![Page 1: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/1.jpg)
1 23
Reaction Kinetics, Mechanisms andCatalysis ISSN 1878-5190 Reac Kinet Mech CatDOI 10.1007/s11144-013-0627-5
Kinetic study of competitive catalytictransfer hydrogenation on a multi-functional molecule: 4-benzyloxy-4′-chlorochalcone
Tyler Nguyen, Julia Arciero, JoshuaPiltz, K. Danielle Hartley, TimothyRickard & Ryan Denton
![Page 2: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/2.jpg)
1 23
Your article is protected by copyright and
all rights are held exclusively by Akadémiai
Kiadó, Budapest, Hungary. This e-offprint is
for personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.
![Page 3: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/3.jpg)
Kinetic study of competitive catalytic transferhydrogenation on a multi-functional molecule:4-benzyloxy-40-chlorochalcone
Tyler Nguyen • Julia Arciero • Joshua Piltz •
K. Danielle Hartley • Timothy Rickard • Ryan Denton
Received: 22 May 2013 / Accepted: 26 August 2013
� Akademiai Kiado, Budapest, Hungary 2013
Abstract A combined chemical and mathematical approach was used to study the
reduction process of 4-benzyloxy-40-chlorochalcone involving catalytic transfer
hydrogenation with ammonium formate and palladium on carbon. Several solvents were
investigated to evaluate the effect of solvent-type on competitive reduction rates. A
mathematical model based on Michaelis–Menten reaction kinetics was developed using
a least squares approximation method to estimate several model parameters according to
time-dependent reduction data. The conjugated alkene was found to reduce fastest in all
solvents except alcohols, which is likely related to the solvent’s ability to support par-
tially-charged intermediate species in the RDS of aryl chloride hydrogenolysis. Sub-
strate concentration and pH dependence of the reduction were also investigated to
confirm prior mechanistic findings. Hydrogen transfer from the formate to palladium
appears to be rate determining in multiple reactions. Also, catalyst poisoning by HCl may
affect aryl chloride reduction more significantly than reduction of other functionalities.
Keywords Palladium � Catalytic transfer hydrogenation � Reduction � Multi-
functional � Mathematical modeling
Introduction
Catalytic transfer hydrogenation (CTH) is a proven technique in synthetic organic
chemistry that provides significant advantages over traditional hydrogenation
T. Nguyen � J. Piltz � K. D. Hartley � T. Rickard � R. Denton (&)
Department of Chemistry and Chemical Biology, Indiana University-Purdue University
Indianapolis, 402 N Blackford Street, Indianapolis, IN 46202, USA
e-mail: [email protected]
T. Nguyen � J. Arciero
Department of Mathematical Sciences, Indiana University-Purdue University Indianapolis,
402 N Blackford Street, Indianapolis, IN 46202, USA
123
Reac Kinet Mech Cat
DOI 10.1007/s11144-013-0627-5
Author's personal copy
![Page 4: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/4.jpg)
methodologies, which often involve special equipment and safety concerns in the
handling of hydrogen gas. An extensive list of inorganic and organic CTH donors
has been employed effectively. Formate salts are perhaps the most desirable
hydrogen donors because of an almost irreversible hydride donation caused by the
formation of CO2 [1]. Some reports have suggested avoiding the use of ammonium
formate due to its propensity to decompose into hydrogen gas [2], which could
potentially allow the reaction to proceed through a non-CTH mechanism. However,
it is widely accepted that little or no hydrogen gas is produced in the presence of a
suitable substrate [3, 4]. Ammonium formate was employed in this study due to its
increased solubility in organic solvents, such as formic acid, compared to sodium
and potassium formate salts [5]. CTH can utilize a wide variety of heterogeneous
and homogeneous catalysts, including palladium, platinum, and nickel species.
Palladium on carbon (Pd/C) has been recommended as the most active heteroge-
neous catalyst when using formate salts as the hydrogen donor [4]. The CTH of
numerous functionalities has been reported, including aryl (or similar) halides,
alkenes, a,b-unsaturated ketones, nitro groups, cyano groups, benzyl protecting
groups, and various carbonyl functionalities. Several in-depth reviews over the past
four decades summarize these findings [6–14].
