The oxidation of carbohydrates with platinum on carbon as
catalyst
Citation for published version (APA):
Dirkx, J. M. H. (1977). The oxidation of carbohydrates with platinum on carbon as catalyst. Technische Hogeschool Eindhoven. https://doi.org/10.6100/IR114072
DOI:
10.6100/IR114072
Document status and date: Published: 01/01/1977
Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:
www.tue.nl/taverne
Take down policy
If you believe that this document breaches copyright please contact us at:
openaccess@tue.nl
THE OXIDA TION OF CARBOHYDRA TES
WITH PLA TINUM
ON CARBON AS CAT AL YST
THE OXIDATION OF CARBOHYDRATES
WITH PLATINUM
ON CARBON AS CAT AL YST
PROEFSCHRIFT
TER VERKRIJGING VAN DE GRAAD VAN DOCTOR IN DE TECHNISCHE WETENSCHAPPEN AAN DE TECHNISCHE HOGESCHOOL EINDHOVEN ,OP GEZAG VAN DE RECTOR MAGNIFICUS, PROF.DR.P.VAN DER LEEDEN, VOOR EEN COMMISSIE AANGEWEZEN DOOR HET COLLEGE VAN DEKANEN IN HET OPENBAAR TE VERDEDIGEN OP DINSDAG 22 NOVEMBER 1977 TE 16.00 UUR
DOOR
JACOBUS MATHIAS HENDAlKUS DIRKX
GEBOREN TE ROGGEL
DIT PROEFSCHRIFT IS GOEDGEKEURD DOOR DE PROMOTOREN: Prof. Drs. H.S. van der Baan
Ter herinnering aan mijn vader Aan mijn moeder
Contents
1. INTRODUCTION 9
1.1 Carbohydratès 9
1.2 Oxidation of monosaccharides; literature data 10
l. 2.1 Non-catalytic oxidation 10 1.2. 2 Catalytic oxidation 11
1.3 Properties and applications of the oxidation
products 14
1.4 The aim and structure of this thesis 16
2. ANALYSIS 19
2.1 Introduetion 19
2.2 Thin layer chromatography and thin layer
electrophoresis 20 2.3 Gas-liquid chromatography 20 2.4 Ion-exchange chromatography 21 2.4.1 Experimental 21 2.4.2 Results 23 2.5 Isotachophoresis 30
J. EXPLORATIVE EXPERIMENTS IN THE BATCH-REACTOR 33
3.1 Introduetion 33
3.2 The catalyst 35
3.3 Experimental
3.4 Mass transfer in the batch-reactor 3.5 Oxidation of gluconic acid
3.5.1 Experiments in which starting procedure A was applied
3.5.2 Experiments in which starting procedure B was applied
39 40 42 43 46 3.6 Oxidation of glucose 52 3.7 Preliminary conclusions and discussion 57
4. TBE ROLE OF OKYGEN IN TBE OKIDATION OF
GLUCOSE AND GLUCONIC ACID 62
62 62 62 4.1 Introduetion
4.2 Oxidation of glucose
4.2.1 The initial reaction rates
4.2.2 The reaction rate as a function of the reaction time
4.2.3 Deactivation of the catalyst 4.2.4 The selectivity
4.3 Discussion
4.4 Oxidation of gluconic acid
65 67 77 78 86
5. PRODUCT DISTRIBUTION IN THE OKIDATION
OF CARBOHYDRATES 93
93 93 5.1 Introduetion
5.2 Oxidation of gluconic acid
5.3 The influence of the pH on the oxidation
rate of glucose and gluconic acid 100 5.4 The oxidation of some other carbohydrates 102
6. THE TRICKLE-BED REACTOR
6.1 Introduetion 105 105 107 109 109 112 6.2 Experimental 6.3 Characterisation of the TBR 6.3.1 The liquid hold-up
6.3.2 Residence time distribution
6.3.3 Mass transfer 115 6.4 Buffering of the reaction mixture 118 6.5 Deactivation of the catalyst 119 6.5.1 Oxidation of glucose 119 6.5.2 Oxidation of gluconic acid 123
6.6 Reaction kinetica 127
6.6,1 Oxidation of glucose 128 6.6.1.1 Differentlal experiments 128 6.6.1.2 Integral experiments 133 6.6.1,3 The rate of deactivation ~37
LITERATURE 143
APPENDIX I STRUCTURE FORMULAE OF THE VARIOUS SUBSTRATES AND REACTION PRODUCTS LIST OF SIMBOLS SUMMARY SAMENVATTING LEVENSBERICHT DANKWOORD 149 151 155 159 163 164
CHAPTER 1
Introduetion
1.1 CARBOHYDRATESCarbohydrates are produced every year in large quan-tities by photosynthesis. The main forms in which they are found in nature are cellulose, starch and sucrose, which all can be hydrolyzed to monosaccharides.
Starch and sucrose are mainly used as food, while cellulose is mainly used for the manufacture of paper and in the form of wood as fuel and construction material. Only a very small part of the available carbohydrates is used as a feedstock for the chemica! industry.
Nowadays the organic chemica! industry is mainly based on oil, natural gas and coal. However, many chemieals can be produced starting from carbohydrates instead of oil, and due to the streng increase of the price of oil it is getting economically more attractive to produce certain chemieals from carbohydrates insteadof oil (1),
Since the appearance of the first report of the Club of Rome (2) the realisation grows that the supplies of oil, natura! gas and coal are not unlimited. Therefore, it is relevant to look for alternative raw materials for the chemical industry. Carbohydrates sometimes can offer such alternatives, the more so because they are produced every year again in large quantities by photosynthesis. In this respect it is interesting to note that during World War II the chemica! industry in SWeden was mainly based on wood
(3). Surveys of the industrially interesting reaction of carbohydrates are given in the literature (4-13).
Furtherrnore, many chemieals can be produced from carbohydrates which are applied or can be applied because of their specific product properties. One class of these chemieals are the oxidation productsof carbohydrates {14).
This thesis deals with the platinum catalyzed dation of some monosaccharides, and especially the oxi-dation of glucose to gluconic acid and glUcaric acid.
0 11 C-H I H- C -OH I HO-C -H I H- C -OH I H- C -OH I H- C -OH I H D-glucose 0
,,
C -OH I H- C -OH I HO-C- H I H- C -OH I H- C- OH I H -C- OH I H D-gluconic acid 0 11 C-OH I H- C -OH I HO-C -H I H- C- OH I H- C-OH I C,,
-OH 0 D-glucaric acid1.2 OXIDATION OF MONOSACCHARIDES; LITERATURE DATA 1.2.1 NON-CATALYTIC OXIDATION
Although this thesis deals with the catalytic oxidation of carbohydrates, a very condensed survey of the literature data on the non-catalytic oxidation is presented as an intro-duction.
Glucose can be oxidised to glucaric acid with nitric acid as oxidising agent (15). Mustakas et.al. (16) studied this process on a pilot plant scale and obtained a yield of 44% (as K-H-glucarate). Truchan (17) obtained a yield of 64-68% by treatment of glucose with ammonia, fellewed by oxidation with nitric acid.
