Bachelor Thesis Scheikunde
Kinetic studies of the CoMeTAA-catalyzed
cyclopropanation of ethyldiazoacetate and methylacrylate
By application of Reaction Progress Kinetic Analysis
door
M.B. Brands
18-01-2016
Studentnummer 10409858 OnderzoeksinstituutVan 't Hoff Institute of Molecular Sciences Onderzoeksgroep
Homogeneous Catalysis
Verantwoordelijk docent prof. dr. Bas de Bruin Begeleider
Populair wetenschappelijke samenvatting
Veel chemische reacties verlopen niet vanzelf. Er kan bijvoorbeeld hitte nodig zijn voordat de reactie verloopt, zoals bij de verbranding van benzine. Omdat warmte een vorm van energie is en er dus een bepaalde hoeveelheid in de reactie moet worden gestopt voordat deze verloopt, wordt er ook wel gezegd dat de reactie een bepaalde energiebarrière moet overwinnen voor deze kan verlopen. Deze kan worden overwonnen door hitte toe te voegen, maar de energiebarrière kan ook worden verlaagd door bepaalde stoffen toe te voegen die de reactie makkelijker laten verlopen. Deze stoffen heten katalysators. Nog een voordeel van katalysators is dat ze niet opgebruikt worden tijdens een reactie, oftewel dat één katalysator-molecuul dezelfde reactie meerdere malen kan katalyseren. De werking van katalysators is afgebeeld in Figuur 1, waarbij de man die de berg beklimt de reactie zonder katalysator voorstelt en de vrouw in de auto de reactie met katalysator weergeeft. Uiteindelijk zullen zij allebei bij de zwart-witgeblokte vlag uitkomen, maar de vrouw in de auto heeft een stuk minder moeite hoeven doen dan de man.
Iedere katalysator die aan een bepaalde reactie wordt toegevoegd, werkt op een andere manier. Sommige werken helemaal niet, terwijl andere er voor kunnen zorgen dat een reactie nog maar een seconde duurt in plaats van uren. Een voorbeeld van een gekatalyseerde reactie is de verbranding van suiker bij lichaamstemperatuur. Normaal gesproken is de energiebarrière voor deze reactie te hoog om spontaan te verlopen, maar gelukkig zijn er katalysators in ons lichaam, genaamd enzymen, die ervoor zorgen dat deze reactie wel kan verlopen. Enzymen zijn vaak een inspiratie voor het bedenken van nieuwe katalysators, omdat zij tot de meest efficiënte groep katalysators behoren.
Tijdens dit onderzoek is de werking (kinetiek) onderzocht van een specifieke katalysator, genaamd CoMeTAA, die een reactie katalyseert waarbij cyclopropaan-derivaten (moleculen met een drie-atomige koolstofring) worden gevormd als reactieproduct. Uit de kinetiek kan namelijk de weg die een katalysator aflegt worden afgeleid, wat ook wel het mechanisme wordt genoemd. Uit een eerdere studie, op basis van computationele berekeningen, is een mechanisme voorgesteld. Of dit mechanisme echt klopt, is onderzocht door een aantal verschillende experimenten uit te voeren en de data die hieruit kwam te vergelijken met dit eerdere onderzoek. Hieruit bleek dat het voorgestelde mechanisme overeenkomt met het mechanisme afgeleid van de experimenten.
Figuur 1: een schematische weergave van katalyse.
Table of Contents
Abstract ... 7
Introduction ... 9
Results and discussion ... 16
Reaction Progress Kinetic Analysis ... 17
Determining the reaction order in catalyst concentration... 17
Catalyst deactivation and product inhibition ... 20
Determining the reaction orders in substrate concentrations ... 23
Experimental versus theoretical rate laws ... 25
Method limitations ... 26
Preliminary kinetic studies using CoTPP ... 26
Conclusions ... 28 Outlook... 28 Acknowledgement ... 30 Experimental ... 32 Experimental conditions ... 32 Reaction procedure ... 32 Kinetic experiments... 32 Workup ... 33 Data processing ... 34 Product Analysis ... 34 Selection of data ... 35 References ... 40
Abstract
This study tries to determine if the proposed mechanism on the cyclopropanation of ethyldiazoacetate (EDA) and methyl acrylate catalyzed by CoMeTAA is accurate, by comparing kinetic parameters derived from earlier DFT calculations from our group to experimentally determined values. Reaction progress kinetic analysis (RPKA) is used for analysis of the experimentally obtained data.1 This data is obtained by using a device that monitors the pressure increase in time. From the experimental data is concluded that the reaction is first order in catalyst and EDA and zero order in methyl acrylate, which agrees with the DFT-calculated orders. The values of the Gibbs free activation energy of the rate determining step (ΔG‡) calculated by DFT is compared with the ΔG‡ derived from the experimental data using the Eyring-equation. Comparison of the data suggested that the proposed mechanism is correct. The analysis of kinetic data also led to the discovery that the catalyst can be both deactivated and inhibited by the product throughout the reaction.
Introduction
Cyclopropanes were discovered by August Freund in 1881 through a reaction between dibromopropane and sodium, leading to an intramolecular Wurtz reaction, directly yielding the cyclopropane.2 They have a remarkable structure as a result of the three-membered ring, consisting of sp3-hybridized carbon atoms. The bond angle between the different carbon bonds is 104, which is smaller than the thermodynamically favored bond angle of 109.5, but higher than the geometrical 60. This leads to a ring strain (also known as Baeyer strain3) that coincides with a higher reactivity of the molecule when compared to its linear analogue, which is caused by the weaker carbon bonds (banana bonds). This makes the cyclopropane ring interesting for further reactions, especially when it is substituted with functional groups.4 A reaction of substantial synthetic interest is the ring-opening of the cyclopropyl group by addition of an electrophile, leading to open-chain compounds with up to three asymmetric centers.5 Furthermore, the cyclopropyl group itself appears abundantly in both natural and synthetic organic compounds. Some examples are shown in Figure 1. 6,7
Figure 1: a) curacin A, a natural substance found in cyanobacteria with possible antitumor applications 6 ; b) 1-sulfonylcyclopropane is used for growth regulation of wheat plants 7 ; c)
trans-chrysanthemic acid is used as an insecticide. 7
Over the years, several methods were developed for the synthesis of cyclopropanes. One possible route the use of free carbenes in combination with alkenes as substrates. 8 A disadvantage ofthis method is that free carbenes are unstable in most cases, with dimerization as the main problem.9 Another possible synthetic route is by using ylides with alkenes as substrates,10 also known as the Johnson-Corey-Chaykovsky reaction.11 A third, widely applied synthesis method is by the use of carbenoid reagents, including diazo compounds. However, diazo compounds react rather slowly and with poor selectivity with alkenes without a catalyst. The use of transition metal complexes as catalysts in combination with diazo compounds and olefins has been studied extensively.