CTH is often highlighted as a valuable technique for the selective reduction of a
single functionality on a multi-functional molecule. The ability to tune the various
reaction properties (solvent, hydrogen donor, temperature and catalyst) allows
greater chemoselectivity than many traditional methods. Solvent choice is
especially crucial since heterogeneously catalyzed hydrogenations involving
complex multiple-phase reaction systems are strongly influenced by the solvent’s
properties [15]. However, these effects are very complex, and thus relatively few
systematic studies have been performed. Rajadhyaksha and Karwa [16] summarized
several links between the following four solvent properties and their mechanistic
outcomes in heterogeneous hydrogenations: (1) different solubility of hydrogen in
the reaction mixture, (2) competitive adsorption of the solvent molecules on the
active sites of the catalyst, (3) agglomeration of catalyst particles, and (4)
nonbonding interactions between reactant or product molecules with the solvent.
The nonbonding substrate-solvent interaction was found to be most critical to the
rate when using a variety of alcohol and hydrocarbon solvents. Specifically,
increased substrate-solvent interactions were found to decrease substrate adsorption
ability. A similar study [17] used the substrate activity coefficient from the
Universal Quasi-Chemical Functional Group Activity Coefficients (UNIFAC)
contribution method [18] to explain this solvation effect on the substrate. Fajt
et al. [15] found the basic character of the solvent and its Hildebrand cohesion
energy density to be the most important solvent traits influencing hydrogenation of a
carbonyl functionality. Yet another report by Aittamaa and co-workers [19]
concluded that the hydrogenation of toluene with a nickel catalyst in various
hydrocarbon solvents was primarily dependent on hydrogen solubility in the
reaction. To summarize, solvent effects on heterogeneous hydrogenation rates have
been shown to depend primarily on the solvent used, substrate structure, and catalyst
type [20].
Reac Kinet Mech Cat
123
Author's personal copy
![Page 5: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/5.jpg)
Competitive multi-functional CTH, as presented in this study, are even more
complex because the solvent’s effect on the adsorption of one functionality will be
unique from the adsorption of a different functionality [21]. Overall, prior reports
seem to suggest that solvent effects on heterogeneous hydrogenations are specific to
the system being investigated and cannot be overly generalized. Further data on
solvent interactions in complex CTH systems may aid in bolstering these findings.
Mathematical modeling is a tool that has been used to understand the dynamics
of complex chemical interactions [22, 23]. The models tend to be complicated and
highly nonlinear, leading to challenges in solution techniques and parameter
estimation [24]. To study these systems, the models are often reduced to include
only primary reactions. For example, Black and Leff [22] used the law of mass
action to simplify a multiple pathway process for prescription drug-based agonist
binding into a single key chemical equation. Similarly, Doktorov and Kipriyanov
[23] used a simple model to examine binary chemical encounters in liquid solution
before developing a larger model. Several notable models have been utilized to
investigate the kinetics of reduction processes and solvent effects on reduction. Fajt
et al. [15] compared the Abraham–Kamlet–Taft (AKT) and Koppel–Palm (KP)
models to describe the effects of multiple solvents on the hydrogenation of 6-ethyl-
1,2,3,4-tetrahydroanthracene-9,10-dione. Glasser [25] used the Langmuir–Hinshel-
wood model to describe solvent effects on alkene hydrogenation using Pd/C. The
above models were not utilized in the current study primarily due to the lack of
hydrogen depletion data from the CTH of ammonium formate. Drougard and
Dececroocq [26] developed a model to predict solvent effects on homogeneous
catalysis reactions. However, that model is only appropriate for structurally similar
solvents and thus is not useful for the current study. Models based on Michaelis–
Menten kinetics have been applied to reduction reactions employing transition metal
catalysts in recent literature [27, 28]. These models were also employed by
Blackmond [29] in constructing a series of graphical rate equations that enable
analysis of the reaction from a minimal number of experiments; similar models will
be used in the present study.