Another possible route for the manufacture of glucaric acid is the oxidation of starch to polyglucuronic acid by N02
(18-20). However, a low yield of glucuronic acid is obtained in the acid hydralysis of polyglucuronic acid (19,21}, so that this process is not suitable for the production of glucaric acid (glucuronic acid can be oxidised to glucaric acid with a selectivity of almest 100% using Pt/C as
The oxidation of xylose to xylaric acid can be carried out as well with nitric acid (22).
The ketose L-sorbose is oxidised with KMn04 after con-verting L-sorbose into 2,3:4,6-di-O-isopropylidene-L-sorbose
(23), and a yield of 75% for 2-keto-L-gulonic acid is obtained.
De Wilt (24) has reviewed the various methods for the oxidation of glucose to gluconic acid. The main process, which is applied nowadays for the production of gluconic acid, is the biochemica! oxidation, and the yearly produc-tion is estimated to be 20,000 ton (4).
1.2.2 CATALYTIC OXIDATION
Platinum and palladiurrt are used as catalyst for the oxidation of aqueous solutions of carbohydrates with oxygen. The pure metal (i.e. platinum black) can be used as catalyst, but in most cases carbon is used as a carrier to obtain a high catalyst surface area.
De Wilt (25,26) studied the kinetica of the oxidation of glucose to gluconic acid with Pt/C as catalyst. Table 1.1 gives a survey of the reaction conditions used by De Wilt.
[G]o [cat] Po T pH
2
mmol/1 g/1 bar oe
25-200 O.l-1.6 0.2-20 25-65 8-11
Table 1.1 Reaation conditions used by De Wilt for the aatalytia oxidation of glucose.
Pt/C
%
I. 5-10
The maximum selectivity for gluconic acid was about 95%, which was obtained at a low pH (pH=8} or low temperature
(30°C). At a higher pH or temperature the selectivity for gluconic acid was affected by the homogeneaus oxidative degradation of glucose. The rate of the catalytic oxidation
of gluconic acid was about a factor 200 lower than the rate of the catalytic oxidation of glucose under the same reac-tion condireac-tions. The kinetic model, developed by De Wilt, will be discussed in chapter 3.
Heyns and Stöckel (27) oxidised D-glucose, D-mannose, D-galactose, D-xylose and L-arabinose to the corresponding aldonic acids with Pt/C as catalyst. The pentoses were oxidised much more rapidly than the hexoses.
Csürös et.al. (28) used Pd/C as catalyst for the oxidation of glucose, and they obtained a yield of 90% for gluconic acid. Oxalic acid and a
c
4-monocarboxylic acidwere supposed to be formed in the consecutive reactions of gluconic acid. Kinetic conclusions can not be drawn from their experiments, as the mass transfer rate from the gas-to liquid phase has most probably affected their results. Poethke (29) reports that in the oxidation of gluconic acid with Pd/BaS04 as catalyst a number of degradation products are formed: glucaric acid was not detected in the reaction product.
Mehltretter (30) obtained a yield of 54% for glucaric acid (as K-H-glucarate) in the oxidation of glucose with Pt/C as catalyst: thus establishing that the primary alcohol group of glucose can be oxidised to a carboxyl group without cleavage of the molecule.
Also the oxidation of alditols with Pt/C as catalyst to the corresponding dicarboxylic acids has been reported in the literature. Glattfeld and Gershon (31) obtained mannonic-, mannuronic- and mannario acid as reaction prod-ucts in the oxidation of mannitol. Barker et.al, (32) report the formation of galactario acid in the oxidation of galactitol.
Some publications (33,34) deal with the oxidation of glucose to glucuronic acid with Pt/C after blocking the relative reactive aldehyde group. The yield of glucuronic acid, which is obtained in this way, was rather low
(30-40%), so that this processis not suitable for the commercial production of glucaric acid.
under optima! conditions a yield of 65% for 2-keto-L-gulonic acid was obtained. The yield of 2-keto-L-2-keto-L-gulonic acid can be strongly increased by the oxidation of
2,3:4,6-di-O-isopropylidene-L-sorbose foliowed by acid hydrolysis (36).
The literature data on the catalytic oxidation, men-tioned so far, all deal with batch-wise processas and only little information is available on the continuous oxidation of glucose.
Acress and Budd (37) oxidised glucose to gluconic acid and glucaric acid in a trickie-bed reactor. Pd/a-Al 2o 3 is recommended as catalyst if gluconic acid is the desired product, while Pt/C should be more suitable for the oxida-tion of glucose to glucaric acid.
De Wilt (25) used Pt/C for the oxidation of glucose to
gluconic acid in a trickle-bed reactor, and he obtain~d a selectivity for gluconic acid of 90% at a conversion of 70%. Doornbos (52) studied the oxidation of glucose in a trickie-bed reactor with Pd/a-Al2o3 as catalyst. A selectivity for gluconic acid of almost 100% was obtained at a conversion of
70%. Higher converslons resulted in a lower selectivity due to consecutive reactions of gluconic acid. Comparison of the results of De Wilt and Doornbos shows that the activity/g noble metal is about a factor 10 higher for Pt/C than Pd/a-Al
2
o
3 as catalyst, so that Pt/C seems to be more attractive than Pd/a-Al2o3 for the oxidation of glucose in a trickle-bed reactor.Okada et.al. (38) oxidised glucose with Pd/Baso4 as catalyst in a continuously stirred tank reactor and a multi-stage gas-liquid contactor. The conversion was kept below 30% and an empirica! mathematica! description, in which the catalyst deactivation is taken into account, is given.
The oxidation of glucose to gluconic acid on a large commercial scale by catalytic air oxidation has been re-ported (39), but no specific information on this process is available.
are of direct interest for the work described in the pres-ent thesis. Comprehensive reviews of the literature data on the catalytic oxidation of a variety of carbohydrates and their derivates have been given by Heyns et.al. (40,41).
1.3 PROPERTIES AND APPLICATIONS OF THE OKIDATION PRODUCTS The main product property of the oxidation products of monosaccharides is their sequestering capacity: the ability to farm soluble complexes with metal ions. Other applications are the use in the food industry and as a raw material for pharmaceutical products. The oxidation product of L-sorbose, 2-keto-L-gulonic acid, is easily converted inti ascorbic acid (vitamin C) and this process is used as such for the manufacture of vitamin
c.
In a strong alkaline medium gluconic-, glucaric-, galactonic-, galactaric- and xylaric acid are good seques-trants (42,43). Gluconic acid is the only sugar acid which is applied as a sequestrant on a large commercial scale. The sequestering capacity of glucaric acid is about a factor 5 higher than the sequestering capacity of gluconic acid
(3% NaOH, 25°C) (42). However, the relativa high cast of production of glucaric acid by nitric acid oxidation of glu-cose has discouraged the widespread use of glucaric acid as a metal ion sequestrant (16).
In a weak alkaline medium the sequestering capacity of the sugar acids is low, but by complexing the sugar acids with boric acid the sequestering capacity in a weak alkaline medium is strongly increased (44,45). For example, the sequestering capacity of sodium gluconate increases by a factor 13 by complexing it with 1 mol boric acid/mol gluconate (pH=9.5, 25°C) (44).
