 1 ml of reaction solution.
sample (ca. 5 ml) was then taken directly into an 8-ml vial.
Peroxide Analysis The peroxide analysis was conducted using a colorimetric method involving a complex between hydrogen peroxide and titanium(IV) (1,2,12-14,39-41). The
initial absorbance of the samples treated with the titanium sulfate reagent was taken as a measure of hydrogen peroxide. With time, absorbance usually
increased and the difference between initial and final absorbance was taken as a measure of organic peroxides. A straight-line calibration curve was
constructed by analyzing dilute hydrogen peroxide solutions by both the titanium sulfate method and a standard iodide-thiosulfate titration (80) (Fig. 29).
After the peroxide sample was taken from the reactor it was immediately analyzed. An aliquot (1 ml) was added to a 5-ml volumetric flask which Dupli-
contained enough sulfuric acid to give a final pH of either 0 or 1. cate analyses were conducted at each pH level.
After adding the sample and
shaking, 3 drops of the titanium sulfate reagent were added to each flask*. Distilled water was added to the mark and the absorbance of each sample was then taken at 400 nm within 5 minutes of the sampling time. The absorbance
*To avoid precipitation of the titanium sulfate, the sample must be acidified before the titanium sulfate reagent is added.
-86-
0
0 -
a)
0~~~~~~~~~~~~~~~~~~0
co
UN~~~.
0
*HAC)
0
0
o
0
CId'
C\J
0 0
U)
0)g
cr4
O
0)0H
.0) a)H
o
4.
,4i
0)
0 0H
0
0~~~~~~~ac
(0(
0~~~~~~0
.Kt
*
n
N~~~~~~~~~~~~~'
1*~~~~
0
'44
Ulu
oov7
'3Dma~osav
-87-
increase over a period of 2 days was followed with the maximum in absorbance for the pH 0 samples occurring at 24-36 hr. maximum When the pH 0 samples reached
absorbance, the pH 1 samples had an absorbance value of 40-70% of The volume of acid necessary to achieve the desired pH was All absorbance measurements were made against
the pH 0 value.
predetermined by a pH meter.
a distilled water reference as this gives the same result as a titanium sulfate reagent reference (42).
PRODUCT ANALYSIS
Product Sampling and Derivitization The product samples were taken through the same by-pass line as employed for the peroxide samples (Appendix I, Reactor System). Prior to the taking The
of each sample the lines were purged with reaction solution (ca. 1 ml).
samples (ca. 4 ml) were taken directly into tared 8-ml vials containing 1M NaHSO3 (1 ml). The sample sizes were determined gravimetrically and then
internal standard (n-propyl xyloside, as an aqueous solution) was added gravimetrically. Each sample was then deionized through an Amberlite IR-120 The pH of
(H ) column (10 ml) and eluted with distilled water (3 x 8 ml).
the deionized sample was immediately adjusted to ca. 7-8 by the addition of NH40H. The sample was then concentrated in vacuo (350 C) to ca. 1 ml in a The sample was
pear shaped flash and transferred to a 4-ml screw-top vial.
then evaporated in vacuo (35C) to dryness and vacuum evaporated (2X) with 1,2-dichloroethane to remove the last traces of water.
Approximately 30 to 50 mg portions of the product samples were transferred to 4-ml vials fitted with septa and the sample residue dissolved in dimethylsulfoxide (0.5 ml, silylation grade, Pierce Chemical Co.). To
assist dissolution, the mixture was heated in a block heater (65C, 15 min).
-88-
The mixture was then treated with Tri-Sil Concentrate (0.5 ml, Pierce Chemical Co.) and mechanically shaken for 12 hr. The top layer was then
analyzed by either GLC (Appendix I, Conditions C), GLC-MS (Appendix III), or preparative GLC-PMR (Appendix I, Conditions D). One sample, from which
methoxyacetic acid was identified, was subjected to an alternative work-up procedure. No bisulfite or NH40H was used and the sample was derivatized
as a CHC1 3 solution with BSTFA (Pierce Chemical Co.).
Mass Spectral Analysis The major reaction products were identified principally by GLC-MS (see Appendix III for mass spectrometer settings) as their TMS derivatives. Mass calibration was achieved through the use of an internal standard, Perflurokerosene (high mass; PCR, Inc., Gainesville, Florida). The fragmenta-
tion patterns of MBR and MBX were assigned by comparison with the fragmentation pattern of methyl 2,3,4-tri-O-methyl-a-L-arabinopyranoside as discussed by Kochetkov and Chizhov (81). The mass spectra and relative abundance data
of the identified compounds are reported in Appendix III.
Time Distribution of Products The concentration of the methyl C-carboxyfuranosides was examined as a function of time by quantitative GLC. The response factor of these furanoid This estimate was based upon
acids relative to PBX was estimated to be 0.9.
the calculated response factors of MBR and MBX relative to PBX (Appendix I, Table VII). The product samples were derivatized as previously described A sample chromatogram of
and the GLC conditions are given in Appendix I. the products is shown in Fig. 12.