Since nearly all natural products are chiral and numerous natural products contain cyclopropanes, it is well recognized that stereoselectivity is important in cyclopropane synthesis. The first publication on asymmetric catalyzed cyclopropanation was written by Nozaki et al. and dates back to 1966.12 The catalyst they introduced is a copper metal complex with a chiral chelate ligand (see Scheme 1).This discovery was followed by outstanding results in asymmetric catalytic cyclopropanation in the consecutive years, including catalysts that are copper13, rhodium14 and ruthenium15 based. These catalysts worked exceptionally well for the synthesis of chiral cyclopropanes derived from diazoacetates and electron-rich olefins, obtaining high yields and selectivities. However, asymmetric cyclopropanation of electron-deficient olefins remained a challenge, probably caused by the electrophilic nature of the metal-carbene intermediate, inducing dimerization of the metal-carbenes.16,17
Scheme 1: The asymmetric cyclopropanation of 2-diazo acetic acid and styrene using (bis(N-(R)-α-phenylsalicylaldiminato) copper (II)) as catalyst of Nozaki et al. The cis/trans ratio of the obtained
product is 1:2.3. An optical yield of 6% for both the 1R, 2R- and 1S, 2S-acids is reported.12
In 1978, Nakamura et al. published an article on cyclopropanation catalyzed by a cobalt (II)-dioximato complex (see Scheme 2).18 They introduced catalyzed cyclopropanation and presented unique results in the field of cyclopropanation, using electron-deficient olefins that included esters, ethers and carboxyl groups. However, further investigation of this catalyst was discouraged, because of difficulties with catalyst homogeneity when using chiral dioximato ligands.19
Scheme 2: The asymmetric cyclopropanation of Nakamura et al., using 2 mol% of bis[(-)-camphorquinone-dioximato]cobalt (II). This resulted in an enantiomeric excess (ee) of 33% for the 1S, 2S-trans-product. The reported chemical yield of the product is 11%, calculated from the amount
of ethyldiazoacetate.
The discovery of Nakamura was followed by the finding of chiral cobalt(II)-salen complexes for the cyclopropanation of styrene derivatives and ethyldiazoacetate, investigated by Katsuki et al.20 and Yamada et al.21 However, neither of them was able to obtain results that
included high yield, diastereoselectivity and enantioselectivity using electron-deficient olefins (see Scheme 3).
Scheme 3: The asymmetric cyclopropanation of Katsuki et al. of styrene and tert-butyldiazoacetate using a Co(II)-salen complex as catalyst. The chemical yield is 39% and the cis/trans ratio is 17:83.
In 2003, Zhang et al. published unique results, using cobalt (II) with chiral porphyrins as ligands, to form catalysts that produced high yield and selectivity.22 By 2009 Zhang et al. managed to obtain extraordinary results: a chemical yield up to 99%, a diastereomeric excess (de) up to 99% and an enantiomeric excess (ee) up to 95% of the cyclopropane.23 Besides having a high stereoselectivity and yield, these catalysts were also able to catalyze reactions involving electron deficient olefins (see Scheme 4).
Scheme 4: The asymmetric cyclopropanation of Zhang et al. of ethyl acrylate and ethyl α-nitrodiazoacetate using a chiral cobalt (II) porphyrin derivative (Co(3,5-DitBu-ChenPhyrin). All chiral
centers of the catalyst have (S)-chirality. A chemical yield of 62% is obtained, a cis/trans ratio of 56:44 is reported and an ee of the cis-isomer of 88% is reported.23
This unique reactivity was explained by de Bruin et al. in 2010 through a detailed study of the reaction mechanism using DFT calculations and experimental studies.24,25 During their investigations, they concluded that the mechanism of CoII(por) systems is a radical mechanism, which explains the higher reactivity towards electron-deficient olefins, because the radical carbene intermediate is more nucleophilic than the Fisher-type carbene of previously reported catalysts.
Figure 2: CoMeTAA 26
Recently, de Bruin et al. introduced a new catalyst, called CoMeTAA (see Figure 2).26
Compared to the reactivity of Zhang’s cobalt (II) porphyrins, CoMeTAA has a higher activity as reactions finish within an hour compared to an average of 24 hours.23,26 Additionally, CoMeTAA is less expensive to produce compared to cobalt (II) porphyrins. CoMeTAA also works through a radical mechanism, meaning that it is able to use electron-deficient olefins as a substrate, because there is no electrophilic Fisher-type carbene. The cyclopropanation of ethyldiazoacetate (EDA) and methyl acrylate is used as an example (Scheme 5) for the mechanism depicted in Scheme 6, which is based on DFT calculations.24,26
Scheme 5: The CoMeTAA catalyzed cyclopropanation of EDA and methyl acrylate
The reaction proceeds via a stepwise radical addition-substitution pathway, in which the redox noninnocent behavior of the terminal carbene ligand plays a key role.24 The first step
(I) describes the interaction between CoMeTAA (A) and EDA to form the temporary intermediate B, which loses dinitrogen during the rate-determining step II, with the corresponding transition state TS1. This leads to the carbon-centered radical terminal carbene
C that is in equilibrium with the bridging radical carbene C’. The terminal carbene is best
described as a one-electron reduced Fisher-type carbene. The bridging radical carbene C’ is a dormant state of the catalyst and is incapable of forming the cyclopropane.24 The third step in the cycle (III) is an irreversible radical addition of the carbon-centered radical carbene to methyl acrylate, with transition state TS2, where the radical shifts to the γ-carbon of the alkyl moiety. The γ-alkyl radical type then cyclizes to form the corresponding product during step
IV, with the transition state TS3, and the catalyst returns to its original state. This is a concerted
radical type C-C bond formation with simultaneous homolysis of the Co-C bond.24 The barrier of this ring-closure reaction is so low that apart from cyclopropanation no other reactions take place, such as free radical polymerization.24,26 Based on DFT calculations, the reaction is first order in EDA and catalyst and zero order in methyl acrylate.26
To provide experimental evidence for the mechanism of CoMeTAA, the kinetics of the reaction can be studied. This is followed by determination of the rate law that is deduced from experimental reaction rates and comparison of the experimental rate law to the theoretical rate law of the proposed mechanism. This was the goal of the studies described in this report. Herein we describe a detailed kinetic study of the cyclopropanation reaction mediated by CoMeTAA to validate the reaction mechanism shown in Scheme 6 based on DFT. The research question of this project is therefore: does the proposed theoretical mechanism agree with the mechanism derived from kinetic analysis of the experiments?