The most important sequestrants, which are applied nowadays, are citric acid, ethylene diamine tetra-acetic acid (EDTA), nitrilo acetic acid (NTA) and sodium tri-polyphosphate. The choice of the sequestrant for a specific application depends, among other factors, upon the type of
ion to be complexed, pH and temperature. Table 1.2 9ives the sequesterin9 capacity (defined here as 9 ca2+ complexed per 100 9 sequestrant) for a number of sugar acids and the four other sequestrants. The data are adopted from Heesen
(44) and Van der Steen(46).
sequestrant lmol boric acid/ sequestering capscity !mol sequestrant g Ca /100 g sequestrant 2+
(pH=9.5, 25°C) sodium gluconate 0 0.5 I 6.5 disodium glucarate 0 1.8 I 16.5 disodium galacterate 0 1.3 l 14.0 L(+)sodium tartrata 0 2.0 I 8.8 trisodium citrate 0 5.5 l 5.2 Na-NTA 0 16.2 K4-EDTA 0 4.9 sodium tripolyphosphatl 0 12.3
-Tab'le 1. 2 Sequestel'ing aapaaity towards Ca2+ for some sequestrants.
From table 1.2 is concluded that the sequestering capacity of the su9ar acids/boric acid complexes can compete with the sequestering capacity of the more conventional seques-trants.
Sodium tripolyphosphate is used on a large scale
in synthetic detergents, but it causes serious environmental pollution. Therefore, much research work is carried out by various industrial laboratories to find alternatives for sodium tripolyphosphate (47,48). The main factors, that are important for the judgment of potential alternatives, are sequestering capacity, bio-de9radibility, toxicological- and
cancerous properties and the price. The complexes of sugar acids with boric acid can meet at least some of these re-guirements, but with the available information we can not fully judge whether the sugar acids can compete with sodium tripolyphosphate or the proposed alternatives (47,48).
Besides as alternative for sodium tripolyphosphate, the sugar acids and their complexes with boric acid are poten-tially alternatives for the other seguestrants. This is very much a matter of price, and the choice of a sequestrant for a specific application can be strongly influenced by the fur-ther changes in production costs, which could be a drawback for oil-based products.
1.4 THE AIM AND STRUCTURE OF THIS THESIS
The literature data on the catalytic oxidation of carbohydrates deal for the major part with a phenomeno-logical description of some processes (40,41}: the only study of the influence of the various parameters on the reaction rate and product distribution is given by de Wilt
(25,26). He studled the kinetica of the oxidation of glucose to gluconic acid with Pt/C as catalyst, and the present thesis is partly a continuatien of that work. The major aim of this study is to obtain more insight in the kinetica of the oxidation of glucose and gluconic acid with Pt/C as catalyst.
Two analytica! systems, based on ion-exchange chro-matography and on. isotachophoresis, were developed to obtain fast and accurate information about the composition of the reaction samples (chapter 2).
De Wilt bas developed a kinetic model for the oxidation of glucose with Pt/C as catalyst. In our study of the
kinetica of the batch-wise oxidation of gluconic acid with Pt/C as catalyst, we tried to explain our results with the same model as proposed by de Wilt for the oxidation of glucose. However, we found that oxygen plays a most important and
cu-rious role in the oxidation of gluconic acid, which is not explained by the De Wilt model. Oxygen proved to play a very important role in the oxidation of glucose as well, and a new reaction model for the oxidation of glucose and gluconic acid is proposed (chapter 3 and 4).
The influence of some process parameters on the product distribution in the oxidation of gluconic acid was investiga-ted. Besides, some attention was paid here to the oxidation of some ether carbohydrates (chapter 5).
Finally, the oxidation of glucose and gluconic acid in a trickle-bed reactor was studied. Especially on this field, hardly any information is available in the literature. At-tention was paid here to kinetic aspects and the deactivation of the catalyst (chapter 6).
CHAPTER 2
Analysis
2.1 INTRODUCTION
In a kinetic study a fast and accurate analysis of the various reaction products is an essential requirement. The quantitative and even the qualitative analysis of the various reaction products has received only little atten-tion in the literature on the catalytic oxidaatten-tion of carbohydrates. Therefore, much attention was paid to this subject.
During the catalytic oxidation of D-glucose, the following reaction products can be formed:
- D-gluconic-, D-glucuronic-, L-guluronic- and D-glucaric acid
- c
1 -c
5 mono- and dicarboxylic acids- keto-acids
- D-fructose and D-mannose (due to isomerisation)
Thus, the composition of reaction samples can be quite com-plicate. However, depending upon the information required in a specific experiment, not in each experiment a complete analysis is necessary: the determination of a few components will suffice in most cases.
Among ethers, the following techniques can be applied for the analysis of monosaccharides and their oxidative reaction products:
- thin layer chromatography (TLC) and thin layer electrophoresis (TLE)
- gas-liquid chromatography - ion-exchange chromatography - isotachophoresis.
2.2 THIN-LAYER CHROMATOGRAPHY AND THIN-LAYER ELECTROPHORESIS
Pe Wilt (25) studied the use of TLC for the analysis of the reaction products of the homogeneaus oxidative dag-radation of glucose, but the results were poor. Because the composition of the samples of the catalytic oxidation of especially gluconic acid are even more complicate, no further attention was paid to this method.
TLE has been used for the analysis of glucaric acid in the early stage of the study of the oxidation of
gluconic acid (49,50). A separation between gluconic acid and glucaric acid can be obtained, but this technique is time consuming and the reproducibility moderate
<±
15%). The various side products of the oxidation of gluconic acidcould not be separated by this technique, and possibly some of the side products were detected tagether with glucaric acid. Large deviations between GLC and TLE (49) were found in the analysis of gluconic acid in reaction samples. This was explained by the assumption that some of the reaction products were not separated from gluconic acid by the TLE procedure that was used. Consequently, TLE has nat been used in the experiments described in the present study.
2.3 GAS-LIQUID CHROMATOGRAPHY
Verhaar and De Wilt (51) developed a GLC procedure for the separation of glucose, gluconic acid and the oxidative degradation products of glucose. A somewhat revised proce-dure (52) was used for the separation of glucose and gluconic acid in the early stage of the present study. Good results were obtained, provided that the analytical procedure was stricktly standardized, but the GLC procedure is very time consuming and can nat be adapted for automation.
Samples of the catalytic oxidation of gluconic acid gave very complex chromatograms with a number of unidenti-fied peaks (49,50). Glucaric acid could not be determined by
this GLC procedure because only very low and non-reproducable signals were obtained. This was probably caused by incom-plete silylation and/or breakdown of the silylated compo-nents on the column. Several attempts to analyze glucaric acid by various GLC procedures were not succesfull (49), and therefore we did not pursue this approach for the analy-sis of glucaric acid and the other reaction products any further.
