Institute The of Paper Chemistry
Appleton, Wisconsin
Doctor's Dissertation
The Oxidation of Simple Organic Compounds With Aqueous Chlorine Dioxide Solutions
Roger A. Somsen
June, 1958
THE OXIDATION OF SIMPLE ORGANIC COMPOUNDS WITH AQUEOUS CHLORINE DIOXIDE SOLUTIONS
A thesis submitted by Roger A. Somsen B. S. 1953, University of Wisconsin M. So 1955, Lawrence College
in partial fulfillment of the requirements of The Institute of Paper Chemistry for the degree of Doctor of Philosophy from Lawrence College, Appleton, Wisconsin
June, 1958
TABLE OF CONTENTS INTRODUCTION PRESENTATION OF PROBLEM EXPERIMENTAL PROCEDURES Organic Compounds Used Chlorine Dioxide Generation and Purification Preparation of Reaction Mixtures Determination of Chlorine Dioxide Reaction Rates Determination of Organic Reaction Products Nongaseous Reaction Products Carbon Dioxide EXPERIMENTAL RESULTS Kinetic Studies at pH 7 Kinetic Studies at pH 1 Organic Reaction Products Discussion and Summary CONCLUSIONS LITERATURE CITED APPENDICES I.. Preparation and Analysis of Aqueous Chlorii ne Dioxide Solutions 1
3 6
6
7 7
9
10 10 17 19 19
25
31
35
41
43 45
45 48
51 51
II. III.
Quantitative Transfer of Chlorine Dioxide Solutions Gas Chromatography Apparatus Physical Components Detection Components Column Preparation Operation
55 56 59 63 68 75
IV.
V.
Kinetic Data at pH 7 Kinetic Data at pH 1 Calculation Methods for Kinetics Constants at pH 1
VI.
INTRODUCTION It was originally thought that chlorine dioxide was specific in its
bleaching reactions, attacking only the noncarbohydrate material present in the cellulose. This concept was based upon the work of Schmidt and Graumann
(1) who reported that cellulose, oxidized cellulose, sugars, and other similar compounds were not attacked appreciably during treatments with chlorine dioxide at room temperature for twenty-four hours (2, 3), 'however, Staudinger and co-workers
found that the degree of polymerization (D.P.) of cellulose
was lowered by treatment with chlorine dioxide, while Jeanes and Isbell (4) showed that aldose sugars were attacked slowly. Marked modification of cellulose was reported by Lapeze and Dardelet (I) with the application of 40% chlorine dioxide based on the cellulose at 70C. Their results, under these relatively drastic experimental conditions.
showed that marked decreases in D.P. and alpha-cellulose content andinmcreases in copper number and carboxyl content occurred over the pH range 2 to
7. When conventional amounts of applied chlorine dioxide were used on
normally bleached sulfate and sulfite pulps, however, no appreciable change in either D.P. or hot alkali solubility were noted (6-8). On the other hand,
the carboxyl content of a strongly overbleached pulp (treated with an acid chlorine bleach liquor to produce carbonyl groups) increased greatly upon treatment with chlorine dioxide, while the carbonyl content decrease. Therefore, it is apparent that while the hydroxyl groups in cellulose are quite resistant to attack by chlorine dioxide, any carbonyl groups resulting from overbleaching are quite susceptible to attack.
-2Ethanol is the only simple compound containing a functional grouping present in cellulose whose reaction with chlorine dioxide has been studied (2) * Upon adding a solution of potassium chlorate dropwise to a mixture of equal volumes of ethanol and sulfuric acid, ethyl acetate, chloroform,
chlorohydrin, chloral, and chloroacetic acid were identified as reaction products Since the reaction conditions in this ethanol oxidation are far more severe than would be the case with an oxidation in dilute aqueous solution, the reaction products of the oxidations may be quite different.
-3PRESENTATION OF PROBLEM Although the importance of chlorine dioxide as a bleaching agent increases each year, little is known about its reactions with carbohydrates In order that the most intelligent use can
and other types of molecules.
be made of its bleaching properties, it is desirable that a complete understanding of its chemical reactions be obtained. A thorough study of the reactions of chlorine dioxide with cellulose is difficult because this reaction system is heterogeneous, the analytical methods for the functional groups present in an oxidized cellulose are not good, and the methods of hydrolyzing an oxidized cellulose to small identifiable compounds are open to question. These problems can be eliminated by
a study of simple molecules containing only one of the several functional groupings which have been proposed for cellulose at various degrees of oxidation (10)----glycol, c-hydroxyketone, and a-diketone for the C2 and
C3 positions in the anhydroglucose ring and primary alcohol and aldehyde for the C6 position Such a study would also provide a foundation for further studies with more complex molecules, such as sugars and cellulose. The compounds chosen for study were simple molecules containing these proposed functional groupings. These compounds are identified and compared with the structures of hypothetical oxidized celluloses in Figure lo n-
Butyraldehyde was chosen for study at higher temperatures because of the low boiling point of acetaldehyde. The rates of reaction between chlorine
dioxide and each of these compounds in dilute aqueous solution were to be determined as a function of temperature at pH 1 and 7. The oxidizing
-4properties of chlorine dioxide were to be characterized by the determination of the organic reaction products from the oxidations of diacetyl and 2,3butanediol at pH l.
-5--
-6EXPERIMENTAL PROCEDURES ORGANIC COMPOUNDS USED The source and method of purification for each of the starting organic compounds are listed below: 1. The 2,3-butanediol (EK P3400) was purified by the method of Leslie and Castagne (11)o The fraction boiling at 179-180Co was collected.
2. Acetoin (EK P3788) was purified by the method of Westerfeld (12). The fraction boiling at 141-142C. was collected and stored under refrigerationo 30 Diacetyl (EK 1591) was purified by the method of Westerfeld (12).