Scheme 6: Proposed reaction mechanism of CoMeTAA-catalyzed cyclopropanation of EDA and methyl acrylate. The ∆G values of the different intermediates and transition states can be found
between brackets.24,26
There are a few methods that can translate the experimental data into kinetic parameters, which will be discussed below. The essence of processing kinetic data is finding the rate law of a reaction, which can in turn be compared to the theoretical rate laws, thus determining if the mechanism agrees with the experimental data. The rate law of this reaction has the form of Equation 1, where exponents x, y and z are the reaction orders of respectively EDA, methyl acrylate and CoMeTAA, and k is the rate constant. These reaction orders and this constant can be derived from experimental data while the concentrations of substrate A and B are measured in kinetic experiments.
𝑟 = 𝑘 [𝐸𝐷𝐴]𝑥[𝑚𝑒𝑡ℎ𝑦𝑙 𝑎𝑐𝑟𝑦𝑙𝑎𝑡𝑒]𝑦[𝐶𝑜𝑀𝑒𝑇𝐴𝐴]𝑧
Equation 1: The general rate law for the cyclopropanation of EDA and methyl acrylate, catalyzed by CoMeTAA
In order to determine the unknown parameters, kinetic measurements have to be performed, which can be differential or integral. A differential measurement is a method where the reaction rate is directly measured against time. An example of an in-situ differential method is a calorimeter.1 Integral methods are defined as methods where the concentration of the substrate is directly related to a measurable parameter.1 A few examples of in-situ integral methods are spectroscopic methods, including FTIR, UV and Raman spectroscopy, and gas uptake or release measurements.27 The kinetics of the reaction between ethyldiazoacetate and methyl acrylate catalyzed by CoMeTAA can be obtained by monitoring the release of nitrogen during the reaction and measuring its pressure build-up in a sealed vessel.
Once the pressure data is obtained, a method for processing the data has to be chosen, to determine the rate law. One possibility is to use the isolation method.28 The dependency of the reaction on the catalyst concentration is left out of the following example. When using this method, the concentration of for instance EDA is kept constant by setting it to a value that is much higher than the concentration of methyl acrylate. The concentration of EDA will then have barely changed once the reaction is finished, so the change in rate observed is caused by the different concentration of methyl acrylate. A pseudo zero order in EDA concentration is obtained by this use of a large excess. The rate constant that is observed can then be divided by the concentration of EDA to obtain the rate constant (see Equations 2 and 3).
𝑟 = 𝑘𝑜𝑏𝑠[𝑚𝑒𝑡ℎ𝑦𝑙 𝑎𝑐𝑟𝑦𝑙𝑎𝑡𝑒]
Equation 2: The pseudo first order rate law when using a large excess of EDA28
𝑘𝑜𝑏𝑠 = 𝑘 [𝐸𝐷𝐴]
Equation 3: The relation between the observed rate constant and the actual rate constant 28
Another way to obtain the kinetic data is by using the initial rate method, where the concentrations of different substrates are individually varied while monitoring the initial rates of the reaction.28 It can subsequently be determined how the reaction rate is dependent on the substrate concentrations and what the different constants in the rate equation are.29 However,
using this method requires the monitoring of numerous experiments.
A more efficient way, which requires only a few experiments, is Blackmond’s reaction progress kinetic analysis (RPKA).1 The advantage of using RPKA over of the initial rate
method is that the reaction progress kinetic analysis follows the complete reaction, where the initial rate method only uses information of the beginning of the reaction.1 The analysis (see Scheme 7) works by carrying out essential reactions (the green boxes), followed by processing the data (the black boxes) by plotting certain graphs. Subsequently, it is checked if the graphs of certain experiments overlay (the pink boxes), from which various conclusions can be drawn (the blue boxes). This convenient kinetic analysis will be used for the CoMeTAA-catalyzed cyclopropanation of methyl acrylate and ethyldiazoacetate.
Results and discussion
The reaction that is looked at in detail during this research is the CoMeTAA-catalyzed cyclopropanation of ethyldiazoacetate (EDA) and methyl acrylate (Scheme 5). The progress of this reaction can be monitored in different ways, resulting in data that is useful for determining the kinetics of the reaction. For this project, the progress of the reaction is determined by monitoring the amount of evolved nitrogen. A device that monitors the pressure and temperature against time is used for this (see Figure 3).
The reaction takes place in an oxygen- and water-free 10 mL Schlenk tube (with a total volume of 20.1 mL) that is attached to the measuring device (the sensor) and placed in a thermostatic ethanol bath at 283 K. The Schlenk tube has two openings, one where it can be attached to the sensor, and one where a screw cap can be screwed on, to make the Schlenk tube leak free. A screwcap in combination with a Teflon septum is used to add the reagents in the Schlenk while under nitrogen atmosphere.
Figure 3: Reaction setup
Initial studies are carried out to determine the limits of cyclopropanation, such as order of addition, variations in concentration, or determining the lower limit of catalyst loading. If EDA is added to the reaction mixture after methyl acrylate, the reaction proceeds normally. When this order is reversed, the reaction does not take place, as there is no nitrogen pressure increase. However, the byproduct of the uncatalyzed reaction between EDA and methyl acrylate is observed in the 1H-NMR spectrum (see Scheme 8).26,30 This can be explained by catalyst deactivation caused by the reactive carbene intermediate formed from EDA, which will most likely abstract a proton from the reaction medium in the absence of methyl acrylate.24 If only ethyldiazoacetate is added to the reaction mixture to see if the dimer diethyl fumarate
CoMeTAA. Above experiments were carried out in dichloromethane (DCM), which has a relative low boiling point when compared to other solvents such as toluene. The risk of using DCM as solvent is that it can cause fluctuations in pressure or volume, due to evaporation. This is why was chosen to continue the kinetic studies in toluene, which has a higher boiling point.