Herein, we implement a combined chemical and mathematical approach to
investigate the reduction process of 4-benzyloxy-40-chlorochalcone, using Pd/C and
ammonium formate. The initial investigation of this reduction process was inspired
by an elegant laboratory experiment by Mohrig et al. [30]. To date, few reduction
studies address even simple chalcone systems [31–37]. The multi-functional system
presented here exhibits five potential reduction sites: benzyl protected phenol, aryl
chloride, aromatic rings, conjugated alkene and carbonyl of an a,b-unsaturated
ketone. Mathematical modeling is employed to elucidate the two most competitive
reduction processes (formation of compound Y from alkene reduction and
compound U from aryl chloride reduction, Fig. 1) as well as their secondary
reductions to form compound Z. As expected from previous reports [31], carbonyl
reduction was not observed in our study and will not be discussed further. Similarly,
reduction of the aromatic systems was not observed. Hydrogenolysis of the benzyl
ether, though observed in this study, will not be significantly discussed since it is a
thoroughly investigated deprotection strategy in organic synthesis. Our data for
hydrogenolysis of the benzyl ether (not included) supported accepted solvent
Reac Kinet Mech Cat
123
Author's personal copy
![Page 6: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/6.jpg)
reactivity trends involving increased rates with polar protic solvents like alcohols
and decreased rates for polar aprotic and non-polar solvents [38]. Furthermore, the
reported experimental data was not normalized due to the omission of benzyl ether
hydrogenolysis. The effect of diverse solvent types on these two competitive
reductions was also investigated. The mathematical model used to elucidate these
reductions is based on the law of mass action and Michaelis–Menten kinetics and is
composed of a system of ten nonlinear ordinary differential equations that can be
used to predict time-dependent changes in the concentration of each reactant or
product. Model predictions are compared with experimental kinetic data for each
reduction product and a least squares regression is used to estimate key reaction
rates in the system. Using the model results, the potential dependence of the reaction
rates on solvent polarity is discussed.
Experimental
Kinetic investigation of 4-benzyloxy-40-chlorochalcone reduction
4-Benzyloxy-40-chlorochalcone [39, 40] (compound X, 0.14 mmol), 10 % Pd/C
(Acros Organics AC19503, 10 mg, surface area 800 m2/g, particle size distribution:
d10 = 4 lm, d50 = 21 lm, d90 = 90 lm), and ammonium formate (Acros Organics
AC40115, 1.1 mmol) were reacted at atmospheric pressure and constant temper-
ature (75 �C) in 4.5 mL of various reagent grade solvents (methanol, ethanol,
acetone, ethyl acetate, tetrahydrofuran, dichloromethane, toluene, and cyclohexane)
stirring at high speeds (approximately 1,000–1,200 rpm) to minimize external mass
Fig. 1 Partial reaction network for the reduction of 4-benzyloxy-40-chlorochalcone (compound X) withammonium formate and Pd/C (Bn = –CH2–C6H5)
Reac Kinet Mech Cat
123
Author's personal copy
![Page 7: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/7.jpg)
diffusion limitations. Internal diffusion limitations were not verified due to the
broad range of palladium particle size and use of several solvents. Aliquots from the
well mixed suspension (0.5 mL) at 0.5, 1, 2, 3, 5, 10, 20, and 30 min were passed
through Celite and flushed with 1 mL of additional reaction solvent. Samples were
subjected to analysis on an Agilent 5975C Series GC/MSD using an Agilent HP-
5MS column (190915-433, 30 m 9 0.25 mm 9 0.25 lm). The reaction progress
was monitored by changes in the relative percent peak area. Variables such as the
effect of substrate concentration (0.03 and 0.06 mM) and acidic/basic additives
(1 mL of 0.1 M HCl or NaOH in ethanol solvent, performed as above) were
independently tested.
Mathematical modeling and parameter estimation
Complex reaction networks are composed of a variety of elementary steps [41]; in
particular, CTH may include: (1) chemisorption of the substrate to the catalyst, (2)
chemisorption of the hydrogen donor, (3) hydrogen transfer or hydride donation to
the catalyst surface, (4) formation of the product, and (5) desorption of the product
[4]. In the reduction of a 4-benzyloxy-40-chlorochalcone, these processes occur
simultaneously and yield intermediate products before forming a common product
Z (Fig. 1). Elementary mechanistic steps 1–3 above can be combined into a
collective ‘‘mostly reversible’’ step and steps 4–5 can be combined into a collective
‘‘mostly irreversible’’ step to yield a simplified system (Fig. 2).