-89-
Preparative GLC-PMR Analysis The assignment of the carboxyl function position in the methyl Ccarboxyfuranosides was accomplished by isolating the products.by preparative GLC and taking PMR spectra. An aliquot of run OS-4 (ca. 25 ml) was passed
over an Amberlite IR-120 (H ) column (12 ml) and eluted with distilled water (75 ml). After deionizing the pH was adjusted from 3 to 8 with 1N NaOH.
This solution was then eluted over an Amberlite IRA-400 (OH ) column at a rate of 1.5 ml/min, followed by elution with distilled water (1 liter) at a rate of 3.0 ml/min. elution with 1M NaHCO The acids were then displaced from the column via
3
(1500 ml) at 3.0 ml/min.*
The bicarbonate eluate
was deionized batchwise with Amberlite IR-120 (H ) until the pH reached ca. 3-4. The pH was then brought back to ca. 7-8 with NH40H. This solution was
then evaporated and derivatized as previously discussed (see Product Sampling and Derivatization).
The products were then separated by GLC (Appendix I, Conditions D) employing a thermal conductivity detector for peak detection. As the methyl
C-carboxyfuranosides eluted from the gas chromatograph, they were collected in glasswool-packed glass tubes cooled by an ice bath. The samples were
washed from the tubes with CDC13 into NMR tubes and analyzed within one hour by FT-PMR.
OXYGEN SOLUBILITY
The solubility of oxygen at the reaction conditions was determined by employing equipment developed by Green and Thompson (32). Reaction solutions
*The use of NaHCO 3 should be avoided if possible as it leads to severe deionizing problems. NaOH should be able to elute all but the dicarboxylic acids of the acids encountered in oxygen-alkali-methyl glycoside studies (15).
-90-
were withdrawn from the reactor by an hydraulic piston which then transferred the solution into a pressurized trap (N2 , 200-250 psig). In the trap,
the reaction solution was allowed to react with Winkler solutions (82) while the nitrogen pressure insured that the oxygen would not come out of solution. pressure. The final sodium thiosulfate titration was done at atmospheric
-91-
ACKNOWLEDGMENTS
The author greatly appreciates the guidance and support of this work by the advisory committee: and R. D. McKelvey. Drs. S. N. Thompson, Chairman; L. R. Schroeder
Special gratitude is extended to Drs. Thompson and The
Schroeder for numerous helpful conversations and votes of confidence. author also acknowledges the invaluable assistance given by numerous faculty and staff members.
The generous financial support provided by the member companies of The Institute of Paper Chemistry was greatly appreciated.
The atmosphere of friendship provided by the faculty, staff, fellow students and their wives and children, made this work more enjoyment than drudgery.
Finally I especially thank the following persons who encouraged me in my work: to the parents of my wife and me for their continued support; to
my sons, David and Walt, who were a welcomed distraction; and most importantly to my wife, Claire, for her unending encouragement and support of my cause.
-92-
LITERATURE CITED
1.
Millard, E. C. The degradation of selected 1,5-anhydroalditols by molecular oxygen in alkaline media. Doctoral Dissertation. Appleton, WI, The Institute of Paper Chemistry, 1976. Millard, E. C., Schroeder, L. R., and Thompson, N. S., Carbohyd. Res. 56:259(1977). Malinen, R., and Sjostrom, E., Cellulose Chem. Technol. 9:231(1975). Samuelson, 0., and Stolpe, L., Tappi 52(9):1709(1969). Kolmodin, H., and Samuelson, 0., Svensk Papperstid. 73(4):93(1970). Rowell, R. M., Pulp Paper Mag. Can. 72:T236(1971). Ericsson, B., Lindgren, B. 0., and Theander, O., 74:757(1971). Malinen, R., Paperi.Puu 57(4a):193(1975). Lindberg, B., and Theander, O., Acta Chem. Scand. 22:1782(1968). Kolmodin, H., and Samuelson, 0., Svensk Papperstid. 76:71(1973). Ericsson, B., Lindgren, B. 0., and Theander, 0., Cellulose Chem. Technol. 7:581(1973). McCloskey, J. T., Schroeder, L. R., Sinkey, J. D., and Thompson, N. S., Paperi Puu 57:131(1975). Sinkey, J. D., and Thompson, N. S., Paperi Puu 56:473(1974). McCloskey, J. T., Sinkey, J. D., and Thompson, N. S., Tappi 58(2):56 (1975). Ericsson, B., Lindgren, B. 0., 8:363(1974). and Theander, 0., Cellulose Chem. Technol. Svensk Papperstid.
2.
3. 4. 5. 6. 7.
8. 9. 10. 11.
12.
13. 14.
15.
16. 17. 18. 19.
Samuelson, 0., and Stolpe, L., Svensk Papperstid. 72:661(1969). Minor, J. L., and Sanyer, N., J. Polymer Sci., Part C, 36:73(1971). Brooks, R. D., and Thompson, N. S., Tappi 49:362(1966). Weaver, J. W. The alkaline hydrogen peroxide reaction of methyl S-Dglucopyranoside and methyl 4-O-methyl B-D-glucopyranoside. Doctoral Dissertation. Appleton, WI, The Institute of Paper Chemistry, 1976. Weaver, J. W., Schroeder, L. R.., 48:C5(1976). and Thompson, N. S., Carbohyd. Res.