2.4 ION-EXCHANGE CHROMATOGRAPHY
Samuelsen et.al. frequently reported the separation of respectively aldonic acids (53-58), alduronic acids (54-56) and aldaric acids (59-63) by ion-exchange chromatography. The separation of the acids is mostly aarried out on a column with a strongly basic anion exchange resin (i.e. Dowex 1-X-8) and with salt solutions (acetate, sulphate, phosphate,
borate) as eluant. Most literature data deal with the separa-tion of mixtures which contain only one class of compounds: literature data on the analysis of samples which contain both aldoses, aldonic-, alduronic- and aldaric acids are not
available. We developed some analytica! systems for the analy-sis of glucose and its main oxidation products (gluconic-, guluronic-, glucuronic- and glucaric acid) and its main iso-merisation products (fructose, mannose) by ion-exchange chromatography.
2.4.1 EXPERIMENTAL
The analytica! system was based on the Technicon Auto-Analyzer and a block diagram of the self-asse~hled
system is given in figure 2.1.
The eluant, which was degassed by heating to a temper-ature of 95°C, was pumped with an Orlita membrane pump
(type PMP1515l. For the !njection of a sample, a Chromatranix injection valve (type SVA8031} with pneumatic actuator was
used. The sample loop (volume 10 ~11 was filled with the sample by pumping it from a sample bottle through the sample loop of the injection valve with a peristaltic pump.
precolumn reactîon thermostal electron ie integrator I
ËJ
colorimeterFigure 2,1 Btook diagram of the Ziquid ahromatograph,
A precolumn (15 cm x 4 mm), which was filled with a cheap ion-exchange resin (Aminex AG-1-X-8 from Biorad), was mounted between the eluant pump and the injection valve to remove any impurities that might be present in the eluant. The separation column, heated by water circulation from a thermostat, was slurry packed after suspending the
ion-exchange resin in 1 M NaCl at the required temperature. An adjustable spindle was mounted on the top of the column to reduce the dead volume to a minimum. Aminex A-25 (particle diameter 17.5
±
2 ~m) and Aminex A-27 (particle diameter 13.5±
1.5 ~m) were used as ion-exchange resins (styrene-divinylbenzene resin, 8% crosslinking, functional group: quaternair ammonium -R4N+, ion-exchange capacity:3.2 meq/g dry weight).
Continuous detection taak place by reaction of the eluted components with either H2
so
4;K2cr2o
7 reagent(composition: 40 v/v% H2
so
4; 60 g K2cr207/l) or orcinol(3,5 dihydroxy toluene) reagent (composition: 70 v/v%
H2
so
4 ; 1 g orcinol/1). The reagent flow (3.35 ml/min) was segmented with air bubbles (0.42 ml/min) and justbe-fore detection the reaction stream was debubbled. The reaction taak place at a temperature of 95°C and the reac-tion time was, unless otherwise stated, 10 minutes. The reagent, segmentation air, cuvet stream (1.25 ml/min) and the sample were pumped by a Technicon peristaltic pump
(type PPI). The äetection was performed with a Technicon Single Channel colorimeter {at 420 nm for the orcinol reagentand at 600 nm for the chromic acid reagent). The signals were recorded and the peak areas were measured with an electronic integrator (Infotronics, type CRS 204).
The analytica! system was adapted for continuous
analysis. The automation basically consisted of the rotating disc of a fraction collector which could contain 150 sample bottles. A needle with a pneumatic activator (Martonair) was connected by a teflon capillary with the peristaltic pump via the sample loop of the injection valve. By means of a time clock and a number of programme discs the various operations, which are involved in the injection of a sample, were carried out periodically.
2.4.2 RESULTS
analysis of samples containing glucose, gluconic acid and glucaric acid, and a chromatagram is given in figure 2.2
(concentration of each component: 100 mmol/1}.
eluant eluant flow ion-exchange resin column dimensions column temperature detection O.J6 M Na2
so
4 0. 75 ml/min Aminex A-27 25 cm x 4 mm 75°C chromic acidTable 2.1 Analytiaal aonditions for the separation of glucose, gluconic acid ànd glucaria acid.
2
0 15 20 25 30 35
-time(minutes)
Figure 2. 2 Chromatogram of a sample aontaining glucose (11~
gZuaonic acid (21 and glucaric acid (3J,
A good separation is obtained and the time required for one analysis is about 35 minutes, which is reduced to about 20 minutes in the case of continuous automatic analysis.
Figure 2.3 gives the calibration curves (peak area A versus concentration) of the three components. As is seen in figure 2.3, a non-linear relationship between peak area and concentratien is obtained. The deviation from linearity can
Cl Qlucose
0 gluconic acid
0 glucaric acid
20 60 80 100
--conaenlratiOn(mmol/1)
Figure 2.3 CaZibration curves of gZuaose. gZuconia
acid and gZucaric acid.
be ascribed to the formation of
co2
during the"oxidation with chromic acid, The following relation can be used to calculate the concentratien of the various components in a reaction sample:c
=
a.Ab where:C = concentration, A
=
peak area, and a and b are constants which can be calculated from the peak areas of standard samples with known concentrations.Figure 2.4 gives a chromatagram of a reaction sample of the oxidation of gluconic acid; the conditions are those given in table 2.1. Besides gluconic- and glucaric acid, the reaction sample contains a number of other components. The keto acids 2- and 5- keto gluconic acid elute àfter gluconic acid. In most experiments the concentratien of the keto acids was only very low, so that they were not measured
quantitatively. Following glucaric acid, a number of other dicarboxylic acids elute: xylaric- + arabinaric acid, tartaric acid and tartronic acid. No signal was obtained of oxalic acid. The concentrations of the dicarboxylic side products were not measured by ion-exchange chromatography.
3 2
0 15 20 25 30 35 40
-time(minutes)
Figure 2.4 Chromatogram of a sample of the aatalytia oxidation of gluaonia acid 1: gluaonia- + guluPonia acid, 2: keto gluaonia aaids, 3: gluaaria acid, 4:
xylaria-+ aPabinaria acid, &: taPtaria acid.
Both the concentratien and rnalar response of these components are rather low, and to obtain a full separation between the side products, the time required for one analysis will in-crease drastically. As isotachophoresis (section 2.5) could be used successfully for the analysis of the side products we have abandonned ion-exchange chromatography for these components. Erythronic- and arabonic acid elute slightly after gluconic acid (not shown in figure 2.4). The concen-tratien of these components is mostly very low, so that they hardly interfere the measured concentratien of gluconic acid. Glycolic- and glycerinic acid elute between gluconic-and glucaric acid (not shown as well in figure 2.4). In most experiments the concentratien of these components is rather
low and because their molar response is low, only very low signals of these components are obtained,
L-guluronic acid, the intermediate product in the
3
0 15 20 25 0 15 20 25
-time(minutes) -time(minutes,
Figu:re 2. S Chromatogram of a sample Figure 2. 6 Chromatogram of a sample
of the aatalytia oxidation of gZu- of the aatalytia oxidation of
gZu-aonia aaid with orainoZ aa reagent aoae with orainoZ aa reagent. 1: guluronia aaid~ 2: 5-keto gluao- 1: gZuaoae, 2: guZuronia aaid,
nia aaid. 3: gZuauronia acid.
oxidation of gluconic acid to glucaric acid, elutes almost simultaneously with gluconic under the conditions given in table 2.1. A separation between gluconic- and guluronic acid can be obtained by the use of NaAc as eluant (56). This agrees with our experiences, but glucaric acid elutes after a long time when NaAc is used as eluant. Orcinol reagent reacts highly specifically with uronic acids, and gives no signal with aldonic- and aldaric acids. Thus, orcinol reagent can be used for the analysis of guluronic aèid in the pres~
ence of gluconic acid. Figure 2.5 gives a chromatagram with orcinol as reagent for the same sample (after diluting 1:4) as in figure 2.4. The eluant flow was reduced to 0.6 ml/min: the other conditions were the same as given in table 2 .1. A strong signal of guluronic acid is obtained, and the second signal has most probably to be ascribed to 5-keto gluconic acid, which gives a strong signal with orcinol reagent as well. A linear relation between peak area and concentration was obtained in the use of orcinol reagent.