Using these simplifications and the assumption that hydrogen ions (H) are in
excess, a mathematical model was developed to investigate the two pathways that
yield compound Z: (1) the initial reduction of the alkene (compound Y, rates k1 and
k2) and secondary reduction of the aryl chloride (compound Z, rates k5 and k7), and
(2) the initial reduction of the aryl chloride (compound U, rates k3 and k4) and the
secondary reduction of the alkene (compound Z, rates k6 and k8). The rate of each
elementary step in this simplified model is given by ki, with i = 1, -1, 2, 3, -3, 4,
5, -5, 6, -6, 7, and 8. Using the law of mass action, a system of ten ordinary
differential equations (ODEs) is defined that describes the rate of change of each
compound:
Fig. 2 Simplified reaction model where compounds X, Y, U, and Z are as in Fig. 1, H representshydrogen (from HCO2NH4), P is palladium, and C1–4 are catalyst-substrate complexes
Reac Kinet Mech Cat
123
Author's personal copy
![Page 8: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/8.jpg)
dx
dt¼ �k1PHX þ k�1C1 � k3PHX þ k�3C2
dp
dt¼ �k1PHX þ k�1C1 þ k2C1 � k5YPH þ k�5C3 þ k7C3 � k3PHX þ k�3C2
þ k4C2 � k6UPH þ k�6C4 þ k8C3
dH
dt¼ �k1PHX þ k�1C1 þ k2C1 � k5YPH þ k�5C3 � k3PHX þ k�3C2 þ k4C2
� k6UPH þ k�6C4
dY
dt¼ k2C1 � k5YPH þ k�5C3
dU
dt¼ k4C2 � k6UPH þ k�6C4
dZ
dt¼ k7C3 þ k8C4
dC1
dt¼ �k2C1 þ k1XPH þ k�1C1
dC2
dt¼ �k4C2 þ k3XPH þ k�2C2
dC3
dt¼ �k7C3 þ k5YPH þ k�5C3
dC4
dt¼ �k8C4 þ k6UPH þ k�6C4
ð1Þ
Positive terms in the ODEs represent formation of a product and negative terms
represent the reverse process. As mentioned above, adsorption processes to the
metal (k1, k3, k5, k6) are treated as reversible, while catalytic surface reactions and
desorption (k2, k4, k7, k8) are treated as irreversible [27]. Since hydrogen ions are in
excess, we let dH/dt = 0. The amount of palladium is assumed to be conserved (dP/
dt = 0) at each step of the reaction pathway and thus the sum of the rate of change
of catalyst concentration with time is zero:
dP
dtþ dC1
dtþ dC2
dtþ dC3
dtþ dC4
dt¼ 0 ð2Þ
Compounds Ci (for i = 1, 2, 3, 4) are assumed to form very quickly and thus are set
to their equilibrium values, that is, dCi/dt = 0. The rates of product degradation in
reversible processes were assumed to be zero (i.e., k-1 = k-3 = k-5 = k-6 = 0).
Given these assumptions, the system reduces to five ODEs for X, Y, U, Z, and H with
8 parameters. The system of ODEs is solved numerically using Matlab [42], and the
rates ki are estimated using a least squares regression to the experimental kinetic
data obtained from 4-benzyloxy-40-chlorochalcone reduction as monitored by time
changes in the relative area percentage of compounds in the reaction mixture. The
rate constants are chosen as the values that minimize the difference between the
normalized model predicted values and experimentally measured values of com-
pounds X, Y, and U at the measured time points. Experimental and model predicted
Reac Kinet Mech Cat
123
Author's personal copy
![Page 9: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/9.jpg)
values were normalized by the standard deviation of data obtained from the multiple
trials that were run for each solvent. For example:
Xexp ¼Xexp
Xstd
and X ¼ Xmodel
Xstd
ð3Þ
Y and U are defined similarly. A constrained optimization is performed to minimize
the following quantity while requiring k1 \ k2 and k3 \ k4:
X8
k¼1
X� Xexp
� �2 þ Y� Yexp
� �2þ U� Uexp
� �2 ð4Þ
This constraint was chosen according to previous reports on alkene [43] and aryl
halide [1, 4] reduction. In particular, the process by which the substrate or hydrogen
adsorb to the metal surface has been suggested as the rate-determining step (RDS).
In the present system, k1 and k3 correspond to the RDS in forming products Y and U,
and thus the model was optimized to experimental data with the constraints that
k1 \ k2 and k3 \ k4 in order to yield chemically relevant results.
Results and discussion
Reduction kinetics and modeling results
The reduction process of 4-benzyloxy-40-chlorochalcone using Pd/C and ammonium
formate was conducted using eight representative polar protic, aprotic, and non-
polar solvents: methanol, ethanol, acetone, ethyl acetate, tetrahydrofuran, dichlo-
romethane, toluene, and cyclohexane. Fig. 3 shows the relative area percentages of
compounds X, Y, U, and Z over time for the aforementioned solvents. For clarity, the
data were divided into two groups: non-polar solvents (tetrahydrofuran, dichloro-
methane, toluene and cyclohexane), which are shown in the left column of Fig. 3
(panels A, C, E, and G), and polar solvents (methanol, ethanol, acetone, and ethyl
acetate), which are depicted in the right column of Fig. 3 (panels B, D, F, and H).
