Subscriber access provided by UNIV OF TEXAS DALLAS
Note
Investigation of a weak temperature–rate relationship in the carbamoylation of a barbituric acid pharmaceutical intermediate Alexander G. O'Brien, Yangmu C Liu, Mark J Hughes, John Jin Lim, Neil S. Hodnett, and Nicholas Falco J. Org. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.joc.9b00411 • Publication Date (Web): 06 Mar 2019 Downloaded from http://pubs.acs.org on March 10, 2019
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Organic Chemistry
Investigation of a weak temperature–rate relationship in the carbamoylation of a barbituric acid pharmaceutical intermediate Alexander G. O’Brien,*§ Yangmu Chloe Liu,*§ Mark J. Hughes,‡ John Jin Lim,§ Neil S. Hodnett,‡ Nicholas Falco§ §GlaxoSmithKline,
1250 S. Collegeville Road, Collegeville, PA, 19426, United States Medicines Research Centre, Gunnels Wood Road, Stevenage, Herts, SG1 2NY, United Kingdom Supporting Information Placeholder ‡GlaxoSmithKline,
ABSTRACT: The rate of reaction between N,N’-dicyclohexylbarbituric acid 1 and ethyl 2-isocyanatoacetate 2 is invariant with temperature. Positive orders in each reactant, and dissociation of triethylammonium salts of 1 and product 3 at elevated temperature indicate that reaction occurs via a catalytic mechanism where changes to the positions of equilibria negate changes in the rate of the turnover-limiting step. A model for the consumption of 1 in a flow reactor accurately predicted the outcome of a laboratory-scale multivariate study.
The development of pharmaceutical manufacturing processes requires a thorough understanding of chemical reactivity to ensure product quality and reliable operation on commercial scale. Determination of the driving forces for a reaction via kinetic measurements is a powerful method for obtaining this understanding. The results of kinetic studies allow reaction optimization through the development of predictive models, facilitate safe transfer into commercial manufacture, aid the control of impurities and contribute to the process understanding required for regulatory submissions.1 In this context, we performed a kinetic study of the reaction of N,N’-dicyclohexylbarbituric acid 1 with ethyl 2isocyanatoacetate 2 and triethylamine (Scheme 1a),2 to give 3 which is an intermediate in the synthesis of a clinical candidate. Development at laboratory scale was carried out in batch reactors but given the relatively short reaction time and homogeneity of the reaction mixture 3 is manufactured in a flow reactor where a solution of 1 in tetrahydrofuran is continuously mixed with 2 and triethylamine and the resulting mixture is heated to give a solution of 3 (Scheme 1b). The solution is typically carried into a series of downstream steps in which residual 1 present at the end of reaction is purged. The goal for this study was to establish a process model which could be used to predict the outcome of changes to process
ACS Paragon Plus Environment
The Journal of Organic Chemistry
parameters on residual 1 during commercial-scale operation and to inform equipment design. Studies of thermal and kinetic behavior are routinely performed early in process development and prior to scale-up3 so we began by determining the effect of temperature on reaction rate. An Arrhenius-type relationship between temperature and reaction rate is typically observed, and fitting of the data allows prediction of thermal behavior at scale. The reaction was studied under the conditions shown in Scheme 1 in a well-mixed and insulated reactor. Monitoring by HPLC provided the reaction time course. Effective control of the reactor jacket temperature mitigated any rise in temperature, ensuring that the reaction ran under essentially isothermal conditions. Preliminary experiments showed that the reaction remained fully homogenous throughout and the rate was not limited by mass-transfer.4 It was therefore considered that measurements obtained in batch reactors were fully representative of the reaction when performed in continuous flow.5 (a) OCN O N
O
CO2Et N
2 (0.37 M) N
O
N
O Et3N (0.44 M) THF, 20–100 °C
O
1 (0.26 M)
OH O
N H
CO2Et
3 (>95 % yield by HPLC)
(b)
2 heated flow reactor
been reported, the behavior has typically been attributed to either a mechanism with a single elementary step with very low activation energy6 or an enzyme-catalyzed mechanism with an off-cycle or pre-equilibrium step where changes in the position of the equilibrium effects the concentration of the active species in the turnover-limiting step.7 The latter case is commonly invoked to rationalize reversible denaturation of enzymes at elevated temperature, and has been used to rationalize complex dependences of rate on temperature in the Soai reaction8 and in Ta/Ir-catalyzed alkane–alkene coupling reactions.9 The concentration driving forces for the reaction were then determined to assess whether this behavior is consistent with either class of known examples. Variable Time Normalization Analysis (VTNA)10 allows determination of the reaction orders using a simple graphical analysis of concentration–time data for two conditions at ‘different excess’11 (see supporting information). Comparison of the normalized reaction profiles obtained by in situ IR monitoring under standard conditions with experiments at different excess (Figure 2) showed that the reaction orders in 1, 2, and triethylamine were approximately 1.1, 0.9 and 1.0 respectively. The reaction was found to be catalytic in triethylamine (complete conversion was observed with only 0.27 equivalents) highlighting that irreversible product inhibition does not occur despite the the increased acidity of 3 relative to 1.12– 14 Furthermore, VTNA returned unreasonably high orders if it was assumed that triethylamine is irreversibly consumed during reaction; a reasonable fit was obtained only if [NEt3] was treated as though it remains constant throughout the reaction.