Scheme 8: Uncatalyzed reaction between EDA and methyl acrylate 26,30
Reaction Progress Kinetic Analysis
The data analysis method of choice for this project is reaction progress kinetic analysis (RPKA) developed by Blackmond, because this method translates information obtained throughout the entire reaction into useful kinetic data.1 RPKA describes the possibility of
having a full kinetic understanding of the reaction, by processing only a few essential reactions, which can be found in Scheme 7. All reactions are performed in duplo or triplo, to assure that the data is reproducible. The experimental details and detailed processing steps of the reactions can be found in the experimental section.
RPKA starts with a reaction at standard conditions, for which concentrations are used that are as close as possible to the ones normally used in the laboratory. It is necessary to set one reaction with its conditions as the reference, because data collected from other experiments will always be compared to this (for the detailed standard reaction conditions, see experimental section). The path of Blackmond’s flowchart is followed to carry out further reactions.
Determining the reaction order in catalyst concentration
The following two experiments have a higher and lower catalyst loading, but equal substrate concentrations compared to the standard conditions. By conducting these experiments, it is possible to determine the order of the reaction in catalyst concentration. If the reaction is first order in catalyst, which is usually the case, the rate of the reaction will be twice as low when the catalyst concentration is twice as low.31 If the reaction is second order in the catalyst concentration, the rate of the reaction is four times as low when the catalyst is twice as low. This can also be seen in the general rate equation (see Equation 1).
Graph 1: Conversion versus time for experiments at different catalyst loadings
During this experiment, it is assumed that the active catalyst concentration is constant, because the amount of catalyst bound to the substrate is negligible, since the reaction happens fast.
The difference in reaction rate can only be caused by the difference in catalyst concentration during these experiments, because the concentration of the substrates remain the same. It can be seen from Graph 1 that different catalyst concentrations influence the reaction rates in different ways, because the reaction reaches a conversion of 70% sooner when using a higher catalyst concentration. A higher catalyst concentration thus accelerates the reaction speed and a lower catalyst concentration decreases the rate. This effect is also visible in Graph 2, where the reaction rate is plotted against EDA concentration. However, it is not possible to see from these graphs whether the reaction is first order or higher in catalyst concentration.
Graph 2: Rate versus [EDA] for experiments at different catalyst loadings
This question can be answered by plotting the turn-over frequency (TOF) (see Equation 4) against the EDA concentration. If the graphs of the different experiments overlap, the reaction is first order in the catalyst concentration. An explanation for this can be found in the
doubled, leading to the same TOF for different experiments at the same EDA concentrations. If the reaction has a different order than one, the TOF will have a different value for different catalyst concentrations.
𝑇𝑂𝐹 = 𝑟𝑎𝑡𝑒 [𝑐𝑎𝑡𝑎𝑙𝑦𝑠𝑡]
Equation 4: Turn-over frequency
As can be seen in Graph 3, the plot of the TOF versus [EDA] at higher catalyst concentration overlaps with the plot of the standard conditions. However, the plot of TOF versus [EDA] at lower catalyst concentration lies slightly lower. This can be caused by the fact that the reaction is not first order in catalyst but this lowering of the plot can also be caused by catalyst deactivation, considering that the higher catalyst concentration curve overlaps with the standard conditions. It is therefore recommended that additional experiments are conducted for the determination of the reaction order in catalyst concentration, but at a higher catalyst concentration. Furthermore, another experiment should be attempted in which the catalyst concentration is decreased below ‘lower catalyst concentration’ to prove catalyst deactivation. When the initial rates of the different experiments are plotted against the catalyst concentration, it can also be seen that the reaction is first order in catalyst, since this plot is a linear curve (Graph 4).
Graph 4: Initial rates (measured at 12% conversion) plotted against the catalyst concentration Catalyst deactivation and product inhibition
To further investigate what happens to the catalyst, reactions at different substrate concentrations are carried out, while keeping the difference between the concentration of methyl acrylate and EDA constant. These reactions are called ‘same excess’ experiments. The ‘excess’ is defined as the difference between the initial concentration of methyl acrylate and the initial concentration of EDA (see Equation 5). Because one molecule of EDA reacts with one molecule of methyl acrylate, this ‘excess’ will be constant during the entire reaction progress.
[𝑒𝑥𝑐𝑒𝑠𝑠] = [𝑚𝑒𝑡ℎ𝑦𝑙 𝑎𝑐𝑟𝑦𝑙𝑎𝑡𝑒] − [𝐸𝐷𝐴]
Equation 5: Definition of excess concentration
The kinetic data extracted from these same excess experiments can show whether the catalyst is inhibited by the product or deactivated. Three experiments at ‘same excess’ are conducted, including the reaction at standard conditions. The other two reactions are designed to have lower initial substrate concentrations that are equal to the concentrations of the substrates at 25% conversion and 50% conversion of the standard reaction (see experimental section). These reactions thus simulate the standard reaction at 25% and 50% conversion, but have a fresh batch of catalyst. This is to see if the catalyst batch of the standard conditions still performs as a fresh catalyst batch and did not deactivate or got inhibited by product after 25% or 50% conversion. The rates of the different experiments can be compared and if there would be no catalyst deactivation or product inhibition, the rate at an initially lower substrate concentration should be the same as the measured rate at that same lower substrate concentration of the reaction at standard conditions. In short, if there is no catalyst deactivation or product inhibition, the plots of the rate versus [EDA] of the three experiments should overlap.
Graph 5: Conversion versus time of the experiments at the same [excess]
The conversion versus the reaction time is plotted in Graph 5. It is noteworthy that the plots overlap, because the reaction rate of the ‘same excess’ experiments is supposed to be lower than the rate of the reaction at standard conditions if there is no catalyst deactivation or inhibition. The reaction rate is namely partly dependent on the substrate concentrations, so lower initial substrate concentrations would result in lower reaction rates. The fact that the reactions with a new catalyst batch at lower initial substrate concentrations are faster than expected, is a first hint towards catalyst deactivation or inhibition.