As stated earlier, benzyl ether hydrogenolysis data was not included and the
remaining reduction data was not normalized. It should be noted that the
concentration maxima observed for compound Z derives from further reduction
processes (e.g. benzyl ether hydrogenolysis) that were not modeled in this study.
Fig. 4 shows the model fits to the reduction data for a representative solvent.
Model fits for the remaining solvents follow similarly. The optimized reaction rates
for the RDS are provided in Table 1 for all modeled solvents. The initial slopes of
the raw data (Fig. 3) agree with the model-predicted reaction rates, indicating that
the model can be used to predict the reduction rates with good accuracy.
Dichloromethane and cyclohexane rates were not included since the collected
experimental reaction data did not show reaction completion (Fig. 3, left column).
As justified previously, k1 and k3 were constrained to be less than k2 and k4,
respectively. Given these constraints, the RDS for the formation of compound Z are
k5 and k6, from compounds Y and U, respectively.
Reac Kinet Mech Cat
123
Author's personal copy
![Page 10: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/10.jpg)
Fig. 3 Experimental data for relative area percentage of compounds X, Y, U, and Z over reaction time.Panels A, C, E and G in the left column represent non-polar solvents: tetrahydrofuran (THF),dichloromethane (DCM), toluene, and cyclohexane. Panels B, D, F, and H in the right column representpolar solvents: methanol, ethanol, ethyl acetate, and acetone
Reac Kinet Mech Cat
123
Author's personal copy
![Page 11: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/11.jpg)
The coefficient of determination (R2) values for compounds X, Y, and U were
calculated between 0.73–0.94 for the modeled solvents (Table 2). A similar study
by Bernas and Myllyoja [27] gave R2 values of 0.7–0.9. It should be reiterated that
Fig. 4 Model fits (solid curves) to experimental data (black dots) for compounds X, Y, and U in ethanol
Table 1 Modeled rates for the formation of Y, U, and Z (% concentration/min)
Compound (RDS) Methanol Ethanol Acetone Ethyl acetate Tetrahydrofuran Toluene
Y (k1) 4.6 14.6 13.3 33.3 20.8 14.9
U (k3) 16.1 22.3 5.8 11.3 2.0 0.9
Z (k5) 32.7 75.3 25.5 91.6 36.5 36.0
Z (k6) 135.3 141.8 114.3 351.6 26.3 36.3
Table 2 The coefficient determination (R2) values for X, Y, and U in all modeled solvents
Compound Methanol Ethanol Acetone Ethyl acetate Tetrahydrofuran Toluene
X 0.8713 0.8425 0.9425 0.9023 0.8984 0.8629
Y 0.8874 0.8582 0.8375 0.8348 0.8543 0.7749
U 0.8343 0.8905 0.8904 0.7266 0.8388 0.7653
Reac Kinet Mech Cat
123
Author's personal copy
![Page 12: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/12.jpg)
model accuracy is limited by the large number of parameters being fit to the data as
well as the limited amount of experimental data available.
Reaction and mechanistic implications
Due to previous mechanistic findings for RDS in similar aryl halide hydrogenolyses
[1] and hydrogenations of a,b-unsaturated ketones [43] the model was constrained
to k1 \ k2 and k3 \ k4 for the formation of compounds Y and U, respectively.
Somewhat expectedly, the secondary reductions showed the rate determining
process to coincide with the ‘‘mostly reversible’’ adsorption step (i.e. k5 and k6).
It was also apparent that secondary reduction rates were higher than primary
reduction rates (i.e. alkene reduction k6 [ k1, aryl chloride reduction k5 [ k3).
Though an argument can be made for increased reactivity of compound U toward
alkene reduction versus compound X, increased reactivity of compound Y toward
aryl chloride reduction versus compound X would not be completely predictable.