The fraction boiling at 89-90C. was collected and stored under refrigeration. 4. nitroge. 5. The n-butyraldehyde (EK 440) was purified by drying three times Acetaldehyde (EK 468) was purified by fractional distillation under
over Drierite and fractionally distilling under nitrogen through a Vigreux column, The fraction boiling at 74Co was collected and transferred to These vials were stored
4-ml. screwcap vials in a nitrogen atmospher.
under refrigeration and one vial was opened for each sample required. 6. The ethanol was Rossville Gold Shield Alcohol, 200 proof, obtained It was purified by the method of
from Commercial Solvents Corporation. McComb and McCready (13 .
-7CHLORINE DIOXIDE GENERATION AND PURIFICATION Aqueous solutions of purified chlorine dioxide were prepared by passing a chlorine-air mixture through two sodium chlorite columns arranged in series. The unreacted chlorine was removed by passing the chlorine-chlorine
dioxide-air mixture through a saturated aqueous solution of barium hydroxide. The chlorine dioxide was then dissolved in triple distilled water containing enough nitric acid to make the pH about lo The presence of acid prevents The concentrations of
hydrolytic disproportionation of chlorine dioxide.
chlorine dioxide and chlorine were determined by titration of a neutral sample and an acid sample with N/10 sodium thiosulfate. found when the generator was operating properly. cedure are given in Appendix I. PREPARATION OF REACTION MIXTURES Reaction mixtures were prepared from three solutions. the preparation of each of these solutions are listed below: 1. Buffer Solution. For reactions at pH 1, 25 mi. of nitric acidThe details of No chlorine was
The details of this pro-
potassium nitrate buffer (105 go potassium nitrate and 51 ml. of nitric acid added to 978 ml. triple distilled water) was added for each 100 ml. For reactions at pH 7, 45 ml. of mono-dipotas-
of total reaction mixtureo
sium phosphate buffer (400 go dipotassium phosphate and 50 ml. 85% phosphoric acid added to 1800 ml. of triple distilled water) was added for each 100 ml. of total reaction mixtureo Buffer catalysis effects, if present,
were shown by adding only 80% of the above amount of buffer at pH 1 and 67%
-8at pH 7.
When buffer concentration was found to affect reaction rate, 40 ml.
of 00775 molal sodium sulfate solution was added to a low buffer concentration mixture to give the finished reaction mixture the same ionic strength as a regular buffer concentration reaction mixture. In this manner, ionic
strength effects upon reaction rate were separated from buffer catalysis effects.
2.
Organic Solution.
The starting organic compound was weighed into The
a volumetric flask and diluted to volume with triple distilled water. volume of this solution necessary to give the required reaction mixture concentration (generally O.O10M 3. was pipetted into the buffer solutio.
Chlorine Dioxide Solution.
Sufficient purified and thermostated
chlorine dioxide solution was added to the mixture with the transfer device described in Appendix II to make the reaction mixture about O.OlOOM in chlorine dioxid. This device minimized transfer loss of chlorine dioxide
because the air space above the solution was kept saturated with chlorine dioxide The buffer solution was added to a volumetric flask at least four hours before the oxidation was begun so that its temperature could equilibrate with that of the bath in which the reaction was to be run. The organic
solution was added just before the chlorine dioxide solution and the total then diluted to volume with triple distilled water at the oxidation temperature. added. Time zero was determined when the chlorine dioxide solution was After thorough mixing, the reaction mixture was ready for either
chlorine dioxide rate studies or organic reaction product analysis.
-9For studies of the effect of surface area upon reaction rate, 0.2 g. of glass fibers was dispersed in the reaction mixtureo These fibers had a
weighted average fiber diameter of 164. and a surface area of 10,840 cm..g. DETERMINATION OF CHLORINE DIOXIDE REACTION RATES Due to the volatility of chlorine dioxide, rate studies required that
the reaction be carried out in a vessel with a variable volume so that there was no gas phase above the reaction mixture effective for this use. inges with aluminum foilo Hypodermic syringes proved
Light effects were minimized by covering the syrThe completed reaction mixture was transferred
to the syringe by the application of air pressure to the surface of the reaction mixture (see Figure 2A). When the syringe was filled, the stop-
cock was closed; the delivery tube attached, and the unit (see Figure 2B) placed in a constant temperature bathe Samples for the determination of
chlorine dioxide concentration were removed from the delivery tube. The following method of iodometric analysis was used for the determination of total active chlorine concentration in the reaction mixture sampleso Ten milliliters of buffer solution (100 go of dipotassium phosphate dissolved in 450 ml. of distilled water and adjusted to a pH of 7o0-7e5 with phosphoric acid), 10--ml of 4N sulfuric acid, and 0.5 g. potassium A portion of this mixture The reaction
iodide were mixed in a 125-ml. Erlenmeyer flask.
was added to a lO-mlo volumetric flask which was then weighe.
mixture sample (about 1 ml.) was then added to this flask and the flask
-10weighed again The contents of the volumetric flask were added quantita-
tively to the Erlenmeyer flask, mixed, and allowed to stand for two minutes* The sample was then titrated with N/100 sodium thiosulfate using Thyodene indicator near the end point. The sample weight was converted to volume by
determining the density of the reaction mixture at 25Co The concentration of total active chlorine in the reaction mixture was calculated from the following equation:
where V was the volume of N/O00 sodium thiosulfate solution, d was the reaction mixture density, and w was the sample weights DETERMINATION OF ORGANIC REACTION PRODUCTS NONGASEOUS REACTION PRODUCTS The scheme for the separation of the completed reaction mixture resulting from the oxidation of diacetyl and 2,3-butanediol at pH 1 into fractions for further analyses is given in Table I. Because gas chromatography
columns are not capable of handling large quantities of water, it was necessary to extract the solution with an organic solvents This extraction with
redistilled Freon-TF solvent was carried out in a continuous Schmall liquidliquid extractor (14) for denser liquids. Each 100 ml. of reaction mixture It was necessary to
was extracted with 200 ml. of redistilled Freon-TF.