Graph 6: [EDA] versus time for experiments at the same [excess]
The [EDA] is plotted against time in Graph 6, in which can be seen that the same excess reactions start at different EDA concentrations. When the plots of the same excess experiments are superimposed on the plot of the standard conditions, they should overlap, which is visualized in Graph 7. This graph shows that the EDA concentration of the same excess experiments decreases faster compared to the standard reaction.
Graph 7: [EDA] versus shifted time for experiments at the same [excess]
Catalyst deactivation or product inhibition is easiest to visualize by plotting the reaction rate versus [EDA] (Graph 8). The plots do not overlap, because at identical EDA concentrations the reaction rate of the standard reaction lies lower than both ‘same excess’ experiments. These ‘same excess’ experiments suggest that a fresh batch of catalyst always performs better and then slowly gets deactivated or inhibited by the product. Furthermore, it can be derived from Graph 7 and 8 that after three minutes almost 50% of the catalyst has deactivated.
Graph 8: Rate versus [EDA] for experiments at same [excess]
In order to find out if the difference in rates is caused by catalyst deactivation or product inhibition, an experiment can be conducted where a known amount of product is added to one of the two reactions at ‘same excess’. In this case, the formed cyclopropane is added to the reaction that starts at 50% conversion compared to the standard reaction. The product concentration corresponds to the concentration in the reaction mixture of the standard reaction at a conversion of 50%.
As can be seen in Graph 9, the experiment where the product is added is compared to the experiment where no product is added and to the experiment at standard conditions. The curve of the experiment with product addition overlaps with neither the standard curve nor the
of the standard conditions, but lower of that of the experiment without product addition. If the decrease in rate compared to the experiment without product addition is purely caused by product inhibition, the graph would have overlapped with the graph of the standard conditions. If it had overlapped with the graph from the same experiment but without product addition, there would be no product inhibition and the decrease in rate would be caused by catalyst deactivation. Looking at the graph, the rate reduction during the reaction is probably a result of a combination of both. However, the curve approaches the curve without product addition more than the standard curve, hence is assumed that the rate reduction will have more influence from catalyst deactivation than from product inhibition.
Graph 9: Rate versus [EDA] for experiments at same [excess] with and without product addition
Determining the reaction orders in substrate concentrations
Finally, the reaction orders in the substrate concentrations will be determined for comparison with the DFT-calculated values. This can be accomplished by conducting experiments at ‘different excess’ concentrations of methyl acrylate compared to the standard reaction. If the reaction is zero order in methyl acrylate, the concentration of methyl acrylate will have no effect on the reaction rate. As can be seen in Graph 10, both reactions reach a conversion of 70% in approximately the same time.
Graph 10: Conversion versus time for experiments at different [excess]
Subsequently, the reaction rate is plotted versus the EDA concentration (Graph 11), which clearly shows that the two curves overlay. The reaction is therefore zero order in methyl acrylate. If the reaction is first order in EDA concentration, the plot of rate versus [EDA] would result in a linear plot. When looking at the graph, it can be seen that the plot is a straight line, until the EDA concentration becomes lower than 0.125 M. The curvature of the graph at low EDA concentrations could be caused by the fact that a polynomial fit instead of an exponential fit is used to fit the experimentally obtained data (see experimental section). The software used during this project was not advanced enough to fit the data to an exponential function. Another reason for this curvature could be the result of catalyst deactivation at high conversions. Since the major part of the plot is linear, it is concluded that the reaction is first order in [EDA]. However, to verify this, it is recommended to fit the data to an exponential curve and see if this will give a straight line in EDA concentration.
Experimental versus theoretical rate laws
The experimentally determined rate law of the cyclopropanation of methyl acrylate and EDA using CoMeTAA as a catalyst is equal to:
𝑟𝑎𝑡𝑒 = 𝑘 [𝐸𝐷𝐴][𝐶𝑜𝑀𝑒𝑇𝐴𝐴]
Equation 6: The rate law for the cyclopropanation of EDA and methyl acrylate using CoMeTAA as catalyst
since the reaction is first order in EDA concentration and first order in catalyst concentration. The reaction is zero order in methyl acrylate, hence the exclusion of methyl acrylate from the rate law.
DFT calculations were performed for a complete mechanistic investigation of this systems. From these calculations is determined that the rate determining step is the reaction of EDA with the catalyst with release from N2 from the metal-bound EDA.26 The corresponding
Gibbs energy of activation (ΔG‡
) was calculated at 10 kcal/mol. The reaction of methyl acrylate with the carbene intermediate has a lower transition state barrier of only 5 kcal/mol, thus it can be concluded that the cyclopropanation is first order in EDA and CoMeTAA concentrations and zero order in methyl acrylate.26
The Gibbs energy of activation (ΔG‡
) of the rate determining step can be compared to the experimental value for the ΔG‡, which can be derived from the Eyring equation (see Equation 7).
𝑘 = 𝑘𝐵𝑇 ℎ 𝑒
−𝛥𝐺𝑅𝑇‡
Equation 7: The Eyring equation
where:
k = reaction rate constant kB = Boltzmann constant
T = absolute temperature h = Planck’s constant
ΔG‡ = Gibbs energy of activation
R = gas constant
The reaction rate constant can be determined from experimental values by using the isolation method. For the isolation method it was necessary to keep the concentration of one of the components constant. Using the steady state approximation, it is most convenient to keep the catalyst concentration constant. The observed rate is then equal to:
𝑟𝑎𝑡𝑒 = 𝑘𝑜𝑏𝑠[𝐸𝐷𝐴]
Equation 8: The observed rate equation
with
𝑘𝑜𝑏𝑠 = 𝑘[𝐶𝑜𝑀𝑒𝑇𝐴𝐴]
The reaction rate constant can be easily derived from a rate versus EDA concentration plot, as it is equal to the slope divided by the catalyst concentration. It was chosen to derive k from the higher catalyst concentration experiment, as catalyst deactivation has the smallest influence on the rate in this reaction. This results in a Gibbs energy of activation of 16 kcal/mol, which is not identical to the calculated energy.
There are multiple explanations that could clarify this difference. Firstly, the experimental k was determined from only one experiment, thus being prone to experimental errors that would result in a deviating ΔG‡. Secondly, k is only determined at one temperature (283 K). To get a more accurate version of ΔG‡, a temperature variation graph should be created
with more experiments. Finally, the effect of solvent molecules is not taken into account in the DFT calculations and it might be possible that gas phase DFT calculations differ slightly in energy barriers compared to the energy in solution determined experimentally.