20.
-93-
21. 22. 23.
Haskins, J. F., and Hogsed, M. J., J. Org. Chem. 15:1264(1950). Russell, G. A., J. Chem. Educ. 36:111(1959). McCloskey, J. T. The degradation of methyl 8-D-glucopyranoside by oxygen in alkaline solution. Doctoral Dissertation. Appleton, WI, The Institute of Paper Chemistry, 1971. Rendleman, J. A., Adv. Carbohyd. Chem. 21:209(1966). Rowland, S. P., Carbohyd. Res. 17:393(1971). Saric, S. P., and Schofield, R. K., Proc. Royal Soc. (London) (A), 185:431(1946). Church, J. A. Dissertation. The autoxidation of methyl glycopyranosides. Doctoral Appleton, WI, The Institute of Paper Chemistry, 1964.
24. 25. 26.
27.
28. 29. 30. 31.
McKelvey, R. D., Carbohyd. Res. 42:187(1975). Reiner, D., and Poe, D. P., Anal. Chem. 49(6):889(1977). Dewilt, H. G. J., and Kuster, B. F. M., Carbohyd. Res. 19:5(1971). Tanner, F. W. Model studies on the effect of oxygen in the alkaline hydrolysis of glucosidic bonds. Master's Dissertation. North Carolina State University, Raleigh, NC, 1973. Green, J. W., and Thompson, N. S., unpublished work. Bruhn, G., Gerlach, J., and Pawlek, F., Z. Anorg. Allgem. Chem. 337(1-2):68(1965). Thompson, N. S., personal communication. Crozier, T. C. Studies on the oxidative activation of loblolly pine dioxane lignin toward degradation with oxygen-alkali. Doctoral Dissertation. Appleton, Wisconsin, The Institute of Paper Chemistry, 1978. Frost, A. A., and Pearson, R. G. Kinetics and mechanism. 2, New York, Wiley and Sons, 1961. Laidler, K. J. York, 1965. Chemical kinetics. 2nd ed., Chap.
32. 33.
34. 35.
36.
37.
2nd ed., Chap. 1, McGraw-Hill, New
38. 39. 40. 41.
Sims, E. W., J. Chromat. Sci. 14:65(1976). Satterfield, C. N., and Bonnell, A. H., Anal. Chem. 27:1175(1955). Marklund, S., Acta Chem. Scand. 25:21(1971). Pobiner, H., Anal. Chem. 33(10):1423(1961).
-94-
42.
Sinkey, J. D. The function of magnesium compounds in an oxygen-alkalicarbohydrate system. Doctoral Dissertation. Appleton, WI, The Institute of Paper Chemistry, 1973. Verhaar, L. A., and DeWilt, H. G., J. Chromat. 41:168(1969). Peterson, G., Tetrahedron 26:3413(1970). Isbell, H. S., Adv. Chem. Ser. 117:71(1975). Ericsson, B., Lindgren, B. 0., Theander, 0., and Petersson, G., Carbohyd. Res. 23:323(1972). Michie, R. I. C., and Neale, S. M., J. Polymer Sci., Part 2a:2063(1964). Kano, T., Ishizu, A., Hosoya, S., and Nakano, J., J. Soc. Fiber Sci. Technol. Japan 30(5/6):T260(1974); Jap.; ABIPC 45:A5534. Seyfarth, H. E., Rieche, A., and Hesse, A., Chem. Ber. 100:624-8(1967); Angew Chem. Intern. Ed. Engl. 5:253(1965). McCloskey, J. T., Sinkey, J. D., and Thompson, N. S. Symposium, Madison, WI, 1975. Thompson, N. S., unpublished data. Rendleman, J. A., Adv. Chem. Ser. 117:51(1975). Neuberger, A., and Wilson, B. M., Carbohyd. Res. 17:89(1971). Izatt, R. M., Rytting, Jo H., Hanson, L. D., and Christensen, J. J., J. Chem. Soc. 88:2461(1966). Kabayama, Mo. A., Patterson, D., Can. J. Chem. 36:563-73(1958). Norrman, B., Acta Chem. Scand. 22:1623(1968). Roberts, E. J., Wade, C. P., and Rowland, S. P., Carbohyd. Res. 21:357 (1972). Bock, K., and Pederson, C., Acta Chem. Scand. B29:258(1975). Angyal, S. J., Aust. J. Chem. 21:2737(1968). Vinogradov, S. N., and Linnell, R. H. Hydrogen bonding. York, Van Nostrand Reinhold Co., 1971. Hordvik, A., Acta Chem. Scand. B28(2):261(1974). Brown, C. J., Cox, G., and Llewellyn, F. J., J. Chem. Soc. (A), 1966:922. Wells, H. A., Jr. An investigation of the vibrational spectra of glucose, galactose, and mannose. Doctoral Dissertation. Appleton, WI, The Institute of Paper Chemistry, 1977. Chap. 7, New Nonsulfur Pulping
43. 44. 45. 46.
47. 48.
49.
50.
51. 52. 53. 54.
55. 56. 57.
58. 59. 60.
61. 62. 63.
-95-
64. 65.