concentrate this extract because of the small amount of organic material dissolved in the Freon-TF solvent. The concentration was accomplished using
-12a stainless steel wire-wound fractionating column suggested by Dr. E. F. Thode, which was operated at a reflux ratio of about 4:1 with the distillation pot immersed in a water bath at 58C. The extraction was concentrated
to a final volume of about 25 ml. for further analysi. The volatile fatty acids present in the extractedreaction mixture were removed by steam distillation (200 ml. of distillate was collected for each 100 ml. of reaction mixture distilled). The concentrated distillate was
purified by the addition of ethanol to precipitate inorganic salts. Only three fractions-concentrated Freon-TF extract, steam distillation residue, and steam-distilled extracted reaction mixture--were analyzed for A gas chromatography apparatus (see Appendix III for
organic components
details) was constructed to determine the volatile material present in the concentrated Freon-TF extractions The experimental conditions under which The components of the
these analyses were performed are listed in Table II.
mixture were removed from the gas stream by freezing them out into separate glass U-tubes. A suitable solvent was added to each U-tube and the infra-
red spectrum was obtained by running this solution against the solvent on a Perkin-Elmer Model 21 Recording Spectrophotometer. identified by comparison with known spectra The compound was
The amount of material in
each gas chromatographic fraction was determined from standard Beer t s Law curves, The curves for diacetyl and 2,3-butanediol are given in Figures 3
and 4, respectively. Because the absorption spectrum of the carbonyl material from the 2,3butanediol oxidation was not sufficiently strong for identification, the 2,4-dinitrophenylhydrazine derivative was prepared (15).
-13-
TABLE I
SEPARATION OF ORGANIC REACTION PRODUCTS
Reaction Mixture
&---Cool
to 0Ce
Quenched Reaction Mixture i---K + No 2 S 2 03 Neutralized Reaction Mixture
Saturated Reaction Mixture ---Freon-TF
NIFreon-TF Extract +-Concentrate I Extracted Reac tion Mixture
<-Steam Distill
~c |
Concentrated 1 Freon-TF Extract
Distillate
Steam Distillate
1--KOH
Alkaline Solution (P
9 )
Steam Distilled Extracted Reaction Mixture
Concent:
.~~~~ .
Concentrated Solution <-EtOH
Water Inorganic Salts Water-EtOH Inorganic Salts
EtOH
Purified Solution
|--Concentrate
Concentrated Solution j< EtOH
Purified Solution
, 1--Evaporate t( Dryness
Steam Distillation
Residue 1 Fractions analyzed
-14TABLE II OPERATING CONDITIONS OF GAS CHROMATOGRAPHY APPARATUS Column support material Column liquid Liquid content on support material Column diameter Column length Bath temperature Helium flow rate Chart speed Amplifier volume Bridge voltage C-22 crushed firebrick DC Silicone 710 Fluid 29% (w/w) 1/2 in. 12 2/3 ft. 70, 100C. 200 ml. (S.T.P.)/min. 12 in./min. maximum 6.0 v.
All three of the fractions were analyzed for nonvolatile material by paper chromatography, acidic material. Ethanol:ammonia (100:1) solvent was used to develop
The presence of acids was shown by spraying the sheets
with formaldehyde (4% solution in ethanol) to destroy the ammonium salts, drying the sheets, and then spraying the dried sheets with a methyl orange
(1 g./l. in water) solution Ethyl acetate:acetic acid formic acid water
(18:3:1:4) solvent was used to develop carbonyl materials The presence of carbonyl material was shown by spraying the sheets first with a 2,4-dinitro-
phenylhydrazine (0*5 go in 250 ml. 3N hydrochloric acid) solution and then with a 10% solution of potassium hydroxide in water* The component(s) of
the steam distillation residue was also identified by the infrared spectrum of a potassium bromide pellet
by its weight.
The quantity of this residue was determined
-15-
-16-
I
I
I
-17CARBON DIOXIDE The carbon dioxide evolved during the oxidations of diacetyl and 2;3butanediol was determined with the equipment outlined in Figure 5. When the
oxidation was completed, the Day pinchclamps were removed and nitrogen was bubbled through both a blank reaction mixture (contained no organic materia) and an organic reaction mixture. Excess chlorine dioxide was removed by
washing the gas stream with acidic potassium iodide, and carbonyl material was absorbed in the 2,4-dinitrophenylhydrazine solution. After drying the
gas stream with Drierite and magnesium perchlorate, the carbon dioxide was absorbed quantitatively on Ascarite. The difference in weight increase
between the blank and organic oxidation was a direct measure of the carbon dioxide evolved by the oxidation mixture.