Apart from the difference in the ΔG‡ of 6 kcal/mol, the data seems to coincide sufficiently to consider the experiments as evidence of the proposed and calculated mechanisms. The reaction orders of the components are similar and both methods suggest the same rate determining step.
Method limitations
While conducting experiments in duplo or triplo, it was noticed that repetition of the same reactions seldom led to the exact same results. This can be explained by the fact that kinetic measurements are highly sensitive to experimental errors.
As a first example, the reaction setup sometimes leaked nitrogen slowly, causing the pressure to decrease and to differ when comparing identical reactions. Secondly, a glass syringe was used to add the substrates, which does have a certain measuring error. This means that maybe inaccurate amounts of substrates were added in the different duplo experiments, possibly resulting in different conversions, even though utmost care has been taken to try to reduce these measurement errors. Thirdly, the amount of nitrogen that is dissolved in the reaction mixture is considered negligible. Fourthly, an identical stirring speed for all reactions is important for rate measurements. In general the stirring speed was kept constant at 500 rpm, but an analog stirring plate was used, which is relatively inaccurate. Lastly, the reaction was performed in a glass Schlenk flask and glass is not an excellent heat conductor, resulting in a less than ideal heat transfer from the reaction medium to the surrounding thermostatic ethanol bath. This might have caused the temperature inside the reaction Schlenk flask to be higher than 283 K, due to the exothermicity of the reaction.
Preliminary kinetic studies using CoTPP
For mechanism comparison, initial kinetic studies were performed on CoII
tetraphenylporphirin (CoTPP) (see Figure 5) using the EDA and methyl acrylate. The solubility of CoTPP in toluene is low, which is why was chosen to dissolve the complex in DCM to produce a 0.0015 M stock solution. However, some problems were discovered during the first experiments.
It was soon found out that the low boiling point of DCM was problematic, causing part of the solvent to evaporate, resulting in an unwanted increase in pressure and decrease in volume. Another problem was that the solubility of CoTPP in DCM is quite low as well, leading to a relative dilute solution of the catalyst and substrates and a slow reaction. Secondly, the reaction could not be performed using the thermostatic ethanol bath. The thermostatic
ethanol bath has a maximum temperature of 10ºC and at this temperature the reaction proceeds sluggish and reaches only a maximum conversion of 17%. The low reaction rate led to a low selectivity, because byproducts can be seen in 1H-NMR spectra, including the product from the
uncatalyzed reaction of EDA and methyl acrylate (see Scheme 8).
If kinetic studies are to be performed the catalyzed cyclopropanation using CoTPP, it is necessary to change the conditions in which the reaction is performed. First of all, the temperature at which the reaction takes place should be elevated, to make the reaction proceed faster. Secondly, a different solvent needs to be used, such as chlorobenzene. The complex should have a higher solubility in this solvent and the boiling point should be much higher than the temperature at which the reaction is carried out.
Conclusions
This study set out to determine if the mechanism of the CoMeTAA-catalyzed cyclopropanation of EDA and methyl acrylate derived from experimental kinetic data agrees with the DFT-calculated mechanism. Initial studies showed that the reaction does not proceed if EDA is introduced to the reaction before methyl acrylate and that no reaction takes place without methyl acrylate. Kinetic data of further reactions is translated into kinetic parameters with the help of Blackmond’s reaction progress kinetic analysis.
The first conclusion from the conducted experiments is that the reaction is first order in catalyst. The second finding is the first-order dependency in EDA and the zero-order dependency in methyl acrylate. Both findings are in line with the results from DFT calculations. The third conclusion is that the Gibbs activation energy barriers are not identical. Before any further conclusions can be drawn from the difference between the experimental and theoretical values, the Gibbs energy of activation needs to be calculated from experimental and calculated data at a variation of temperatures. A fourth conclusion that can be deduced from the analysis is that the catalyst is deactivated or inhibited by the product to a certain extent during the reaction. Taken together, these results suggest that the currently presented mechanism for the CoMeTAA-catalyzed cyclopropanation is correct, since the mechanisms deduced from separately performed kinetic studies are in agreement.
Outlook
For full comparison of the two differently determined mechanisms, the Gibbs free activation energy needs to be determined at different temperatures, since the Eyring equation (see Equation 7) is temperature-dependent. The kinetic studies can also be expanded using different measurement techniques. This multi-technique procedure might present supplementary mechanistic information that one technique alone might miss.1 It might be interesting to study a broader substrate scope, including electron-rich olefins, to see if the mechanism is generally applicable for this catalyst.
If any further studies on the mechanism of CoTPP are to be conducted, it is recommended to change the conditions in which the reaction takes place. The temperature at which the data is collected should be elevated to a temperature where the reaction proceeds faster. This higher reaction temperature should be taken into account when choosing the solvent, to ensure that solvent evaporation plays no part in the pressure increase during the reaction. It is advised to use a solvent that dissolves CoTPP well, such as chlorobenzene.
Acknowledgements
Firstly, I would like to thank prof. dr. Bas de Bruin for giving me the opportunity to do my bachelor internship at Homkat, and for his advice and input. Secondly, I would like to thank my supervisor, drs. Andrei Chirila, for sharing his knowledge and for his patience and guidance. Furthermore, I would like to thank prof. dr. Jan van Maarseveen for being my second corrector and prof. dr. Joost Reek and dr. Jarl-Ivar van der Vlugt for the useful discussions during mini-meetings. I want to thank the people on lab E1.05 for helping me out if I had questions about the lab and the rest of the Homkat group for the discussions and for making it possible to have such a great time here.
Experimental
The chemicals used during this research were purchased from Sigma Aldrich. EDA and methyl acrylate are prior degassed using the freeze-thaw-pump-method. Methyl acrylate is passed through basic alumina before use, to remove radical scavengers that avoid polymerization. Both substrates are kept in the fridge (6C) in J. Young-valve Schlenk flasks. CoMeTAA is not commercially available and is synthesized according to a known procedure.26
All reactions are performed under nitrogen atmosphere.
The ethanol bath used is Neslab ULT-80. The kinetic kit used for N2 pressure increase
measurements is a series X102 kit from Man on the Moon.