Therefore, it is very possible that this result is caused by a saturation effect of
hydrogen on the Pd/C catalyst surface. Presaturation of the catalyst is known to
increase aryl chloride hydrogenolysis rates [2]. In addition, Suzuki and Suzuki [44]
suggest that once significant amounts of hydrogen have transferred to palladium,
excess hydrogen then migrates to the carbon surface and is easily accessible for
reduction. This result may also confirm that hydrogen transfer from the bound
formate to palladium is the RDS in both reductions, as has been previously
suggested with aryl chloride reduction [4]. Solvents also seem to play a role in this
process, as secondary alkene reduction rate increases (k6 [ k1) are most notable in
polar solvents, while secondary aryl chloride reduction rate increases (k5 [ k3) are
most notable in non-polar solvents.
When comparing initial reductions in Table 1, formation of compound U (aryl
chloride hydrogenolysis, k3) is faster than formation of compound Y (alkene
reduction, k1) in alcohol solvents, which are known to increase aryl chloride
reduction rates [2]. It should also be noted that alcohols have been known to act as
hydrogen donors in CTH reactions [45, 46] which may explain why some reductions
are predominantly fastest in ethanol. Observing the higher reduction rate of an aryl
chloride is a somewhat unexpected result, though no similar systematic study
between these two functionalities is known to the authors. With all other solvents,
alkene reduction rate was higher, especially in non-polar solvents. On the contrary,
when comparing secondary reduction rates, alkene reduction rate (k6) is higher than
aryl chloride hydrogenolysis rate (k5) in polar solvents and in non-polar solvents.
The initial reduction of aryl chlorides, with an unsaturated Pd/C catalyst, may place
more importance on solvent interactions in the hydrogen transfer/CO2 decompo-
sition step. It is also possible that the formation of HCl as a byproduct of aryl
chloride reduction is irreversibly poisoning the Pd/C toward further aryl chloride
reduction, even in polar solvents, whereas alkene reduction is less affected by HCl
poisoning [4].
Reac Kinet Mech Cat
123
Author's personal copy
![Page 13: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/13.jpg)
Effect of solvent polarity on rate
Initial reactivity trends seemed to suggest that solvent polarity might play a
significant factor in the RDS, potentially elucidating the increased rate of formation
of compound U compared to compound Y in alcohol solvents. To identify possible
trends between solvent type and reduction, the model predicted reaction rates are
graphed as a function of the solvent polarity index [47] and dielectric constant [48]
in Fig. 5 (constant values in Table 3). Panels A and B compare initial reductions (k1
and k3) and panels C and D compare secondary reductions (k5 and k6).
Table 3 Polarity indexes [47] and dielectric constants [48] of the modeled solvents
Methanol Ethanol Acetone Ethyl acetate Tetrahydrofuran Toluene
Polarity index 6.6 5.2 5.1 4.3 4.2 2.3
Dielectric constant 32.6 24.3 20.7 6.02 7.4 2.44
Fig. 5 Modeled reaction rates (% concentration/minute) graphed as a function of polarity index anddielectric constant. A k1 and k3 versus polarity index. B k1 and k3 versus dielectric constant. C k5 and k6
versus polarity index. D k5 and k6 versus dielectric constant
Reac Kinet Mech Cat
123
Author's personal copy
![Page 14: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/14.jpg)
All reduction rates investigated were highest in either ethanol or ethyl acetate,
with acetone and methanol lagging behind. Acetone rates (as compared to ethyl
acetate) and methanol rates (as compared to ethanol) were notably low when
considering the similar polarity index values. The higher dielectric constant values
in those pairs resulted in lower rates, which may result from a solvent-substrate
interaction preventing catalyst binding, for instance. Reduction rates for non-polar
solvents were generally lower. It should be noted that ammonium formate solubility
may be a factor of some importance. Of particular interest is the rate of k3 as
compared to k1 at high dielectric constants, which was not as prominent when
graphed as a function of polarity index (see Fig. 5, panel B). It could be concluded
that a solvent’s ability to support partially-charged intermediate species in the RDS
of aryl chloride reduction (k3, likely hydrogen transfer to palladium from formate) is
crucial. This appears to be less important for k1, which opens the possibility for an
alternate hydrogen donation mechanism or other compounding solvent-substrate
interactions that would mitigate this hypothesis. By observing the lack of such a
trend in k5 and k6, it appears that accumulation of hydrogen on the catalyst surface
minimizes this effect in aryl chloride reduction.