-18-
-19EXPERIMENTAL RESULTS KINETIC STUDIES AT pH 7 The experimental data upon which the following observations are based are listed in Appendix IV. these observations. The effect of temperature upon the rate of decomposition of blank chlorine dioxide solutions buffered at pH 7 is shown in Figure 6. It is Two Figures and tables are given below to illustrate
apparent that these solutions are quite unstable at this pH value.
reactions have been proposed to account for the decomposition of neutral and acid chlorine dioxide solutions:
It
was reported that in unbuffered solutions the rate of decomposition This is definitely not the case,
was slow in both acid and neutral systems.
however, in solutions buffered at pH 7 with phosphate. A typical reaction plot at pH 7 is illustrated in Figure 7. The amount
of chlorine dioxide consumed in the oxidation of diacetyl appears to be appreciable, but consumption values must be known numerically for a reaction The difference between the blank and organic oxidation
kinetics study.
analyses may be assumed to be a direct measure of the oxidant consumption. The organic oxidation, however, results in a faster buildup of inorganic reaction products as well as a lower chlorine dioxide concentration than
-20that in the blank. Both of these effects result in a lower rate of chlorine As a result, the blank
dioxide decomposition in the oxidation mixture.
solution overcompensates and conclusions based on this difference technique are not theoretically sound. Since chlorate ion is formed in both of the proposed decomposition equations, and since no chlorate ion should be produced during the oxida-
tions of organic compounds, the chlorate ion concentration in a reaction mixture should be a direct measure of the amount of decomposed chlorine dioxide. Techniques for the determination of chlorate, however,, were found
to be both insensitive and lengthy and, therefore, not suitable for use in a kinetics study. Consequently, the difference technique was applied to
several oxidations and several interesting curves were obtained (see Figure 8). The reaction rate would appear to be zero order over an extended
period of the reaction when the rate of consumption of active chlorine is not too rapid and when the difference between the oxidant concentration in the blank and organic oxidation is not too great. When these two conditions It
are not fulfilled, plots similar to Curve B (Figure 8) are obtained. should be noted, however,
that neither of the apparent zero order plots
(Curves A and C, Figure 8) could be extrapolated to zero consumption at zero time This effect might be the result of an initial, very rapid It is this possi-
reaction which is followed by a second, slower reaction
ble second reaction that would be illustrated in Figure 8. The stability of the starting compounds toward chlorine dioxide attack at pH 7 is expressed in Table III in terms of the maximum temperature at
-21-
-22-
-23which no reaction was detectable. Hydroxyl compounds are apparently very This stability
stable to chlorine dioxide attack at pH 7 even at 80C.
decreases rapidly, however, upon the substitution of carbonyl groups for hydroxyl groups. rapidly at 10C. TABLE III STARTING COMPOUND STABILITY TO CHLORINE DIOXIDE AT pH 7 Compound Ethanol 2,3-Butanediol Acetoin Diacetyl Acetaldehyde Temperature, >80 >80 10 <10 <10 C.a To illustrate, both diacetyl and acetaldehyde react
a Maximum temperature at which no reaction was noted.
The effect of buffer concentration upon the rate of oxidant consumption at pH 7 is shown in Figure 9. An increase in reaction rate with increased buffer concentration could be produced either by buffer catalysis or by change in ionic strength. Buffer catalysis provided the only expla-
nation for this effect, however, when it was found that the addition of sodium sulfate to the low buffer concentration reaction mixture did not change the reaction rate0 Such buffer catalysis phenomena have been obser-
ved in many halogen oxidation systems (18, 19) and were shown to be caused by changes in the concentration of the acid form of the buffer.
-24-
-25In summary, the experimental kinetics observations at pH 7 are qualitative in nature rather than quantitative because of the decomposition of the chlorine dioxide solutions and of the presence of buffer catalysis. It
was shown, however, that the hydroxyl compounds are much more resistant to attack by chlorine dioxide than compounds containing carbonyl groups. The
reactions with acetoin and acetaldehyde did appear, however, to be of zero order. KINETIC STUDIES AT pH 1 The experimental data upon which the following observations are based are listed in Appendix V. these observations. The stability of the various functional groupings toward chlorine dioxide attack at 800C is illustrated in Figure 10. These groupings can, primary alcohol> basically the Figures and tables are given below to illustrate
therefore, be ranked in the following order of stability: a-glycol > a-hydroxyketone > a-diketone > aldehyde. same order that was noted at pH 7 (see Table III).
This is
The curve for n-butyraldehyde at 80C. appears to be quite different, however, from that of the other compounds. When the reaction is carried
out at 50C.
(see Figure ll), however, the reaction rate plot shows that the
total reaction is the sum of two successive apparent zero order reactions. The first reaction of this sequence is much more rapid than the second. It is possible to analyze the kinetics of these oxidations directly . from the experimental data because blank chlorine dioxide solutions buffered
-26-
-27at pH 1 were found to be stable. order reaction rate. Each oxidation followed an apparent zero
The rate constants listed in Table IV for these The activation energies
reactions were calculated from rate plot slopes.
and temperature coefficients listed in Table V are based upon these rate constants (see Appendix VI for calculation methods). The activation
energies listed are of the magnitude expected for reactions occurring under these experimental conditions since values of about 20 kg.-cal./mole are normal for reactions at ordinary temperatures (35 p. 1091). The initial
reaction in the n-butyraldehyde oxidations shows a markedly lower activation energy. This result, however, would be expected from the greater reactivity
of the aldehyde group over that of the ketone and hydroxyl groups under the same experimental conditions. In contrast to the buffer catalysis phenomenon noted at pH 7, buffer concentration was found to have no effect on reaction rate at pH 1. Since surface area often affects the rate of zero order reactions, its effect at pH 1 was studied in the oxidation of acetoin at 60 and 80C. When the free
surface area in the reaction syringe was doubled by the addition of finely chopped glass fibers, there was no apparent change in reaction rate. Since
acetoin was thought to be typical of the type of compounds being studied, no surface effects would be expected with the other compounds. In summary, the oxidations at pH 1 were found to follow apparent zero order reaction rates and were not affected by either buffer concentration or surface area. Hydroxyl groups were again found to be more stable to chlorine
dioxide attack than carbonyl groups.