Experimental conditions Reaction procedure
The 10 mL reaction Schlenk flask is set in place and a 1.0 cm cylindrical stirring bar is added, while it is made sure that the system is leak free. The system is flushed with nitrogen, including the small volume between the sensor and the crane. While under vacuum, the flask is heated using a heat gun to vaporize and evacuate traces of water. The addition order of the reactants in the reaction flask is as following: firstly the catalyst solution, secondly the solvent, thirdly the alkene and finally the diazo compound. After addition of the first three reactants, the septum of the reaction flask is replaced. The diazo compound is not yet added to the flask, but collected in a syringe and the needle is introduced through the septum into the system without coming into direct contact with the other substrates. This whole reaction setup is introduced in the thermostatic ethanol bath. The mixture is stirred at 500 rpm and the setup is cooled for 30 minutes. During this time a test measurement is started to check the connection of the sensor to the computer and to monitor the temperature of the reaction Schlenk. After reaching the desired temperature the crane is turned towards the pressure sensor. The test measurement is stopped and a new measurement is started, which will monitor the pressure and temperature of the reaction.The entire amount of the diazo compound is added at once, after which the reaction starts. Subsequently the needle is removed quickly from the septum to avoid nitrogen leaking through the syringe. Once the nitrogen evolution has reached a plateau, the measurement is stopped.
Kinetic experiments
All reactions took place in toluene. The reaction mixture always had a total volume of 2 mL. The different conditions can be found in the table below. A 0.01 M stock solution of CoMeTAA in toluene was used for all experiments. A 0.474 M product stock solution in toluene is used.
Table 1: Reaction conditions for RPKA
Reaction CoMeTAA EDA Methyl
acrylate Toluene Product Standard conditions 750 µL 0.0075 mmol 0.015 eq 53 µL 0.5 mmol 1 eq 90 µL 1 mmol 2 eq 1.1 mL Higher catalyst concentration 1000 µL 0.01 mmol 0.02 eq 53 µL 0.5 mmol 1 eq 90 µL 1 mmol 2 eq 857 µL Lower catalyst concentration 500 µL 0.005 mmol 0.01 eq 53 µL 0.5 mmol 1 eq 90 µL 1 mmol 2 eq 1357 µL Same excess at 25% conversion 750 µL 0.0075 mmol 0.015 eq 40 µL 0.375 mmol 1 eq 79 µL 0.875 mmol 2.33 eq 1132 µL Same excess at 50% conversion 750 µL 0.0075 mmol 0.015 eq 29 µL 0.25 mmol 1 eq 68 µL 0.75 mmol 3 eq 1156 µL Same excess at 50% conversion with product addition 750 µL 0.0075 mmol 0.015 eq 29 µL 0.25 mmol 1 eq 68 µL 0.75 mmol 3 eq 629 µL 530 µL Different excess 750 µL 0.0075 mmol 0.015 eq 53 µL 0.5 mmol 1 eq 135 µL 1.5 mmol 3 eq 1062 µL Workup
Two workup methods were used for the different experiments. Method 1 was used for the initial CoMeTAA experiments. Both method 1 and method 2 were used for comparison for the CoTPP experiment series. For the CoMeTAA experiments both workup methods were used as well, but method 2 resulted in higher isolated yields.
Method 1
A 15 mL disposable filter is filled with roughly 2-3 cm silica (60-200 m) that is impregnated with DCM. The reaction mixture is poured on top of the silica and the small column is eluted with DCM (50 mL). The filtrate is collected, which is followed by evaporation of DCM using a Rotavap. An NMR sample is prepared in CDCl3 using mesitylene (1 equivalent
compared to EDA) as internal standard. Method 2
The reaction mixture is transferred into a round bottom flask and the reaction Schlenk is washed with the solvent. The solvent is evaporated on the rotavap and the remaining residue is extracted thrice with pentane. Pentane is evaporated on the rotavap and the remaining product is analyzed using 1H-NMR spectroscopy and mesitylene as an internal standard.
Data processing
The data presented in Graph 12 is directly obtained from the kinetic setup used. For this reaction, the EDA was added after one minute, and an increase in pressure can be observed. At a pressure increase of approximately 0.35 bar, the system always gave the same pressure drop of 0.12 bar, due to changing the pressure measuring regime of the sensor, which is corrected in further processing of the experiment.
This data was transferred to Origin, a software for scientific graphing and data analysis. For further processing, the data between 5% to 70 % conversion is used, the rest of the data is discarded. At least two experiments were averaged, by using the “Average multiple curves” function. The data that comes out of this function is a concatenate of multiple curves, so all the points are merged. This means that there is no line yet that is the average of the different experiments. This was achieved by fitting the concatenate data to a polynomial of 500 points.
Once an average curve of the different experiments was determined, in the form of a time versus conversion curve, different quantitative parameters were calculated using the ideal gas law, including the conversion and EDA concentration. From the EDA concentration, the rate was derived using the option “Differentiate” and “Derivative order: 1”.
Product Analysis
1H-NMR spectra were taken from the products of all reactions to assure product
formation and yield, which was determined using mesitylene as internal standard. To characterize the product and determine the yield, an 1H-NMR spectrum was taken with mesitylene as internal standard (see Graph 13 below). The product shifts are as follows: 1
H-NMR (300 MHz, CDCl3): δ = 4.185 (q, J = 6 Hz, 2H), 3.72 (s, 3H), 2.21 (m, 2H), 1.47 (m,
2H), 1.30 (t, J = 6 Hz, 3H). The shifts of the internal standard mesitylene are: 1H-NMR (300 MHz, CDCl3): δ = 6.84 (s, 3H), 2.31 (s, 9H) -0,1 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0 5 10 15 20 25 N itr o ge n p re ssur e in cr e ase (b ar ) Time (min)
Initial data
Example of experimentGraph 13: The1H-NMR spectrum of the product with mesitylene as an internal standard
Gas chromatography (GC) samples were taken to determine the cis-trans ratio of the product (see Figure 6). GC samples were first taken from the reaction mixture after silica filtration, but it was soon found out that the cis-product was lost on the silica column, which is why was decided to analyze the raw reaction mixture. The average trans/cis ratio was 97:3. The GC used is a Shimadzu 17A with a Supelco SPB TM – 1 Fused Silica Capillary Column with a length of 30 m, a diameter of 0.32 mm and a film thickness of 2.0 µm. The method uses a rate of 7 ºC/min from 70 ºC to 250 ºC. The retention time of the trans-product is 14.45 and that of the cis-product is 14.78 minutes.