Kuhn, L. P., J. Am. Chem. Soc. 74:2492(1952). Stoddart, J. F. Stereochemistry of carbohydrates. Wiley-Interscience, 1971. p. 54-8. Toronto,
66.
Behar, D., Czapski, G., Rabini, J., Dorfman, L. M., and Schwarz, H. A., J. Phys. Chem. 74:3209(1970). Hayday, K., and McKelvey, R. D., J. Org. Chem. 41:2222(1976). Schroeder, L. R., personal communication. March, J. Advanced organic chemistry: reactions, mechanisms, and structure. p. 804-9. New York, McGraw-HIll, 1968. Carey, F. A., and Sundberg, R. J. Advanced organic chemistry. Reactions and synthesis. p. 324-8. New York, Plenum, 1977. Part B:
67. 68. 69.
70.
71.
Skoog, D. A., and West, D. M. Fundamentals of analytical chemistry. 835 p., 2nd ed., New York, Holt, Rinehart, and Winston, 1969. Perrin, D. D., Armarego, W. L. F., and Perrin, D. R. Purification of laboratory chemicals. 362 p., New York, Pergamon Press, 1966. Bauer, N., and Lewin, S. Z. In Weissberger's Techniques in organic chemistry. Vol. I, Part 1, p. 136., New York, Interscience, 1959. Debruyne, C. K., and Loontinens, F. G., Nature 209:396(1966). Fischer, E., Ber. 28:1157(1895). Jeanloz, E. L., Fletcher, H. G., Jr., and Hudson, C. S., J. Am. Chem. Soc. 70:4055(1948). Durette, P. L., and Horton, D., Carbohyd. Res. 18:403(1971). Jackson, E. L., and Hudson, C. S., J. Am. Chem. Soc. 63:1229(1941). Bishop, C. T., and Cooper, F. P., Can. J. Chem. 41:2743(1963). Kolthoff, I. M., and Sandell, E. B. Textbook of quantitative inorganic chemistry. 3rd ed., p. 574, New York, MacMillan, 1956. Kochetkov, N. K., and Chizhov, O. S., Adv. Carbohyd. Chem. 21:29(1966). American Public Health Association. Standard methods for the examination of water and waste water. p. 474, New York, APHA, 1971. Brandon, R. E. Alkaline degradation of 1,5-anhydrocellobiitol. Doctoral Dissertation, Appleton, WI, The Institute of Paper Chemistry, 1973.
72.
73.
74. 75. 76.
77. 78. 79. 80.
81. 82.
83.
-96-
APPENDIX I GAS-LIQUID CHROMATOGRAPHY
Quantitative GLC was accomplished through the use of an appropriate internal standard. Molar response factors relative to the internal standards
were calculated according to Equation (32).
F
x
= A
r
x M
r
(32)
where
F
x
= response factor for compound X relative to the internal standard
Ar = ratio of peak area of compound X to the peak area of the internal standard Mr = mole ratio of the internal standard to compound X
The response factors were experimentally determined by preparing a series of solutions with varied molar ratios and subjecting the solutions to the appropriate work-up procedure. The solutions were then analyzed by GLC in tripli-
cate and.the response factor calculated as an average of.the values obtained.*
Table VII lists the various GLC conditions used in this work and Table VIII gives GLC retention times and response factors of various compounds associated with this work.
Through the use of Equation (33) the response factors were used to calculate the concentration of compound X in the reaction samples.
[X] = 103 (Ar) (IS) p/(S) (F ) x
(33)
*This assumes that the response factorwas constant over the range of molar ratios and concentrations examined. Although the response factors did not seem to be affected by molar ratio, at lower concentrations of MBR and methanol deviation from the reported response factors occurred. However, these concentration ranges were not, generally, experienced in this work.
-97-
where
[X] p
=
concentration of compound X, moles/liter density of reaction mixture, g/ml
-
S = reaction sample, g IS = internal standard, moles
TABLE VII GAS-LIQUID CHROMATOGRAPHIC CONDITIONS
Conditions Column type Derivative
A SE-30a Acetylated
B Carbowax 20M
C OV-17 C Trimethylsilylated
Dd
Underivatized
Trimethylsilylated
Column temp. programming,
C
170
70
60-+160
@ 2/min
60-+150
@ 20 /min
Injector temp., C Detector temp., C N 2 flow rate, ml/min
260
150
260
200
260
265
260
200
15
35
20
20
aStainless steel column (5 ft x 0.125 inch) rigged for on-column injection and packed with 10% SE-30 on 60/80 mesh DMCS-AW Chromosorb W.
bStainless steel column (4 ft x 0.125 inch) rigged for off-column injection and flash vaporization and packed with 5% Carbowax 20M on 80/100 mesh Chromosorb 101. CStainless steel column (15 ft x 0.125 inch) rigged for on-column injection and packed with 3% OV-17 on 80/100 mesh Supelcoport. dThermal conductivity bridge detector - 100 ma current. eStainless steel column (10 ft x 0.25 inch) rigged for on-column injection and packed with 5% OV-17 on 80/90 mesh Anakrom ABS.