The concentratien of gluconic acid in a sample was calculated from the total signa! obtained with chromic acid and the guluronic acid concentratien as measured with
orcinol.
In the oxidation of glucose, L-guluronic- and D-glucu-ronic acid are the main side products. Under the conditions given in table 2.1, the uronic acids are separated from glucose, but the separation between guluronic- and glucuro-nic acid is very poor. A better resolution is obtained by the use of NaAc as eluant. Figure 2.6 gives a chromatagram of a reaction sample of the catalytic oxidation of glucose
(eluant: 0.1 M NaAc, eluant flow: 0.6 ml/min, detection: orcinol reagent, other conditions as given in table 2.1), and a reasonable separation between glucose, guluronic- and glucuronic acid is obtained.
Under certain reaction conditions (especially high pH and temperature) fructose and mannose are formed by isomeri-sation of glucose. Boric acid is most frequently used as eluant for the seJ?aration of sugar mixtures with ion-exchange chromatography (64-681. Most ltterature data deal with the separation of relattve complex mixtures of sugars, result!ng in long separation times between the various components. We developed an eluant which gives a fast separation between glucose, fructose and mannose, and the con di t i ons which we re applied are gi ven in table 2. 2.
Figure 2.7 gives a chromatagram of the three augars glucose, fructose and mannose, showing that a good resolution between the three augars is obtained.
eluant eluant flow ion-exchange resin column dimensions column temperature detection reaction time 0.38 M H3Bo3, 0.024 M Na 2B4o7, 0.01 M NaCl, 0.01 M HAc 0. 75 ml/min Aminex A-27 15 cm x 4 mm 75°C orcinol reagent 4 minutes
TabZe 2.2 Conditions appZied for the separation of
gZuaose~ jruatose and mannose.
2 3
0 5 10 15 20 25 30 35
- time(minutes)
Fig~ 2.( Chromatagram of a sample aontaining mannose (1).
fructose (2) and glucose (3).
In the routine automatic analysis four standard samples of different concentrations, each containing the components of interest, were used, and after each 15 reaction samples the standard samples were injected again. A good reproduci-bility was obtained in the automatic analysis: the deviation between standard samples over a period of 24 hours was less than 5%, provided that the various flows and temperatures were kept constant over this period.
2,5 ISOTACHOPHORESIS
Isotachophoresis is most usefull for the separation of ionic components. The principle and theoretica! backgrounds of isotachophoresis have been described by Everaerts et.al.
(69,70) and will notbedealt with here. The use of iso-tachophoresis for the analysis of the oxidative degradation products of sugars has been described by Pe Wilt (25) and by Everaerts and Konz (71}. The conditions, used in the present study, are given in table 2.3 and a conductivity detector, as described by Everaerts and Verheggen (72), was used for detection.
leading electrolyte terminator
capillary current strength injection volume time for analysis
0.01 N Cl- (pH ~ 6.02} 0 •. 005 M morfoline ethane sulphonate (pH= 6.10} teflon: 20 cm x 0.5 mm 100 !JA 1 ).1]. (after diluting I :4} 10 - 15 minutes
TabZe 2.3 Conditions used for the anaZyaia by iaotaahophoreeis.
A typical isotachopherogram of a reaction sample of the catalytic oxidation of gluconic acid is given in figure 2.8
(I= integral signal, D = differentlal signa!}. The quali-tative information is given by the height of the integral signa! while the quantitative information is given by the distance between two peaks of the differentlal signa!.
Gluconic-, guluronic- and keto- gluconic acids were not separated under the conditions given in table 2.3:
the concentrations of these components were measured by ion-exchange chromatography. The signal belonging to
c
4 dicar-boxylic acids most probably consists of erythraric- and threaric acid, and is referred to as tartaric acid. Xylaric-and arabinaria acid were neither separated under the abovementioned conditions. Therefore, the signal beloninging to
c
5 dicarboxylic acids is referred to as xylaric- + arabi-nario acid, as both isomers are supposed to be formed in the oxidation of gluconic acid.-
8 7 1 10I\
4 3 2 5 9 6r'Ln-fi
U~n
I----
""
DFigure 2.8 Isotaahopherogram of a sample of the aatalytia oxidation of gluaonia aaid 1: oxalia aaid~ 2: tartronic acid~ 3: tartaria acid. 4: :x:ytaria- + arabinaria acid, 5: gluaaria acid, 6: glycol-ia acid, 7: glyaerinia acid~ 8: e~thronia acid .. 9: arabonia acid. 10: gluaonia-. guZuronia- and keto gZuaonic acids.
CHAPTER 3
Explorative experiments in the batch-reactor
3.1 INTRODUCTION
De Wilt (25) developed a kinetic model for the catalytic oxidation of glucose, which was based on the dehydrogenation mechanism (73,74}.
Pt H I H I + R- C- OH I OH Pt. (R-C-OH} I OH H I R- C- OH. I OH H I Pt.(R-C-OH) I OH 0 11 (hydratation) {adsorption) Pt, H2 + R- C- OH (dehydro-genation)
---;>,..
Pt + H 20 (oxidation}A peroxide mechanism has been proposedas well (25,40), but the dehydrogenation mechanism is accepted by most authors. An analoguous mechanism has been proposed for the
Pt-catalyzed oxidation of ethanol (74}. The oxidation of gluco-nic acid to glucaric aci.d is supposed to occur via the inter-mediate product L-guluronic acid,
+ (dehyd:rogena-tion)
The aldehyde L-gulu:ronic acid wil! be oxidised acco:rding to the sarne mechanism as used for the oxidation of glucose. Rottenberg and ThU:rkauf (75) concluded f:rom tracer experi-ments that the Pt-catalyzed oxidation of ethanol and 2-propanol does not involve :reversible dehyd:rogenation as an initia! step. They assume that a simultaneons activatien of the entering oxygen molecule and the leaving hydragen atoms occurs on the catalyst surface.