-29-
-30TABLE IV REACTION RATE CONSTANTS AT pH 1 Compound Ethanol 2,3-Butanediol Acetoin Diacetyl n-Butyraldehyde Temperature, 80 70 80.5 80 70 80.5 80 60 80.5
80
C.
ka x 105 1.6
0.65 6.0 5.7 2.6
26 21
60 80b 60 b 50 c 40C
4.9 64 56 9.8
6.0
1.8
6.0
4.3
a (Equivalents total active chlorine)/(liter)(minute). b Based on final rate, Figure 11. c Based on initial rate, Figure 11. TABLE V ACTIVATION ENERGIES AND TEMPERATURE COEFFICIENTS AT pH 1 Compound Ethanol 2,3-Butanediol Acetoin Diacetyl n-Butyraldehyde Temperature, C. a a Temperature Coefficient 2.5 2.2 2.1 2.9 1.4 1.7
70, 80
70, 80
60, 80 60, 80
40,
21 19 18 21
60,
a Kilogram-calories/mole. b Based on initial rate. c Based on final rate.
80
50 b
6.8
14
c
-31ORGANIC REACTION PRODUCTS The results from the chromatographic analyses of the three fractions from the oxidations of diacetyl and 2,3-butanediol at pH 1 are listed in Table VI and Table VII, respectively. Acetic acid, isolated as potassium
acetate, is the only nongaseous reaction product that was found in large enough quantities to be identified by infrared spectra and paper chromatography in each oxidation. The formation of acetic acid indicates that these
functional groupings are oxidized with a cleavage of carbon-carbon bonds. This cleavage results in the formation of acetic acid from both a-diketone and a-glycol groupings. Gas chromatographic analysis of the concentrated Freon-TF extraction from the oxidation of 2,3-butanediol showed the presence of a very small amount of material of low retention volume, approximately that for diacetyl. The low retention volume indicates that the material had a low boiling point (below 100C.) and the shape of the peak (blip), which appears on the recorder, shows that it probably did not contain either hydroxyl or carboxyl groups. Infrared spectrum analysis was difficult because the extremely The
small amount of material present produced only small absorption peaks. spectrum, however, did show the three major absorption bands of diacetyl (5.81, 7.38, and 8.99P). The carbonyl band at 5.81l,
however, appears as a
shoulder on a stronger band at 5.77L. Absorption at this wavelength may be attributed to a-halogen substituted ketones, a-halogen substituted acids, or saturated aliphatic aldehydes. The remaining bands in the spectrum were not The
of sufficient strength to allow correlations of spectra with structure.
2,4-dinitrophenylhydrazine derivative(s) of this material was prepared and,
-32TABLE VI REACTION PRODUCT ANALYSIS FOR DIACETYL OXIDATION a Fraction Concentrated Freon-TF extract Steam distillation residue Steam distilled extracted reaction mixture Experimental Results Gas Chromatography Paper Chromatography Unreacted diacetyl present - - No acid spots No carbonyl spots KOAcb present No carbonyl spots No acid spots No carbonyl spots
a Oxidation carried out at pH 1 and 60Co for 8 hro bPotassium acetate
TABLE VII REACTION PRODUCT ANALYSIS FOR 2,3-BUTANEDIOL OXIDATIONa Fraction Concentrated Freon-TF extract Steam distillation residue Steam distilled extracted reaction mixture a Oxidation carried out at pH 1 and 80C bPotassium acetate Experimental Results Gas Chromatography Paper Chromatography Very small amount carbonyl material present - - No acid spots No carbonyl spots KOAcb present No carbonyl spots No acid spots No carbonyl spots
for 19 hr.
-33without purification, was found to char at 288-292'C. dinitrophenylhydrazone) chars at 315C. Diacetyl bis(2,4-
Although no irrevocable proof for
the presence of diacetyl has been established, both the infrared spectrum of the gas chromatographic fraction and decomposition point of the derivative indicate that it may be present in this trace material.
The possible presence of diacetyl in this oxidation mixture indicates that the mechanism of the 2,3-butanediol oxidation involves the conversion of at least part of the a-glycol groupings to a-diketone groupings before the molecules were split into carboxylic acid molecules. The concentration
of diacetyl could never become very large, however, because of its rapid oxidation by chlorine dioxide to acetic acid and carbon dioxide. The material balance summaries for the oxidations of diacetyl and 2,3-butanediol at pH 1 are presented in Table VIII and Table IX, respecteJy*. These data show that carbon dioxide as well as acetic acid have been isolated as reaction products. The two successive apparent zero order
reactions noted for the oxidations of n-butyraifehyde (see Figure 11) can be explained on the basis of the formation of a carboxylic acid and carbon dioxide. The initial rapid reaction could be the oxidation of n-butyralde-
hyde to n-butyric acid while the second slower reaction could be the degradation of n-butyric acid to carbon dioxide. The following three points should be noted in considering the quantitative data presented in Table VIII and Table IX: 1. There was undoubtedly some loss of diacetyl during the concentration of the Freon-TF extract during Run 68, and unreacted no 2,3-butanediol
-34TABLE VIII MATERIAL BALANCE SUMMARY FOR OXIDATION OF DIACETYLa
Run no.
Initial concentration CO 2 Diacetyl Total volume, ml. Reaction time, hr. C102 consumed, mmol. Compounds determined, mmol.
Acetic acid C02 Unreacted diacetyl
68
0.0103M 0.0100M 1000 8 9.6
6.3 1.0 32 --
71
0.0139M O.0100M 1000 8 - - 2.4 - - 6
Compound yields
Acetic acid CO2
Unreacted diacetyl Total yield, %
10 48c
- -
C Oxidations carried out at pH 1 and 60C. b Molar percentage of initial diacetyl. c Total yield could not be determined from any single run (p.ll).
TABLE IX MATERIAL BALANCE SUMMARY FOR OXIDATION OF 2,3-BUTANEDIOLa Run noo. Initial concentration C102 2,3-Butanediol Total volume, ml. Reaction time, hr.
C102 consumed, mmol.