-98-
TABLE VIII RETENTION TIMES (Tr ) AND RESPONSE FACTORS (F ) Conditions A
Compound Methyl P-D-xylopyranoside Methyl B-D-ribopyranoside n-Propyl B-D-xylopyranoside
Tr , min 7.2 7.7 12.3 2.9 7.8 14.5 19.4 21.4 38.4 55.0 56.1 57.8 58.9 62.9 90.9
F
x
0,774 0.011 a
a 0.756 0.013
B
Methanol Ethanol
0.535 + 0.008 1.ooob
-
d'ooa
b
C
Methoxyacetic acid Lactic acid Glycolic acid Glyceric acid Methyl 6-D-ribopyranoside Methyl 3-C-carboxy-8-Dtetrafuranoside Isomeric methyl 2-C-carboxyB-D-tetrafuranosides Methyl B-D-xylopyranoside n-Propyl B-D-xylopyranoside
1.044 i0. 04 0 ac
0.894 0.019 a
0.900d 0.9008 0.900d 0.881 0.016 a 1.000 a
Calculated relative to n-propyl 8-D-xylopyranoside at the specified conditions. bCalculated relative to ethanol at the specified conditions. Calculated from triplicate injections of a single sample. dEstimated response factor.
-99-
APPENDIX II EXPERIMENTAL DATA
TABLE IX DEGRADATION OF METHYL S-D-XYLOPYRANOSIDE (MBX) (ca. 0.1M) IN 1.25M NaOH AT 120C, 0.682 MPa 02
Time, min
MBX, x 10 2 M
MeOH, x 103 M
Mole %, MeOH
Reaction 7MBX 0 15 30 60 120 240 420 660 930 1428 1914 2730 3360 4200 9.77 9.61 9.43 8.95 8.14 6.98 6.19 5.36 4.94 4.21 3.70 3.17 2.82 2.41
1.2 3.6 5.7 10.4 17.5 23.6 27.8 34.3 35.4 36.0 39.6 44.0 45.7
75 106 70 64 63 66 63 71 64 59 60 63 62
Reaction 8MBX 0 20 40 60 120 240 420 660 900 1410 2100 2850 3540 4290 9.57 9.44 9.11 8.69 7.79 6.88 5.99 5.31 4.79 4.05 3.41 2.94 2.58 2.25
0.9 3.3 6.3 10.3 17.0 23.6 29.7 33.4 38.2 39.6 41.7 45.3 45.8
69 72 71 58 63 66 70 69 69 64 63 65 63
-100-
TABLE X DEGRADATION OF METHYL B-D-RIBOPYRANOSIDE (MBR) (ca. 0.1M) IN 1.25M NaOH AT 120C, 0.682 MPa 02
Time, min
MBR x 10 M
MeOH, x 103 M
Mole %, MeOH
Reaction 5MBR 0 10 30 50 70 90 120 175 300 480 810 1380 2100 2880 9.66 9.30 7.90 6.65 5.68 5.20 4.54 3.72 2.85 2.37 1.91 1.46 1.11 0.88
3.7 11.4 17.8 22.5 26.5 31,6 36.6 42.1 45.8 49.5 49.7 53.7 54.1
103 65 59 57 59 62 62 62 63 64 61 63 62
Reaction 6MBR 0 5 30 50 70 90 120 177 300 480 840 1440 2130 2880 9.65 9.65 8.25 7.07 5.94 5.20 4.34 3.51 2.70 2.15 1.62 1.23 0.94 0.72
0.7 10.7 14.5 22.0 25.3 30.9 35.7 43.1 49.4 51.3 49.8 53.0 50.1
76 56 59 57 58 58 62 66 64 59 61 56
-101-
TABLE XI DEGRADATION OF METHYL 3-D-XYLOPYRANOSIDE AND METHYL B-D-RIBOPYRANOSIDE IN 1.25M NaOH AT 120C AND 0.682 MPa 0 2 a
Time, min
MBX, x 102 M
MeOH, x 103M
Mole %, MeOH
Reaction 3MBX 0 60 120 180 308 420 600 890 1200 1530 1900 2582 10.18 9.70 9.75 9.68 9.60 9.30 8.92 8.30 7.64 6.95 6.25 4.92
0.0
'0.5 1.0 1.3 2.5 3.8 6.0 9.8 14.0 19.1 22.6 30.8
13 30 33 52 54 52 55 57 61 59 60
Reaction 3MBR 0 70 120 250 480 942 1295 1560 1800 9.83 9.64 9.71 9.63 8.87 7.65 6.86 6.31 5.74 0.0 0.4 0.6 1.9 5.1
11.1
16.4 21.2 24.0
21 50 95 53 51 55 60 59
aOxygen deficient degradations, i.e., limited agitation.