De Wit (76) found that at a high pH (pH
=
13.5) glucose is dehydrogenated in a N2 atmosphere to gluconic acid and H2• This reaction occurs most p:robably according to the above mechanism, but under our reaction conditions (pH 7-11) the rate of the latter reaction is a factor 103 - 10 4 lower than the rate of the reaction in ano
2 atmosphere.De Wilt described the oxidation of glucose as a paral-lel system of two monomolecular reactions on two different types of active sites on the catalyst surface. The kinetics of the main reaction (glucose + gluconic acid) was descriped
by a first order model in which a streng adsorption of both glucose and gluconic acid on the catalyst surface was
supposed. r
= -
.!![§]_ = dt [ G] [G] + [GOZ] • [ cat] (3.1}The reaction rate constant kr increased with increasing [cat], and decreased with increasing [G]
0 • This was ex-plained by the assumption that the added glucose contained some poison, which deactivates a part of the catalyst surface.
The rate of the proposed side :reaction (g.lucose +
glucaric acidl decreased strongly in the beginning of an experiment, because the active sites concerned were
sup-posed to be deactivated by the streng adsorption of some reaction product.
An analoguous model as that of De Wilt for the cata-lytic oxidation of glucose can be postulated for the catalytic oxidation of gluconic acid.
r = d[GOZ]
dt
=
rsite 1 + (3.2)Site 1 represents the active sites which are deactivated by SOOE reaction product, and the main reaction is supposed
to occur on site 2. This two site model was our starting point for the investigation of the catalytic oxidation of gluconic acid. Results of some explorative experiments are discussed in this chapter.
3.2 THE CATALYST
Van Laarhoven (77) studied two procedures for the prep-aratien of Pt/C catalysts for the oxidation of glucose: the procedure of Heyns (35) and the procedure of Zelinskii
(78,79). The best results were obtained by the latter method, and our procedure for the preparatien of the catalyst was based on the procedure of Zelinskii. This procedure for the
preparatien of a 5% Pt/C catalyst was as fellows: A salution of 10 g hexachloroplatinum acid
(H2PtC16 .6H20) in 100 rnl-H2
o
was added to 72 g active carbon(Norit PK 10x30), and about 100 rnl H
2
o
was added to obtain a complete wetting of the carbon. During the adsorption of platinum acid on the active carbon, which took place at room temperature, nitrogen was bubbled through the suspen-sion. After a period of 5 hours the adsorption equilibrium was reached, and the suspension was caoled to a temperature of 0°C. After the addition of 170 ml 35% formaldehydesolution, the platinurn acid was reduced to platinurn metal by the slow addition of 90 ml 30% KOH over a period of 1 hour. After the addition of the KOH salution and standing
during a night, the catalyst was filtered off and washed with distilled water until the filtrate was neutral. After drying at a temperature of 50°C the catalyst was sieved to remove the fine powder. If the catalyst had to be used in a batch-wise experiment, the catalyst was ground in a mortar to obtain a fine powder. If the catalyst had to be used in a trickle-bed reactor, the active carbon was ground to the required sieve fraction before the preparation was started.
The platinum content of the catalyst was determined by dissolving the platinum in aqua regia, complexing it with SnC1 2 and measuring the extinction of the solution at 403 nm (80-82).
Some Pt-loss (10-20%) takes place during the prepara-tion of the catalyst, what is probably caused by the re-versible adsorption of H2PtC16 on active carbon. During the reduction, the H2PtC16 in solution is reduced as well. This causes desorption of a part of the adsorbed H
2PtC16•
The above procedure gave a catalyst with a reproducable activity. Figure 3.1 gives the initial reaction rate in the
10
Figure 3,1 Catatyat aativity as a funation of the Ft-aontent.
oxidation of glucose (pH=lO, T=55°c,
[Gl
=100 mmol/1, 100% 0o
2 )as a function of the Pt-content of the catalyst. A linearob-tained, which enables a comparison of results obtained with different catalysts.
In some experiments, especially for the oxidation of gluconic acid, a relative large amount of catalyst is re-quired so that it was unpractically to use a freshly pre-pared catalyst for each experiment, Therefore, the used catalyst had to be regenerated. This succeeded fairly well by washing it with 3 1 hot water (90°C). If the catalyst was used for a long period, this procedure did not result in a complete regeneration. A better, although still in-complete, regeneratien was obtained by washing the catalyst with 0.1 N HCl. The souree of this small irreversible deactivation is not completely clear. In any case, it was not caused by Pt-loss during prolonged use, Possibly, this deactivation is caused by some (unknownl reaction product or is related to the formation of Pto2 during an experiment
( chapter 4} •
The catalysts were not characterized by measuring the Pt-surface by o2 or H2 titration. This technique will give information on the total metal surface, but not on the metal surface available for the oxidation reaction. De Wilt (25)
measured the Pt surface of a number of Pt-catalysts, and large deviations (factor 2-3) in activity between catalysts with almost the same Pt-surface was found. The variatien in activity/m2 Pt of catalysts with different Pt contents was even much higher (factor 15-20).
Some catalysts were used for bath the hydragenation of fructose (83) and the oxidation of glucose. Catalysts which showed a large difference in activity in the oxidation
reaction (catalysts prepared according to the above procedure and catalysts supplied by Johnson Matthey) showed only a
small difference in activity in the hydragenation of fructose. Besides, a strong irreversible deactivation of the catalyst was observed in the hydragenation of fructose. However, a catalyst which had been strongly deactivated in the hydrage-nation reaction, showed only a minor deactivation in the oxidation reaction. Thus, the total metal surface is not a good yardstick for the catalyst activity in these reactions.
w ()0 ~
1
~ ~ b:l 1:)"'"
t
~-\1) ~ ~<i-"'·
~ 0> ~<i-~
water burette OI'P eenü
I 1D
recorder I I I~--0
power relais oxygen pump 0:2-out temperafure controller0
:~--- ~
~--
j
I I I stirringmu
reactor KOH burette magnette valve---L----
--o
pH c~troller Ir--0
1 racdrder I I--~--0
oxygen analyzer3.3 EXPERIMENTAL
A sketch of the batch-wise reaction system is given in figure 3.2. The reaction mixture was stirred with a six-bladed turbine stirrer at a rate of 2000 r.p.m. The pH of the reaction mixture was controlled by automatic titration with 4N KOH by means of a Radiometer TTT ld titrator. The oxygen consumption during an experiment could be measured by cantrolling a closed gas circuit at a constant pressure, and the oxygen concentratien in the liquid phase could be measured with a Beckmann Fieldlab oxygen analyzer. The oxygen supply could be effected in two ways: either via a capillary in the liquid or via a sintered bottorn plate.Rela-tively small gas bubbles are created by the latter method, thus increasing the mass transfer rate from the gas- to liquid phase.
The experimental procedure was as follows: The required amount of catalyst + 450 ml water were heated in the reactor to the required temperature. Then, 50 ml of a concentrated substrate solution, which was at the required temperature and pH, was added. During an experiment a number of samples
(about 5 ml) were taken with a syringe. After sampling, the catalyst was filtered off, and stared in a refrigerator un-til analysis. The measured concentrations were corrected for the dilution with KOH and sampling.
Two different starting procedures were used: Starting proeedure A:
The catalyst suspension was brought to the required temperature in the oxidation gas atmosphere (gas flow 1 1/min) , and the experiment was started by introducing the concentrated substrate salution into the reactor.