69 O.OU9M 0.0101M 1000 19
10.3
70 0.012o1M 0.0103M 1000 19
-
Compounds determined, mmol. Acetic acid
CO2 Unreacted 2 3-butanediol
45
- b - 22 -- b
- 0.8 - -2
-
Compound; yields
Acetic acid C02 Unreacted 2,3-butanediol
Total yield, % a b c d
24d
Oxidations carried out at pH 1 and 80Co None found but test sensitivity low. Molar percentage of initial 2,3-butanediole Total yield could not be determined from any single run (page 11).
-35was recovered during Run 69. would be low. 2. While the technique for determining carbon dioxide was quantitative Consequently, the total yield values reported
only 74% of the acetic acid present in a solution of known concentration could be recovered in the steam distillation residue. for recovered acetic acid would probably be low. 3. The yields of acetic acid and carbon dioxide could not be reported for the same reaction mixture. As a result, there is some question as to Therefore, the values
whether or not the same degree of oxidation was accomplished for each of the determinations. Bearing these points in mind. Table X was constructed upon the assumption that the oxidations followed the equations below: 5 CH COCOCH3 + 2 C102 + 6 H2 0 -3 10 CH COOH + 2 HC1 3
5 CH3COCOCH 3 + 18 ClO-->20 C02/+ 6 H2 0 + 18 HC1 5 CH CHOHCHOHCH 3 + 6 C10 2-3 5 CH CHOHHCHOHC 3 It
3
10 CH3COOH + 6 HC1 + 2 H20 H20 + 22 HC1
+ 22 C10 ->20 C02/+ 1 2
is immediately apparent that in no instance has more than 50% of the
consumed chlorine dioxide or unrecovered organic starting material been determined. If chlorine dioxide in acid solution is reduced to chloride
ion (17), the assumed equation should be correct and, therefore, the low consumption and recovery values must be the result of incomplete recovery of acetic acid and carbon dioxide and/or formation of other reaction products which were not detected. DISCUSSION AND SUMMARY Reactions exhibiting zero order mechanisms are not commonplace
-36-
-37in physical chemistry. Many free radical reactions, however, have been Free radicals are defined as
shown to proceed by zero order mechanisms.
molecules or atoms having an odd number of electrons which produce an unbalanced electron spin and, thereby, permanent magnetic moments (20). Electron diffraction studies of chlorine dioxide in the gas phase have shown that the chlorine dioxide molecule can be represented mainly by the following double bonded structure (21):
00
8C1
Therefore, the structure of the chlorine dioxide molecule fits the definition presented for a free radical and, consequently, a free radical reaction mechanism could be a possible explanation of these zero order reaction rates. No experimental proof for this hypothesis can be presented t
however, since the mechanism of the possible free radical formation in aqueous solution is not known. The two successive apparent zero order reactions measured for the oxidation of n-butyraldehyde at pH 1 can be explained as resulting from the rapid formation of n-butyric acid followed by the slower formation of carbon dioxide. Only a single reaction rate could be determined for each of This result could originate from the fact
the other compounds, however.
(1) that only the initial reaction in which the acid is formed was rapid enough to be measured, or (2) that the rates of formation of the two
reaction products are not sufficiently different to allow their separate determination.
-38The formation of acetic acid during the oxidations of diacetyl and 2,3butanediol shows that chlorine dioxide is a sufficiently strong oxidizing agent to cleave carbon-carbon bonds. This result was expected because the
presence of carbonyl and hydroxyl groups on the C2 and C3 positions would make these positions more liable to oxidative attack. agents--peracetic acid (22), periodic acid (23), Three other oxidizing
and calcium hypochlorite
(24)--have also been shown to produce acetic acid from diacetyl. The formation of carbon dioxide during these oxidations is probably the result of chlorine dioxide attack upon the carboxylic acid formed during the oxidation. If this is the case, hydroxyl, carbonyl, and carboxyl groups are
unstable in chlorine dioxide solutions at pH 1. This oxidation of molecules containing carboxyl groups was unexpected because, generally, such oxidations of the lower fatty acids have been found to be difficult (5-27). The stability of the various functional groupings toward chlorine dioxide attack at pH 1 can be ranked in the following order: primary alcohol > a-glycol > a-hydroxyketone > a-diketone > aldehyde. Essentially
the same ranking order seems to also hold for oxidations at pH 7. Since tentative evidence has been presented to show that the a-glycol grouping is oxidized to the a-diketone grouping before the molecule is acetic acid, it cleaved to form
is likely that the a-hydroxyketone grouping also would be cleaved. Since,
oxidized to the a-diketone grouping before this molecule is
in general, the nonspecific oxidation of a hydroxyl group is more easily accomplished than cleavage of a carbon-carbon bond, it would appear anomalous
that the a-hydroxyketone is more stable to attack than the a-diketone.
-39Diacetyl, however, dissociates readily to acetyl free radicals in the preSince chlorine dioxide molecules are free radicals,
sence of light (28).
this same dissociation could occur in their presncee.. Once acetyl free radicals have been formed, they should be readily oxidizable to acetic acid, thereby. explaining the more rapid oxidation of the a-diketone. The results from the kinetic and reaction product studies on the oxidations of these simple compounds may be related to the possible chlorine dioxide oxidation of cellulose in the following ways: 1. Carbonyl groups were found to react more readily with chlorine
dioxide than hydroxyl groups at both pH 1 and 7, This observation correlates directly with the slight cellulose degradation produced by chlorine dioxide on a normally bleached pulp and the marked degradation of an overbleached pulp. 2. Carbonyl compounds were found to react readily with chlorine This
dioxide at pH 7, while hydroxyl compounds did not react at this pHo
result would mean that during chlorine dioxide bleaches near the neutral point, the reaction with cellulose should be limited primarily to those points in the anhydroglucose chain which had already been oxidized to carbonyl groups in the previous bleaching steps. Therefore, if the previous
bleaching operations have not damaged the pulp, there should be little cellulose attack near pH 7. 3. Hydroxyl groups were found to be stable to chlorine dioxide attack Consequently, more oxidative
at pH 7 but were oxidized slowly at pH lo
attack on the hydroxyl groups in cellulose could be expected as the bleach liquor pH decreases.