-102-
TABLE XII DEGRADATION OF METHYL B-D-XYLOPYRANOSIDE (MBX) (ca. 0.1M) IN 1.25M NaOH AT 120C, 0.0 MPa 02 (N2 atmosphere)
Time, min
MBX x 10 M
Time, min Reaction 10MBX
MBX, x 102 M
0 370 1025 1640
9.66 9.61 9.89 9.69
2375 3080 3840 4380
9.57 9.68 9.61 9.50
TABLE XIII DEGRADATION OF METHYL B-D-RIBOPYRANOSIDE (MBR) (ca. 0.1M) IN 1.25M NaOH AT 120C, 0.0 MPa 02 (N2 atmosphere)
Time, min
MBX, x 102M
Time, min
MBX, x 102 M
Reaction 7MBR 0 365 1015 1550 10.18 10.27 10.20 10.14 2420 2990 3860 4360 10.02 10.11 10.04 10.19
-103-
TABLE XIV PEROXIDE FORMATION DURING DEGRADATION OF METHYL a-D-XYLOPYRANOSIDE (MBX) (ca. 0.1M) IN 1.25M NaOH AT 120 C, 0.682 MPa 02
Time, min
H202, x 104 M
ganic Peroxides, OrE x 104 M
4 15 30 45 60 105 210 420 600
Reaction 7MBX 0.4 2.9 3.8 3.6 3.2 2.7 1.8 0.9 0.8 Reaction 8MBX
0.3 0.3 0.5 0.8 1.1 2.2 5.8 12.4 15.3
10 20 30 40 50 60 75 105 180 300
2.0 3.5 3.8 3.9 3.7 3.5 3.4 2.7 2.2 1.5 Reaction
9 MBXab
0.2
0.4 0.7 0.6
0.8 1.3 2.3 4.7 8.5
5 15 30 40 50 65 85 110
1.1 1.6 2.3 3.2 3.1 2.8 1.8 1.3
aperoxides determined at test pH of 1.9. No organic peroxides detected after 36 hr of hydrolysis;
-104-
TABLE XV PEROXIDE FORMATION DURING DEGRADATION OF METHYL a-D-RIBOPYRANOSIDE (MBR) (ca. 0.1M) IN 1.25M NaOH AT 120C, 0.682 MPa 02
Time, min
H2 0 2 , x 104 M
Organic Peroxides, x 104 M
Reaction 5MBR 2 10 20 30 40 50 68 105 117 172 0.7 7.4 10.2 10.1 8.9 7.7 6.4 4.8 3.8 2.5 Reaction 6MBR 5 15 30 50 70 90 150 2.3 7.2 8.2 8.8 7.6 6.3 3.3
0.7 1.2 2.0 2.8 3.9 5.4 6.6 8.7
0.6 1.2 2.2 3.6 4.3 7.3
-105-
TABLE XVI PRODUCT DISTRIBUTION FROM THE DEGRADATION OF 0.1M METHYL XYLOSIDE AND METHYL RIBOSIDE IN 1.25M NaOH AT 120C and 0.682 MPa 02
Time, hr
% Reacted Glycoside
Producta , x Reaction 6MBR
Product ), x 103M
Productb
103 M
0.3 0.5 1.0 1.5 2.0 4.0 10.0 35.4
7 15 33 46 55 66 77 90
c -0.4 2.1 2.8 3.1 3.3 3.8 Reaction 8MBX
0.2 1.6 3.7 4.6 5.8 7.5 7.9 10.3
0.3 0.6 1.8 3.8 5.0 6.2 6.1 6.5
0.5 1.5 3.0 6.0 9.0 15.0 35.0 59.0
3 14 23 34 41 50 64 73
-0.4 0.7 1.1 1.4 1.6 1.8 1.4
0.4 1.4 2.5 3.7 4.6 5.3 6.2 5.8
-0.5 0.9 1.3 1.7 2.0 1.8 0.7
aMethyl 3-C-carboxy-a-D-tetrafuranoside. Isomeric methyl 2-C-carboxy-8-D-tetrafuranoside. CUnable to determine due to tailing methyl riboside peak.
-106-
APPENDIX III MASS SPECTROMETRY ANALYSIS
The equipment utilized for the mass spectrometry was previously described (see Experimental). The GLC conditions were identical to Conditions
C (Appendix I, Table VII) except helium was used as the carrier gas instead of nitrogen. in Table XVII. The operating conditions for the mass spectrometer are given
TABLE XVII MASS SPECTROMETER OPERATING CONDITIONS Jet separator temperature: Separator block temperature: Oven temperature: Source temperature: Ionization voltage: Vacuum: Filament: Scan: 115C 180C 70 eV 300C 300C
2 x 10- 7 torr GC mode
10 sec/decade
The mass spectral data are presented in Tables XVIII-XXV.
Fragmenta-
tion patterns for the methyl glycosides and the methyl C-carboxyfuranosides are presented in Fig. 30 and 31, respectively. The product numbers referred
to in the table headings correspond to the product numbers assigned in the Results section.
-107-
TABLE XVIII MASS SPECTRAL DATA FOR METHYL $-D-XYLOPYRANOSIDE (MBX) AND METHYL 8-D-RIBOPYRANOSIDE (MBR) (TMS DERIVATIVES) Relative Abundance, m/e
%
m/e MBX
Relative Abundance, %
73 74 75 89 101 102 103 116 117 129 133 135 146 146 148
93 10 6 7 5 34 5 4 5 7 42 5 7 27 6 MBR
149 189 191 201 204 205 206 217 218 219 233 305 349
5 5 4 6 100 20 11 81 11 5 4 4 1
73 74 75 89 101 103 116 117 129 131 133 134 143 146 147 148 149
63 11 13 12 8 5 10 4 7 4 56 6 5 4 33 6 5
159 189 191 201 204 205 206 217 218 219 233 247 305 349 365
3 8 3 4 69 20 8 100 35 20 4 3 6 <1 <1
aFigure 30 gives a fragmentation pattern for some of the major ions.