Substrate concentration
Time course experiments with two concentrations (0.03 and 0.06 mM) of compound
X were performed using the procedure described above to elucidate potential
differences in relative rates of reduction. Acetone was chosen as the solvent because
modeled rates for the formation for Y and U were somewhat similar. The formation
of compounds Y, U, and Z were relatively consistent between substrate concen-
trations, which further supports that the starting material was not involved in the
RDS for either reduction process. However, it should be noted that hydrogenolysis
of the benzyl ether was observed only in small amounts at 0.03 mM of compound X,
yet was accumulated in significant amounts at 0.06 mM of compound X. Benzyl
ether hydrogenolysis was ultimately occurring at both concentrations as evidenced
by further reduced products, but no accumulation of the phenol product was
Fig. 6 Acidic and basic additives show significant effects on reaction rate (A Compound Y, B CompoundU)
Reac Kinet Mech Cat
123
Author's personal copy
![Page 15: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/15.jpg)
observed at 0.03 mM. This data suggests that hydrogenolysis of the benzyl ether
may involve the substrate in the rate determining process.
Acidic and basic additives
From trends noted above, k5 (aryl chloride hydrogenolysis) is comparatively slower
than k6 (alkene reduction) during the secondary reduction than the primary
reduction (k3 vs. k1). One of the potential explanations involves HCl as a known
poison to the Pd/C catalyst [4]. Therefore, the reductions were performed with
aqueous acidic and basic additives to investigate this possibility (Fig. 6). As
evidenced in panel A, the addition of increased levels of HCl does not initially affect
the reduction of compound Y compared to ethanol. The addition of NaOH
significantly affects reduction in panel a, but likely because of the competitive rate
increase of compound U with the addition of NaOH in panel B. This supports results
that HCl poisoning of Pd/C is reversible with an aqueous base [4]. The addition of
HCl in panel B produced little increased poisoning effect compared to HCl
produced from aryl chloride reduction. It should also be mentioned that the effect of
water concentration on CTH reduction of the a,b-unsaturated ketones [49] and aryl
chlorides [4] is of some importance, which was not specifically explored here. As
the poisoning effects of HCl in aryl chloride hydrogenolysis have been well
established, we only sought to confirm or refute its presence in our study.
Conclusions
The relative rates of reduction of CTH of 4-benzyloxy-40-chlorochalcone were
modeled for the two most competitive processes, where conjugated alkene reduction
was found to be faster than the aryl chloride reduction in all solvents except
alcohols. Initial trends indicated that higher reaction rates correlated with increased
solvent polarity, though generalization of solvent effects on CTH to a single solvent
trait was only possible in a few instances. Secondary reductions to form compound
Z were significantly faster than subsequent primary reductions, which supports the
accumulation of hydrogen on the catalytic surface, where hydrogen transfer from
the formate to palladium was the RDS. To that end, substrate adsorption was
confirmed not to be rate determining in either reduction reaction. Catalyst poisoning
by HCl affects aryl chloride reduction significantly more than conjugated alkene
reduction.
Acknowledgments We would like to acknowledge Jeremy Martin, Tumare Iqbal, leaders/participants
from the 2010 CWCS Teaching Guided-Inquiry Organic Chemistry Labs Workshop and C344 students
from 2011-2013 for their valuable contributions to this work.
References
1. Rajagopal S, Spatola AF (1995) J Org Chem 60:1347–1355
2. Anwer MK, Sherman DB, Rooney JG, Spatola AF (1989) J Org Chem 54:1284
Reac Kinet Mech Cat
123
Author's personal copy
![Page 16: jarciero/Publication_14.pdf · solvents except alcohols, ... However, that model is only appropriate for structurally similar solvents and thus is not useful for the current study](https://reader033.vdocuments.us/reader033/viewer/2022042001/5e6d91b7612f2862547d77f6/html5/thumbnails/16.jpg)