1 in THF
0.3
3 in THF
Strikingly, only a slight increase in the reaction rate was observed between 20 and 62 °C. Furthermore, increasing the reaction temperature to 100 ºC in flow reactor at elevated pressure, showed no discernable impact on the initial rate (Figure 1). A point at which inversion of the temperature–rate relationship occurred was not found; the low solubility of 1 and 3 below 20 °C prevented the collection of reliable kinetic data below this temperature. 0.3 0.2
20 ºC
50 ºC
62 ºC
100 ºC
0.2 0.1 0 0
0.25
0.5
0.75
∑[1]1.1Δt
0.4 [2] (M)
Scheme 1. (a) Carbamoylation of 1 by reaction with 2-ethyl isocyanatoacetate; (b) Schematic for scaleup of the reaction in continuous flow.
[1] (M)
Et3N
[1] (M)
0.2 0 0
1
2
3
∑[2]0.9Δt
4
5
0.1
0.4 0 0
10
20 30 time (min)
40
50
Figure 1. Reaction progress for the consumption of 1 at varying temperatures determined by HPLC.
[1] (M)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 5
0.2 0
Examples of weak or inverse dependences of rate on temperature are uncommon and where similar examples have
ACS Paragon Plus Environment
0.00
5.00
10.00
∑[NEt3]1.0Δt
15.00
20.00
Figure 2. VTNA for the determination of orders in 1, 2, and triethylamine. Red: data under standard conditions; blue: data at different excess. Overall, these data rule out a reaction occurring via a single elementary step and are consistent with the mechanism shown in Scheme 2, wherein reversible deprotonation of 1 by triethylamine gives the nucleophilic acid–base complex 1•NEt3 which undergoes slow irreversible reaction with 2. Fast proton transfer then results in formation of 3, which exists in equilibrium with its triethylammonium salt 3•NEt3. NEt3 1
3 K3
K1
3NEt3
1NEt3 k2 2
Scheme 2. Proposed mechanism. To determine whether this mechanism is consistent with the observed invariance between temperature and reaction rate, the temperature dependence of equilibrium constants K1 and K3 was determined. 1H NMR was initially explored as a method to directly determine K1 and K3, however difficulty in resolving protonated and deprotonated forms of 1 and 3 due to fast exchange between tautomers led us to explore UV spectroscopy as a method for estimating the equilibrium constants by determining the relative concentrations of protonated and deprotonated forms at different temperatures. Solutions of 1 and 3 in THF were held the temperatures studied in Figure 1 and triethylamine was added in aliquots. Monitoring of the UV absorbance for 1•NEt3 at 262 nm and the OPC component of the UV spectrum corresponding to 3•NEt3 gave an estimate of the concentration of the deprotonated species, from which K1 and K3, both expressed as associative equilibria, could be calculated respectively (see Supporting Information). Increasing the temperature led to suppression of the signal for the triethylammonium salt (as shown for 1•NEt3 in Figure 3, left), which was reflected in a decrease in K (Figure 3, right). A shift in the equilibrium position to favor protonated 1 at elevated temperature, reducing [1•NEt3] in the turnoverliming step, could counteract any increase in k2, and would contribute to the thermal insensitivity of the reaction rate. 1 1•NEt3 0.05 0.03 0.01
1.5
15
1
10
0.5
5
0 200 250 300 Wavelength (nm)
K3 (L/mol)
0.07
K1 (L/mol)
Absorbance (mAu)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The Journal of Organic Chemistry
0 20 45 70 Temperature (°C)
Figure 3. Left: UV spectra of for 1 and excess triethylamine in THF at 20 °C (red line) and 68 °C (blue line). Right: Temperature dependence of equilibrium constants. Green
squares: K1; Red circles: K3. Both are expressed as associative equilibria. It is acknowledged that the origin of the temperature invariance may be more complex, and that the explanation presented assumes that changes in the rate due to variation in K1 and K3 effectively balance changes in rate due to variation in k2. However, the mechanism above was sufficient to enable prediction of conversion with acceptable accuracy when the reaction was performed in continuous flow at increased scale. A model for the conversion of 1 to 3 was constructed using commercially available process simulation software15 and was used to describe the reaction progress in a plug flow reactor.16 The reaction progress data above, obtained at different temperatures and different reactant excess for VTNA, were fitted