-108-
Ur) -. %
--
5%
-4.
z
00 U coE 0-02
Lx
o
N
0-
E
0
'0 4)
x
fn
U .0
'0
I
U)
U
'I
0 U)
0
4)
0~ U I-
4)
"S
2-
z U=0
I C)
+
T--
ON%
U)
4
2 1-X-~
U-0
.II
C-)
U)
4)
0)
x
C-)
I
7
-S
*0
to
U)
i0
U 0 .4!00
0
U I.0
X IU-0 Ul) U0
II
I-. U-0
U)
0
4)
E..
I.-
Figure 30.
Fragmentation Pattern for Some Major Ions in the MBX and MBR Mass Spectra
-109-
TABLE XIX MASS SPECTRAL DATA FOR METHOXYACETIC ACID (TMS DERIVATIVE) (MBR AND MBX PRODUCT O ) Relative Abundance, % Known MethoxyProduct acetic Acid Bb Aa b .Aa 100 14 42 15 12 4 8 75 3 3 37 6 17 5 3 6 4 4 9 4 3 4 75 26 14 9 17 14 27 6 3 72 35 23 100 87 70 8 67 16 10 7 80 31 60 6 4 51 10 6 4 12 3 11 94 17 9 100 78 64 100 13 8
m/e 73 74 75 83 85 86 87 88 89 90 91 100 101 102 103 104 105 113 114 115 116 117 118 119 130 131 132 133 146 147 148 162,P 163 164 177 178 179 191 192 193 205 206 207 208
11
5
57 4
4 72 51 11
3
4 21 44 7 10 3
14
27 4 3
70 14 1
25 10 5
21 6 3 70 27 18 10 4
aDerivatized without the use of NaHSO 3 . NaHSO3 used in samp le work-up.
-110-
TABLE XX MASS SPECTRAL DATA FOR LACTIC ACID (TMS DERIVATIVE) S (MBR AND MBX PRODUCT )a
Relative Abundance, m/e 73 74 75 89 101 117 118 119 131 100 12 18 7 3 49 5 3 3 m/e 133 147 148 149 190 191 192 219
Relative Abundance,
%
6 70 9 6 7 8 4 2
Spectrum taken on MBX product sample.
TABLE XXI MASS SPECTRAL DATA FOR GLYCOLIC ACID (TMS DERIVATIVE) (MBR AND MBX PRODUCT )a
Relative Abundance, m/e 73 74 75. 89 101 102 103 115 117 131 133 134 135
%
Relative Abundance, m/e 147 148 149 96 59 28 25 4 43 5 5 4 63 10 8 1
100 18 22 3 4 4 10 6 9 13 18 5 3
161
162 177 178 179 191 205 206 207 220,P
aSpectrum taken on MBX product sample.
-111-
TABLE XXII MASS SPECTRAL DATA FOR GLYCERIC ACID (TMS DERIVATIVE) (MBR AND MBX PRODUCT Q)a
Relative Abundance, m/e 73 74 75 89 102 103 107 131 133
%
Relative Abundance, m/e 147 148 189 205 292 307
%
100 7 14 3 7 10 7 3 4
21 5 8 3 4 1.
aSpectrum taken on MBX product sample.
-112-
TABLE XXIII MASS SPECTRAL DATA FOR METHYL 3-C-CARBOXY-g-DTETRAFURANOSIDE (TMS DERIVATIVE) (MBR AND MBX PRODUCT Q )a
Relative Abundance, m/e 73 74 75 89 101 113 116 117 127 129 131 133 134 143 147 148 149 157 159 162 174 177 189 191 201 204 205
%
Relative Abundance, m/e 207 215 216 217 219 221 231 232 233 243 244 245 259 276 277 291 316 319 334 335 336 363 379 380 381 394
%
100 12 15 12 3 3 9 4 3 10 14 10 7 7 54 7 9 4 3 4 4 10 3 4 8 5 6
4 14 7 28 3 7 3 3 4 4 5 8 5 5 3 11 3 6 45 12 13 1 8 4 4 1
aMass spectra taken from the MBR and MBX systems were identical within the operat :ing limits of the mass spectrometer.
-113-
TABLE XXIV MASS SPECTRAL DATA FOR METHYL 2-C-CARBOXY-B-DTETRAFURANOSIDE (TMS DERIVATIVE) (MBR AND MBX PRODUCT O ) a,b
Relative Abundance, m/e 73 74 75 89 101 103 116 117 118 119 127 128 129 132 142 147 148 149 157 163 173 177 190 191 201 207 208
%
Relative Abundance, m/e 215 216 217 219 221 229 230 231 233 244 245 246 247 262 276 283 289 291 308 319 320 334 335 336 363 379 394
%
100 11 15 11 3 3 4 3 3 6 4 4 4 16 5 41 18 10 5 16 7 6 6 3 6 4 6
6 19 10 4 4. 5 5 3 4 6 10 4 17 4 16 3 5 10 3 7 3 67 19 11 2 19 2
aMass spectra from MBR and MBX systems were identical within the operating limits of the mass spectrometer. Fragmentation pattern for some major ions are given in Fig. 31.