Starting proeedure B:
The catalyst was brought to the required temperature in a nitrogen atmosphere. After introducing the concentrated substrate salution into the reactor, the suspension was kept in a nitrogen atmosphere for 10 minutes. Thereafter, the nitrogen flow was stopped, and the experiment was
started by introducing oxygen or the oxygen containing gas at a flow of 1 1/min.
3.4 MASS TRANSFER IN THE BATCH-REACTOR
In the three phase system gas-liquid-solid the following mass transfer steps can influence the reaction rate (84): 1. Oxygen transfer from the gas- to the liquid phase
2, Ditfusion of reactantsfrom the bulk of the liquid phase to the surface of the catalyst
3. Pore diffusion of the reactants.
The initial reaction rate as a function of the amount of catalyst is given in figure 3.3 for the oxidation of
2 4 6 8 -icatl{g/1)
Figure 3.3 InitiaL reaction rate in the oxidation of gLucose versus cataLyst concentration.
0
glucose (pH·=lO, T=55 C, 100% 0 2, 5% Pt/C, [G]0=100 nunol/1,
starting procedure B). A linear relation between initia! reaction rate and the amount of catalyst is obtained, so that is concluded that step 1 does not influence the reac-tion rate under the above menreac-tioned condireac-tions. This is in agreement with the observation that the reaction rate is not increased by increasing the gas flow or the stirrer speed. Invsome experiments, oxidation of gluconic acid at a low %
o
2 , the reaction rate was influenced by step 1,which resulted in an increase of the reaction rate. This phenomenon will be discussed in section 3.5.
The diameter of the catalyst particles in the batch-wise experiments was about 50 ~m. As is seen in figure 3,4
o pOwdered catalyst • granulee
20 40 60 80 100 120
-time(minutesl
Fig~ 3.4 Oxidation of gZuaose with powdered aataZyst and granuZes.
almest the same reaction rate is obtained with catalysts with an average partiele diameter of respectively 50 and 520 vm (both 20 9 0.30% Pt/C of the same batch of catalyst, ether conditions the same as used for experiments in figure
3. 3) •
Catalysts with a Pt-content of <10% have been used in the batch-wise experiments. The mass transfer rate from the bulk of the liquid phase to the external catalyst surface is proportional to the external catalyst surface. From fig-ure 3.4 it is concluded that the reaction rate is net affected by step 2 in the batch-wise experiments.
The Thiele modulus is defined as fellows:
where: K
=
a constant, dp=
partiele diameter, kr=
the reaction rate constant, Cs=
surface concentratien of the reactant and n the order of the reaction.The reaction rate is proportional to the platinum content of the catalyst ( [Pt] ) • kr
=
k' r .JPt] (3.4) ~=
K d Vk' IPt] cn-l • p r' • s (3.5) dp 520 ]Jin, [Pt] 0.30%dp.~
=
285 d=
50 ]Jin, [Pt] 10%dp·~·
=
158. pThus, the Thiele modulus in the experirnents with powdered catalysts was smaller than the Thiele modulus in the experi-rnents with granules in figure 3.4. Frorn this and figure 3.4 it is concluded that the pore diffusion rate did not influ-ence the reaction rate in the batch-wise experirnents.
3.& OXIDATION OF GLUCONIC ACID
Unless otherwise stated, the standard reaction condi-tions as given in table 3.1 were used for the catalytic oxidation of gluconic acid.
[GOZ] [ cat
J
V % 02 T pH 0 0 mmo1/1 g/1 1 in oxidation oe gas 200 40 0.500 100 55Tabte 3.1 Standard reaction conditions for the cataZytic oxidation of gtuconic acid.
10
Pt/C %
5
The rate of the non-catalytic oxidation of gluconic acid is negligibly low under the above rnentioned conditions.
3.5.1 EXPERIMENTSIN WHJCH STARTING PROCEDURE A WAS APPLIED
Figure 3,5 and 3.6 give a first order presentation of the influence of
[GOZ]
0 and [cat] on the oxidation rate of
2 3 4 5 2 3 4 5
-time(hours) -time(hours)
Figure 3.5 Oxidation of gtuconia aaid at various ~Zj
0•
Figure 3.6 Oxidation of gluaonia aaid at varioua [aat] .
gluconic acid (starting procedure Al. After an initial peri-cd, in which a relative streng decrease of the reaction rate occurs, the experiments are described by a first order model.
r = _ d.[GOZ] =
dt kr·IGOZ] ,[cat] (3.6)
The initial deviation from the first order model can be ex-plained by the two-site model as proposed by De Wilt for the oxidation of glucose: in the beginning of an experiment the reaction takes place on both types of active sites, but
during reaction some product is formed which adsorbs strongly on one type of the active sites (site 1), so that aftersome reaction time (0.5 - 1 hour) the reaction mainly 'occurs on site 2. The first order behaviour of the reaction on slte 2 suggests a weak adsorption of gluconic acid on site 2. This
is in disagreement with the model of De Wilt for the oxida-tion of glucose, in which a streng adsorpoxida-tion of gluconic acid on the catalyst surface was supposed. This contradietien will he discussed in a later stage.
De Wilt reports an order of 0.2 - 0.3 in oxygen in the oxidation of glucose, and this is in agreement with the dehydrogenation mechanism. In a patent of Hoffman La Roche
(36) it is reported that the catalyst is deactivated after one experiment if pure oxygen is used as oxidation gas for the oxidation of diacetone-L-sorhose. On the other hand, the activity of the catalyst was preserved in prolonged use if an oxidation gas with a low 0
2 content was used. We carried out an experiment with air instead of 100%
o
2 as oxidation gas (starting procedure A), and we found that the use of200 50 2
•
3 o oxygen A atr 4 5 - time(hours) 6 7Fi~e 3.7 Oxidation of aZuconic acid with air and oxygen as oxidation gas.
air as oxidation gas results in a much higher reaction rate (figure 3.7). This can not he explained in termsof the dehydrogenation mechanism: the rate of the oxidation of hyd-rogen is supposed to he fast compared to the rate of the other reactions, so that a low positive order in oxygen is expected.
The use of a larger amount of catalyst in the experi-ment with air as oxidation gas did not further increase the reaction rate. This indicates that a mass transfer limita-tion from the gas- to liquid phase occurs under these con-ditions. This was concluded as well from the low expertmen-tal oxygen concentratien in the liquid phase (3 - 5%) in the experiment with air.
The use of 7.5 g of catalyst gave the same reaction rate at 21 and 100%
o
2• Addition of an extra amount of catalyst during an experiment with 7.5 g catalyst in which air was used as oxidation gas, resulted in a more than pro-portional increase of the reaction rate. The reaction rate after the addition of the extra amount of catalyst was even higher than the initial reaction rate/g catalyst. This is illustrated in figure 3.8 in which the eaustic soda
consump-7. 5 g catalyst added 100 ets 07.5 -time(minutes) 200
Figure J.B eaustic soda aonsumption versus reaction time in the o~idation of gluaonia acid with air.
tion is given as a function of the reaction time: the extra amount of catalyst was added aftera reaction time of 125 minutes.
The reaction rate (average conversion over the first hour) as a function of the % 0 2 in the oxidation gas is given in figure 3.9 (15 g 5% Pt/C, starting procedure A).