to the mechanism in Scheme 2. The temperature dependences of equilibrium constants K1 and K3 can be estimated from van’t Hoff plots of the data in Figure 3 (Supporting Information) and were incorporated into the model as their free energies, The inputs for the model were the reactant feed concentrations and relative flow rates of the reagents (which together control stoichiometry), reaction time and temperature. The output of the model was the concentration of 1 at the end of the plug flow reactor under a given set of conditions. Concentrations were expressed as weight fractions (mg/g) to enable comparison with precisely established purging limits for 1 in the downstream process. Experimental data for comparison with model predictions were obtained from a statistical design of experiment (DoE) study.17 Input parameter ranges (Supporting Information) were selected to represent a wide operating winow around the target conditions. The experiments were performed in a laboratory-scale flow reactor and the concentration of 1 in the output of the reactor was determined by HPLC once the system had reached steady state as determined by repeated sampling and in-line IR spectroscopy. Simulations of these conditions were performed and the predicted and measured values for [1] in the reactor output were compared (Figure 4). Results from simulations using the model based on the mechanism in Scheme 2 were in excellent agreement with the experimentally-determined values, even at very low [1]. 12 Predicted [1] (mg/g)
Page 3 of 5
10 8 6 4 2 0 0
2
4 6 8 Measured [1] (mg/g)
10
12
Figure 4. Parity plot of model predictions against experimental data for the consumption of 1 under flow conditions. Each point corresponds to the steady-state concentration of 1 under a given set of conditions
ACS Paragon Plus Environment
The Journal of Organic Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
The agreement between simulations and experiment, together with the kinetic studies and measurements of the temperature dependence of equilibrium constants shown above suggests that the weak relationship between rate and temperature can be explained by the mechanism in Scheme 2. The model was subsequently used to support optimization of conditions and enable transfer of the process to commercial manufacture. Mechanistic understanding of the origin of this weak temperature effect was crucial for development of this robust predictive model. An alternative approach using an empirically-derived power law and fitting of temperature–rate data to the Arrhenius equation alone would be expected to give a very low activation energy with high error, making it unsuitable as a predictive tool. In conclusion, kinetic studies uncovered an unusually weak relationship between rate and temperature in the carbamoylation of 1 to give 3. A study showing positive order in all three reaction components suggested that the reaction is likely to operate through a catalytic mechanism where addition of the nucleophilic triethylammonium complex of 1 with triethylamine to 2 is the turnover limiting step. A study of the temperature dependence of equilibria between 1, 3 and their respective triethylammonium complexes found that these complexes become increasingly dissociated at higher temperatures. It is proposed that changes in the positions of these equilibria with changing reaction temperature mask changes in the rate of the turnover-limiting step.18 The findings herein were used to develop a model to accurately predict the performance of the reaction when scaled up in continuous flow, providing essential data for commercialization of the process. That the simple mechanism presented (Scheme 2) is prevalent in catalytic reactions, and the that response of equilibria to temperature is fundamentally understood indicates that ‘nonArrhenius’ behavior may be more commonplace than literature reports suggest. The effect may be overlooked unless the reaction is subjected to detailed kinetic studies. This underlines the value of mechanistic investigations in understanding even seemingly simple reactions during pharmaceutical process development.
EXPERIMENTAL SECTION Materials: N,N’-dicyclohexylbarbituric acid 1 was obtained as detailed in a previous report from our group.19 Ethyl 2isocyanaotacetate 2 is commercially available. Triethylamine was HPLC grade and anyhydrous THF (