Figure 5: The cis- and trans-products Selection of data
The conversion versus time graphs of different experiments in duplo, triplo or quadruplo were compared to see if the obtained data was reliable. Some of the experiments were discarded because they had deviant data for different reasons that could include leakage or stirring problems. The detailed selection for all experiments can be found below.
Standard conditions
Graph 14: Standard conditions experiment selection
In this case, it was chosen to discard MBB034 and MBB030, because they respectively had deviant data or stirring problems. The average is a 4th order polynomial.
Higher catalyst loading
Graph 15: Higher catalyst loading experiment selection
Lower catalyst loading
Graph 16: Lower catalyst loading experiment selection
In this case, MBB026 was discarded. The reason for this is that the rate is visibly higher throughout the whole reaction and this reaction reached a conversion that was considerably higher than the other two experiments (>7%). The average calculated from ACH212 and MBB036 is a 6th order polynomial.
Same excess at 25% conversion
Eventhough the conversion of ACH210 is higher than that of MBB027 and MBB028, the average was calculated using all three reactions, because the rate overlaps for a part of the reaction. The average is a 4th order polynomial.
Same excess at 50% conversion
Graph 18: Same excess at 50% conversion experiment selection
ACH213 is discarded for the calculation of the average conversion. The average is a 4th order
Same excess at 50% conversion with product addition
Graph 19: Same excess at 50% conversion with product addition experiment selection
All three experiments are taken into account for the calculation of the average conversion. The average is a 3rd order polynomial.
Different excess
Graph 20: Different excess experiment selection
References
1. Blackmond, D. G. Angew. Chem. Int. Ed. 2005, 44, 4302 – 4320.
2. a) Freund, A. J. Prakt. Chem. 1881, 26, 367 – 377; b) Freund, A. Monatsh. Chem.
1882, 3, 625 – 635.
3. Wade, L.G. Organic Chemistry 6th edition; Pearson Prentice Hall: Upper Saddle River, New Jersey, 2006; 103 – 122.
4. March, J.; Smith, M.B. March’s Advanced Organic Chemistry 6th edition; John Wiley and Sons, New Jersey, 2007; 216 – 220.
5. Reissig, H. in The Chemistry of the Cyclopropyl Group; Rappoport, Z., Eds. John Wiley and Sons: New York, 1987; 1, 376 – 434.
6. Moriarty, R. M.; Tyagi, S.; Kinch, M. Tetrahedron 2010, 66, 5801 – 5810. 7. Salaün, J.; Baird, M.S. Curr. Med. Chem. 1995, 2, 511 – 542.
8. Fedoryński, M. Chem. Rev. 2003, 103, 1099 – 1132.
9. Grundmann, C. Justus Liebigs Ann. Chem. 1938, 536, 29 – 36. 10. Li, A.; Dai, L.; Aggarwal, V. K. Chem. Rev. 1997, 97, 2351 – 2356.
11. a) Johnson, A.W.; LaCount, R.B. J. Am. Chem. Soc. 1961, 83, 417 – 423; b) Corey, E.J.; Chaykovsky, M. J. Am. Chem. Soc. 1965, 857, 1353 – 1364.
12. Nozaki, H.; Moriuti, S.; Takaya, H.; Noyori, R. Tetrahedron Lett. 1966, 7, 1839 – 1844.
13. Lo, M.; Fu, G. C. J. Am. Chem. Soc. 1998, 120, 10270 – 10271.
14. Davies, H. M. L.; Bruzinski, P. R.; Lake, D. H.; Kong, N.; Fall, M. J. J. Am. Chem. Soc. 1996, 118, 6897 – 6907.
15. Che, C.; Huang, J.; Lee, F.; Li, Y.; Lai, T.; Kwong, H.; Teng, P.; Lee, W.; Lo, W.; Peng, S.; Zhou, Z. J. Am. Chem. Soc. 2001, 123, 4119 – 4129.
16. Paul, N. D.; Chirila, A.; Lu, H.; Zhang, X. P.; Bruin, de B. Chem. Eur. J. 2013, 19, 12953 – 12958.
17. Chen, Y.; Ruppel, J. V.; Zhang, X. P. J. Am. Chem. Soc. 2007, 129, 12074 – 12075. 18. Nakamura, A.; Konishi, A.; Tatsuno, Y.; Otsuka, S. J. Am. Chem. Soc. 1978, 100,
3443 – 3448.
19. Doyle, M.P. Angew. Chem. Int. Ed. 2009, 48, 850 – 852.
20. Niimi, T.; Uchida, T.; Irie, R.; Katsuki, T. Adv. Synth. Catal. 2001, 343, 79 – 88. 21. Ikeno, T.; Sato, M.l Sekino, H.; Nishizuka, A.; Yamada, T. Bull. Chem. Soc. Jpn.
2001, 74, 2139 – 2150.
22. Huang, L.; Chen, Y.; Gao, G.; Zhang, X. P. J. Org. Chem. 2003, 68, 8179 – 8184. 23. Zhu, S.; Perman, J. A.; Zhang, X. P. Angew. Chem. Int. Ed. 2008, 47, 8460 – 8463. 24. Dzik, W. I.; Xu, X.; Zhang, X.P.; Reek, J. N. H.; de Bruin, B. J. Am. Chem. Soc.
2010, 132, 10891 – 10902.
25. Lu, H.; Dzik, W. I.; Xu, X.; Wojtas, L.; de Bruin, B.; Zhang, X.P. J. Am. Chem. Soc.
2011, 133, 8518 – 8521.
26. Chirila, A.; de Bruin, B. University of Amsterdam. Unpublished work, 2015. 27. vallance.chem.ox.ac.uk, reaction kinetics lecture notes, consulted on 9th of January
2016.
28. Rothenberg, G. Catalysis – concepts and green applications 1st edition; Wiley-VCH,
Weinheim, 2008; 61 – 62.
29. Casado, J.; López-Quintela, M. A.; Lorenzo-Barral, F.M. J. Chem. Educ. 1986, 5, 450 – 452.
30. Krishna, P. R.; Empati, R. S.; Mongin, F. Tetrahedron Lett. 2008, 49, 6768 – 6772. 31. Blackmond, D. G. J. Am. Chem. Soc. 2015, 137, 10852 – 10866.