3. Bar R, Sasson Y (1981) Tetrahedron Lett 22:1709
4. Wiener H, Blum J, Sasson Y (1991) J Org Chem 56:6145–6148
5. Groschuff E, Ber D (1903) Chem Ges 36(1783):4351
6. Brieger G, Nestrick TJ (1974) Chem Rev 74:567–580
7. Haertner H, Knotakte D (1980) J Mol Catal 1:3–10
8. Johnstone RAW, Wilby AH, Entwistle ID (1985) Chem Rev 85:129–170
9. Ram S, Ehrenkaufer RE (1988) Synthesis 91:5
10. Zini CA, vonHolleben MLA (1992) Quimica Nova 15:40–54
11. Rajagopal S, Anwer MK, Spatola AF (1994) Peptides : design, synthesis, and biological activity.
Birkhauser, Boston
12. Zheng C, Zhang J, Wang R (2004) Gongye Cuihua 12:29–35
13. Mehta NH, Manyar HG, Pawar NN, Gham NO (2004) Chem Indust Digest 17:55–60
14. Su C, Zheng C, Wang R, Zhang J (2004) Huaxue Tongbao 67:731–739
15. Fajt V, Kurc L, Cerveny L (2008) Int J Chem Kinet 40:240–252
16. Rajadhyaksha RA, Karwa SL (1986) Chem Eng Sci 41:1765–1770
17. Lo HS, Paulaitis ME (1981) Am Inst Chem Eng J 27:842–844
18. Fredenslund A, Jones RL, Prausnitz JM (1975) Am Inst Chem Eng J 21:1086–1099
19. Rautanen PA, Aittamaa JR, Krause AOI (2000) Ind Eng Chem Res 39:4032–4039
20. Cerveny L, Ruzicka V (1981) In Advances in Catalysis, 1st edn. Academic Press, New York
21. Cerveny L, Ruzicka V (1982) In catalysis reviews—science and engineering, 1st edn. Marcel Dekker,
New York
22. Black JW, Leff P (1983) Proc R Soc London B202:141–162
23. Doktorov AB, Kipriyanov AA (2007). J Phys Condens Matter. doi:10.1088/0953-8984/19/16/065136
24. Oklno MS, Mavrovouniotis ML (1998) J Am Chem Soc 98:391–408
25. Glasser L (1979) J Chem Ed 56:22–23
26. Drougard Y, Decroocq D (1969) Bull Soc Chem Fr 9:2972–2983
27. Bernas A, Myllyoya J, Salmo T, Murzin DY (2008) J App Cat 353:166–180
28. Kohler Agnes Gamez Jochem, Bradley John (1998) Catal Lett 55:73–77
29. Blackmond Donna G (2005) Angew Chem Int Ed 44:4302
30. Mohrig JR, Hammond CN, Schatz PF, Davidson TA (2009) J Chem Ed 86:234–239
31. Goldberg R, Glen J (1994) Ind Eng Chem 33B:163–165
32. Keinam E, Greenspoon N (1986) J Am Chem Soc 108:7314
33. Nielsen SF, Kharazmi A, Christensen SB (1998) Bioorg Med Chem Lett 6:937
34. Sajiki H, Ikawa T, Yamada H, Tsubouchi K, Hirota K (2003) Tetrahedron Lett 44:171
35. Basu B, Bhuiyan MH, Das P, Hossain I (2003) Tetrahedron Lett 44:8931
36. Tian F, Lu S (2004) Syn Lett 39:1953–1956
37. Andrade CKZ, Silva WA (2006) Lett Org Chem 3:39–41
38. Kocienski PJ (2005) Protecting groups. Thieme, Munster
39. Rashad AA, El-Sabbagh OI, Baraka MM, Ibrahim SM, Pannecouque C, Andrei G, Snoeck R,
Balzarini J, Mostafa A (2010) Med Chem Res 19:1025–1035
40. Robinson TP et al (2005) Mioorg Med Chem Lett 13:4007–4013
41. Rader CP, Smith HA (1961) J Am Chem Soc 84:1443–1449
42. MathWorks (2013) Matlab. vol R2013, Student Version edn
43. Ram S, Spicer LD (1992) Synth Commun 22:2683–2690
44. Suzuki S, Suzuki T (1965) Japanese J Sci Tech 38:2020
45. Kleiderer EC, Kornfield EC (1948) J Org Chem 13:455–458
46. Andrews MJ, Pillai CN (1978) Ind J Chem 16:465
47. Snyder LR (1978) J Chromatogr Sci 16:223–234
48. Maryott A, Smith E (1951) Dielectric constants of pure liquids. In: Linstrom PJ, Mallard WG (eds)
NIST chemistry webBook, NIST standard reference database number 69. National Bureau of
Standards, Washington, DC http://webbook.nist.gov
49. Li J-P, Zhang Y-X, Ji Y (2008) Selective 1,4-reduction of chalcones with Zn/NH4Cl/C2H5OH/H2O.
J Chin Chem Soc 55:390–395
Reac Kinet Mech Cat
123
Author's personal copy