Over a large range (35 - 100 %
o
2) the reaction rate hardly depends upon the %o
2 in the oxidation gas, while around 20%
o
2 a relative high reaction rate is obtained. The use of a better oxygen inlet system (sintered bottorn plate instead of.06
20 40 60 80 100
- % 02 in oxidation gas
Figure 3.9 Average reaation rate during the firat hour versua oxygen aontent of the oxidation gas.
a capillary) or a lower amount of catalyst results in the same reaction rate at 21 and 100%
o
2, and in this case a lower %o
2 has to be used to obtain an increase of the reac-tion rate. Thus, the reacreac-tion rate increases only at a lower %
o
2 if, at least during the first minutes of an experiment, the oxygen concentratien in the liquid phase is very
low. The oxygen concentratien in the liquid phase does not only depend upon the oxygen concentratien in the gas phase, but as well upon the applied gas inlet system and the reaction rate.
5.5.2 EXPERIMENTSIN WHICH STARTING PROCEDURE B WAS APPLIED The reaction rate does not only depend upon the oxygen concentratien in the liquid phase, but also the starting procedure of an experiment has a most important influence on the reaction rate. By the use of air instead of 100%
o
2 amuch hiqher reaction rateis obtained (fiqure 3.7), but if during an experiment with air the %
o
2 is increased from 21 to 100% a further inarease of the reaction rate is obtained. Thus, the reaction rate does not only depend upon the concen-trations of the various reactants, but also the previous history of an experiment. This is illustrated in fiqure 3.10,o starting proQ!dure A
<> start ing procedure B
a starting procedure A • t •1 h - 1h 10 mill< N2
~1
f
2 3 4 5 6 7
-time(hours)
Figure 3.10 InfLuenae of the starting proaedure on the oxidation rate of gluaonia aaid.
in which the concentratien of gluconic acid as a function of the reaction time is gi ven for both startinq procedure A ar•d B (A: catalyst saturated with 0
2 befare the start of the experiment; B: catalyst toqether with gluconic acid in a N2 atmosphere befare the start of the
o
2 flow). The use of startinq procedure B results in a much higher reaction rate than the use of starting procedure A, while in both experi-ments the experimental conditions are the same. The same reaction rate was obtained when the catalyst was kept to-gether with gluconic acid in a N2 atmosphere during 1 mi-nute instead of 10 mimi-nutes. A reaction rate which was between that of startinq procedure A and B was obtained when theo
2 flow was started immediately after the addition of gluconic acid to the catalyst suspension in a N2 atmos-phere (contact time in the N2 atrnosphere between qluconic acid and the catalyst: about \minute).If starting procedure B is applied, an initially fast reaction takes place immediately after adding the gluconic acid salution to the catalyst suspension in a N2 atmosphere. This reaction stops in less than 1 minute:
guluronic-and glucaric acid (respectively 7 guluronic-and 2 mmol/1) are the main reaction products. Evidently, a reaction takes place between
adaorbed oxygen and gluconic acid. The rate of this reaction is much lower in an oxygen- than in a nitrogen atmosphere: in a sample which was taken after a reaction time of 1 minu-te in an experiment in which starting procedure A was ap-plied, the concentrations of guluronic- and glucaric acid were respectively 2 and 1 mmol/1.
Temporarily stopping the oxygen flow during an experiment by replacing it by a nitrogen flow results in an increase of the reaction rate in case of both starting procedure A and B. This is illustrated in figure 3.10 for the case of starting procedure A.
Obviously, a steady state is not reached during an ex-periment1 but a change in the degree of coverage of the catalyst by oxygen takes place. A low degree of coverage by oxygen results in a high reaction rate.
Oxygen adsorbs very fast and irreversibely on platinum as Pt-0 (85 186). If starting procedure A is applied, the initia! degree of coverage by chemisorbed oxygen (ePt-O) is almest 1, while in the case of starting procedure B, at least initially, 6Pt-O =0.
We measured the rate of adsorption of oxygen on a re-duced catalyst (rere-duced with gluconic acid, 10 g 9.7% Pt/C in 0.500 1 H20, pH= 10, 55°C). The equivalent amount of oxygen
(2.4 mmol) was consumed within about 4 minutes, while the experimental adsorption rate, especially during the first two minutes, can have been affected by the experimental pro-cedure (changing from a nitrogen to oxygen flow).
The influence of the adsorption of oxygen on the reaction rate was established as fellows: The catalyst (10 g 5% Pt/C) was reduced with gluconic acid in a nitrogen atmosphere. Sub-sequently, the catalyst was filtered off in a nitrogen atmos-phere and washed with oxygen free water. This reduced
catalyst was transferred into the reactor. After heating the catalyst suspension to a temperature of 55°C, the catalyst was contacted with oxygen during various times, befare the reac-tion was started by the addireac-tion of a concentrated gluconic acid salution to the reactor content. The results of these experiments are given in table 3.2, in which the initia! re-action rate is given as a function of the contact time T be-tween oxygen and the catalyst befare the addition of gluconic acid.
T r
0
minutes mmol/1 min
0 (s.p. B) 8.0 5 2.5 30 1.4 60 1.1 ISO 0.78 oc. (s.p. A) 0.57
TabZe 3.2 Initial reaation rate as a funation of the aontaat time T between aataZyst and oxygen.
Frorn table 3.2 is concluded that a considerable contact time between oxygen and an initially reducad catalyst is required to deactivate a catalyst by oxygen adsorption. If the adsorp-tion of gluconic acid on the catalyst surface is described by a single-site Langrnuir-Hinshelwood model, we can write for the degree of coverage by gluconic acid, eGOZ:
K
2.[GOZ] 1 + K2 .[GOZ] where:
-1 K2
=
adsorption constant of gluconic acid 1 mrnol aPt-O=
degree of coverage by chemisorbed oxygen(3.7)
gluconic acid, we can state that at a high ePt-o=
(3.8)
Thus, with the above assumptions the react!on rate is proportional to (1-ePt-o>, and this e:xplains the large difference in reaction rate between starting procedure A and B, notwithstanding the fact that oxygen adsorbs
fast on platinum.
The starting procedure does not only influence the reaction rate, it also has a considerable influence on the product distribution. Gluconic acid (GOZ) is oxidised to glucaric acid (GAZ) via the intermediate product
guluronic acid (GLZ): byproducts are formed starting from gluconic- and glucaric acid (chapter 5},
k1
GOZ Jo GLZ ----:>!1-GAZ --...;:;.. ... consecutive products
~side
productsIn figure 3.11 and 3.12 the concentrations of guluronic-and glucaric acid are given as a function of the
concen-.§ ~ ë
"'
g 8l
10 .. guluronic acid o g lucaric acid 100 200 - AIGOZ((mmolll) 100 .. guluronic acid o glucaric acid 100 200 - AIGOZl (mmoltl)Figu:J'e 3.11 {GLZ] and IGAZ] versus Fi{JUI'e 3.12 [GLZ] and [GAZ] Vel'sus [GOZ] (star>ting p:roaedu:J'e A). IGOZ] (starting p:roaedul'e B).