-404. Acetic acid and carbon dioxide were isolated as reaction products from the oxidations of diacetyl and 2,3-butanediol. This cleavage of carbon-
carbon bonds during cellulose oxidations would result in the formation of carboxyl groups at the C2 and C3 positions in the anhydroglucose chains.
-41CONCLUSIONS 1. The general aspects of the chlorine dioxide oxidations studied include: a. Chlorine dioxide solutions buffered at pH 7 (mono-dipotassium Solutions buffered at pH 1 (nitric are stable.
phosphate buffer) are unstable.
acid-potassium nitrate buffer), however, b.
While buffer catalysis effects were very prominent with each
compound oxidized at pH 7, no such effect was found with oxidations at pH 1. c. Surface area was found to have no effect upon the rate of reaction at pH 1. 2. The stability of the functional groupings toward chlorine dioxide primary alcohol > Essentially the same however,
attack at pH 1 can be ranked in the following order:. a-glycol > a-hydroxyketone > a-diketone > aldehyde. ranking order appears to hold at pH 7.
The hydroxyl compounds,
react more readily at the same temperature with chlorine dioxide at pH 1 than at pH 7 while the reverse is true for the carbonyl compounds. 3. All reaction mechanisms for oxidations at pH 1 are apparently zero
order with respect to chlorine dioxide consumption, and there are indicattions that the same is true at pH 7. 4. The activation energies and temperature coefficients calculated from the rate constants for pH 1 oxidations were of the magnitude expected for reactions occurring under these experimental conditions.
-425. The isolation of acetic acid and carbon dioxide as reaction products from the oxidations of diacetyl and 2,3-butanediol at pH 1 shows that there is appreciable chain cleavage during these oxidations of C4 compounds to C2 and C1 fragments. 6. The two successive reactions noted in the oxidation of n-butyraldehyde at pH 1 can be explained on the basis of the formation of n-butyric acid followed by the formation of carbon dioxide. 7. The isolation of a small amount of carbonyl material from the oxidation of 2,3-butanediol which appears to contain diacetyl indicates that the a-glycol is oxidized to diacetyl before the molecule is cleaved to form two molecules of acetic acid.
-43LITERATURE CITED 1. 2. Schmidt, Erich, and Graumann, Erich, Ber. deut. chem. Ges. 54B:1860-73 (1921); C.A. 16:273. Staudinger, H., and Jurisch, J., Zellstoff u. Papier 18, no. 12:690-2
(Dec., 1938); Paper Ind. 19: 1317, 1319(1938).
3.
4.
5. 6.
Staudinger, H. and Dohle, W., J. prakt. Chem. 161, no.8-10:219-40 (Jan. 12, 19435; Translation. Jeanes, Allene
27:125-42(1941). Lapeze, R. P., and Dardelet, S., Ann. Inst. Polytechnique Grenoble, numero special:23-35(1952); Translation. Samuelson, Olof, and Ramsel, Curt, Svensk Papperstidn. 53:155-63(1950). Svensk Kern. Tidskr. 62:197-204(1950)
and Isbell, Horace S., J. Research Natl. Bur Standards
7. Samuelson, Olof, and Hartler, Nil1,
8. Rapson, W. Howard, Tappi 39, no. 5:284-95(May, 1956). 9. 10. Bhaduri, Kshitibhushan, Z. anorg. Chem. 84:113-14(1907). McBurney, L. F. Oxidative degradation. In Ottts Cellulose and cellulose derivatives. 2nd ed. Part 1. p. 145. New York, InterScience Publishers, 1954. Leslie, J. D., and Castagne, A., Can. J. Research 24F:311-19(1946). Westerfeld, W. W., J. Biol. Chem. 161:495-502(1945). McComb, Elizabeth A., and McCready, R. M., Anal. Chem. 24, no. 10:16302(Oct., 1952). Schmall, Morton, Pifer, C. W., and Wollish, E. G., Anal. Chem. 24, no. 9:1446-9(Sept., 1952). Staff of Hopkins and Williams Research Laboratory. Organic reagents for organic analysis. p. 68. Brooklyn, Chemical Publishing Company,
ll1 12. 13. 14. 15.
Inc., 1950.
16. 17. Taube, Henry, and Dodgen, H., J. Am. Chem. Soc. 71:3330-7(1949). Brown, Richard W. Studies on the chemistry of chlorine dioxide. Doctorts Dissertation. Appleton, Wis., The Institute of Paper Chemistry, 1951. 109 p.
4418. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. Fletcher, Harry H., and Taylor, T. Clinton, J. Am. Chem. Soc. 60:3018.25(1938). Bunzel, H. H., and Mathews, A. P., J. Am. Chem. Soc. 31:464-79(1909). Waters, We A. The chemistry of free radicals. Oxford University Press, 1948. 2nd ed. p. 21. London,
Dunitz, J. D., and Hedbert, Kenneth, J. Am. Chem. Soc. 72:3108-12(1950) Boeseken, J., 24:2427. and Slooff, G., Rec. trav. chim. 49:91-4(1930); C.A.