-114-
4.
TMS,
OR
P+, m/e 394
I
' TMSO OT~Ms~ OCH3 .M o COOTMS
I-
CH 3
OTMS
m/e 379 OCH
1- C
TMSO m COOTMS m/e 334 - OTMS m/e 245 CH
I
m/e 319
H3C\ '
OH 3
.
Si H3 C H3
Figure 31.
OTMS m/e 147
Fragmentation Pattern for Some Major Ions in the Mass Spectrum of MBX and MBR Products()
and Q
-115-
TABLE XXV MASS SPECTRAL DATA FOR METHYL 2-C-CARBOXY-S-DTETRAFURANOSIDE (TMS DERIVATIVE) (MBR AND MBX PRODUCTQ )a,b
Relative Abundance, m/e 73 74 75 89 101 114 116 117 119 129 131 132 133 141 142 145 147 148 149 156 161 175 186 190 191 201 205 208 211 215 216 217 218
%
Relative Abundance, m/e 219 221 229 230 231 232 233 244 245 246 247 259 262 277 288 289 291 292 293 319 321 334 335 336 337 347 348 351 363 379 380 381 394 3 6 4 4 12 4 6 6 5 24 7 4 4 21 4 3 8 4 4 15 8 57 19 6 5 14 3 6 2 25 5 5 6
100 20 30 17 3 5 6 4 5 10 10 5 31 4 5 4 58 12 12 5 9 8 5 12 15 7 8 5 4 13 4 13 5
aMass spectra from MBR and MBX systems were identical within the operating limits of the mass spectrometer. Fragmentation patterns for some major ions are given in Fig. 31.
-116-
APPENDIX IV REACTOR SYSTEM
REACTION VESSEL
The reaction vessel was the same as described by Millard (1) and the reader is referred to his thesis for a detailed description of the vessel. The inside of the reactor was completely Teflon lined which minimized the possibility of surface catalysis. Agitation in the reactor was accomplished
by an air-driven magnetic stirrer also described by Millard (1).
OIL BATH AND TEMPERATURE CONTROL APPARATUS
The oil bath pan design was patterned after the pan described by Brandon (83). Insulation was provided by calcium silicate and the reaction vessel The bath was heated
was raised and lowered by a rack and pinion mechanism.
by a Polyscience 73 immersion circulator (0-1000 w; Polyscience Corp., Niles, Illinois) equipped with a pre-set thermostat (120C; J. L. Stortz Div., PSG Industries, Inc., Perkaskie, Pennsylvania) to insure temperature reproducibility. As a safety precaution, the heater was plugged into an Over-Temp
Guard (Instruments for Research and Industry, Cheltenham, Pennsylvania) set at 130C.
Reactor temperature was sensed by an inconel clad, copper constantan, grounded thermocouple (Omega Engineering Inc., Stamford, Connecticutt). The
signal from the thermocouple was referenced to a type T thermocouple reference junction (Charles T. Gamble Ind., Riverside, New Jersey) and the output recorded by a Leeds and Northrup Speedomax Flatbed 628 recorder. As sensed
by the thermocouple, the reactor temperature was controlled to 0.2C.
-117-
SAMPLING SYSTEM
The reactor was sampled under pressure through Teflon lines (0.063 inch OD x 0.031 inch ID, Chromatronix, Inc.) and a series of pressure tight, inert valves (Chromatronix, Inc.) (Fig. 32). The Teflon line was backed by The
stainless steel tubing for the first 12 inches leading from the reactor.
sampling system could be arranged to take either a small sample (kinetic or methanol sample, ca. 0.8 ml) through the sample loop or a large sample (peroxide or product sample, ca. 4-5 ml) through the bypass to the sample loop. In all cases, sample valve ports B2, B3, and A2 were blanked.
GLYCOSIDE AND METHANOL SAMPLING PROCEDURE
1. 2.
The three-way valve was positioned to connect ports C2 and C3. The sample valve was placed in position B. This connected ports
A3, A4, C2, and C3 filling the sample loop with reaction solution. 3. The sample valve was placed in position A. By use of the syringe
attached to port Al, the sample was blown out port B4 into the sample vial. 4. With the sample valve in position A, the sample loop was backflushed by drawing solvents through port B4 and out port Al with a vacuum.
PEROXIDE AND PRODUCT SAMPLING PROCEDURE
1. 2.
The three-way valve was positioned to connect ports C2 and C1. The sample valve was placed in position B. Pressurized reaction
solution then flowed from the reactor through ports A3, A4, C2, and C1 directly into the sample vial. 3. When the desired sample size was obtained, the sample valve was returned to position A.
-118-
Peroxide Sample Position B
From Reactor
Sample Loop
Methanol & Glycoside Samples Position A
Figure 32.
Schematic Representation of Reactor Valve System