Clutterbuck, Percival W., and Reuter, Fritz, J. Chem. Soc. 1935:1467-9. Suknevich, I. F., and Chilingaryan, A. A., Ber. deut. chem. Ges. 69B, 1537-42(1936); C.A. 30:5936. Morette, Andre, and Gaudefroy, Ghislain, Compt. rend. 237:1523-5(1953); C.Ao 48:4941. Polonovski, Michel, Compt. rend. 178:576-8(1924); C.A.18:1271. Knoop, F., and Gehrke, M., Z. physiol. Chem. 146:63-71(1925); C.A. 19: 2844. Walling, Cheves. Free radicals in solution. Wiley & Sons, Inc., 1957. Luther, R., and Hoffmann, R., Z. physik. Chem. 755-69(1931); C.A. 26:1178. p. 547. New York, John
Bodenstein-Festband:
Casciani, Ferri, Paper Trade J. 135, no. 10:22-32(Sept. 5, 1952). Giertz, Hans W., Svensk Papperstidn. 54:469-76(1951). James D. H., and Phillips, C. S. G., J. Sci. Instr. 29:362-3(1952). James A. T., and Martin, A. J. P., Biochem. J. 50:679-90(1952).
Cropper, F. R., and Heywood, A. The analysis of fatty acids and fatty In Destyts Vapour phase alcohols by vapour phase chromatography. chromatography. p. 316. New York, Academic Press, Inc., 1957. Glasstone, S. Textbook of physical chemistry. Van Nostrand Company, Inc., 1946. 2nd ed. New York, D.
35.
-45APPENDIX I PREPARATION AND ANALYSIS OF AQUEOUS CHLORINE DIOXIDE SOLUTIONS
Chlorine dioxide was prepared by passing a controlled mixture of chlorine gas and air through two sodium chlorite (technical grade) beds as illustrated in Figure 12. The unreacted chlorine was removed by bubbling 3o) .
the gas mixture through a saturated barium hydroxide solution (29
The purified gas mixture was then bubbled through triple distilled water containing 3 ml, of 1N nitric acid per liter to produce the chlorine dioxide solution. This procedure is a modification of Pulping Group Procedure 74 of
The Institute of Paper Chemistry. This chlorine dioxide solution was analyzed iodometrically for chlorine dioxide and chlorine by thy neutral and acid titration method of Giertz (31): Twenty milliliters of the liquor was transferred into a mixture of 100 ml. of water, 2 g. potassium iodide crystals, and 20 ml. of buffer solution (100 g. of dipotassium phosphate dissolved in 450 ml. of water and adjusted to a pH of 7.0-7.5 with phosphoric acid). The sample was immedi-
ately titrated with O1lN sodium thiosulfate using Thyodene indicator near the end point. The amount of thiosulfate consumed (ml.) = Tn. Twenty
milliliters of 4N sulfuric acid was added and after 3 minutes the titration was continued titration = Ts. The total amount of sodium thiosulfate consumed (ml.) in the (Since the acidic titration was a direct continuation of the
neutral one, the Tn value also entered into the Ts value.) The neutral titration corresponds to 1/5 of the oxidizing equivalent of
-46chlorine dioxide, and the chlorine reacts quantitatively. The additional
amount of thiosulfate consumed in the acidic titration corresponded to 4/5 of the oxidizing equivalents of chlorine dioxide. Tn) and g./l. C1 02 Thus, C10 2 = 5/4 (T -
1.688 (Ts - Tn)(normality of sodium thiosulfate),
C12 = Ts - 5/4 (Ts - Tn) and g./l. C12 = (3.55) [T - 5/4(Ts - Tn)](normals ity of sodium thiosulfate). If no C12 was present, Ts = -5/4(T - Tn).
-47-
APPENDIX II QUANTITATIVE TRANSFER OF CHLORINE DIOXIDE SOLUTIONS Quantitative transfer of aqueous chlorine dioxide solutions was accomplished with the apparatus illustrated in Figure 13. was operated in the following manner: 1. The bath temperature was controlled at the oxidation temperature This transfer device
with a Fenwal Thermoswitch. 2. Purified chlorine dioxide solution was added to both flasks. The
solution level in the small flask was kept above the bottom of the air inlet tube. 3. The buret was then filled by applying a slight positive pressure to
the air inlet tube after closing the stopcock below the Bunsen valve and opening the three-way vent stopcock to both the atmosphere, and the buret. The level of the solution in the buret was controlled by the source stopcock. 4. The three-way vent stopcock was then closed to the atmosphere. The
stopcock below the Bunsen valve was opened while the chlorine dioxide solution reached bath temperature in order to relieve any pressure formed because of temperature change. This stopcock was closed thereafter only
when chlorine dioxide solution was being removed. 5. When the chlorine dioxide solution reached the bath temperature (about 2 hours), the buret was rinsed and then the correct sample volume for chlorine dioxide analysis or reaction mixture preparation was measured from the buret. This transfer device makes it possible to transfer chlorine dioxide
-49solutions quantitatively because all air entering the system is first bubbled through a chlorine dioxide solution of the same concentration and temperature as the solution being transferred. In this manner, the air
above the solution being transferred is kept essentially saturated with chlorine dioxide at all times, thus minimizing losses.
-50-
-51APPENDIX III GAS CHROMATOGRAPHY APPARATUS Gas chromatography offered both a fast and an efficient analysis technique for organic materials because both the starting organic compounds and the expected reaction products were volatile. The relatively small amount
of organic starting materials and reaction products which would be present in a relatively large volume of Freon-TF solvent, however, required that the column be of relatively large diameter. Consequently, a gas chromatography
apparatus was constructed which used 1/2-in. inner diameter (i.d.) glass columns capable of handling 0.50-ml. samples is presented in Figure 14. PHYSICAL COMPONENTS The details of the physical components of this apparatus are listed below: 1. Helium Cylinder. The helium was research-grade helium (99.99%) A diagram of this apparatus
obtained from The Matheson Company, Inc. 2. Reducing Valve. The reducing valve was a double dial single stage
regulator attached to the helium cylinder. 3...