close

Вход

Забыли?

вход по аккаунту

?

Expanding the scope and utility of the scientific microwave apparatus in organic synthesis: Reaction monitoring, scale-up and new methodology development

код для вставкиСкачать
Expanding the Scope and Utility of the Scientific Microwave Apparatus in Organic
Synthesis: Reaction Monitoring, Scale-Up and New Methodology Development
Jason R. Schmink, Ph.D. Thesis
University of Connecticut, 2010
Abstract
The work contained within this dissertation covers three broad areas in the
continued development of the scientific microwave apparatus. A brief introduction of
pertinent fundamental microwave background and theory is presented in Chapter 1, and
a short background of microwave-assisted organic synthesis is described, including
areas of recent growth, its widespread use, and contains a discussion of the range of
equipment available to the organic chemist.
Three fundamental research areas are discussed. First, the interface of a Raman
spectrometer with a commercially available scientific microwave apparatus is discussed
in detail. The apparatus was used both qualitatively as a reaction-monitoring tool and as
a quantitative tool for rapid kinetic studies. Second, the scale-up of microwave mediated
organic reactions is discussed. Within the pharmaceutical industry, a great number of
synthetic protocol are developed utilizing the scientific microwave. The scope and
limitations of utilizing microwave heating, especially in batch reactors, is thoroughly
examined and presented within this work. Finally, the development of novel synthetic
methodologies utilizing microwave heating is presented.
Expanding the Scope and Utility of the Scientific Microwave Apparatus in Organic
Synthesis: Reaction Monitoring, Scale-Up and New Methodology Development
Jason R. Schmink
A Dissertation
Submitted in Partial Fulfillment of the
Requirements for the Degree of
Doctor of Philosophy
at the
University of Connecticut
2010
ti
UMI Number: 3415563
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
UMT
Dissertation Publishing
UMI 3415563
Copyright 2010 by ProQuest LLC.
All rights reserved. This edition of the work is protected against
unauthorized copying under Title 17, United States Code.
ProQuest LLC
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Copyright by Jason R. Schmink
2010
APPROVAL PAGE
Doctor of Philosophy Dissertation
Expanding the Scope and Utility of the Scientific Microwave Apparatus in Organic
Synthesis: Reaction Monitoring, Scale-Up and New Methodology Development
Presented by
Jason R. Schmink
Major Advisor_
cnbtesJL_Leadbeater
Associate Advisor
Associate Advisor
Associate Advisor
Associate Advisor
£u0 hJ$HiL
Ron Wikholm
Associate Advisor
Edward J. Neth
University of Connecticut
2010
In Memory of
Dr. Paul David Schmink
A man who imbued in me the importance of education, impressed upon me the
difference between knowledge and intelligence, inspired me to demand more of myself
than others do of me, and lastly to always "go look it up!"
Acknowledgements
First and foremost, I'd like to thank my advisor, Prof. Nicholas E. Leadbeater for
giving me opportunities to succeed in the laboratory. The rapport he nurtured both with
me and among the rest of his research group made it easy to come to work each day,
even when the chemistry was not cooperating.
I'd also like to thank Prof. Tyson A. Miller, who convinced me to spend my
doctoral years at the University of Connecticut. I appreciate his passion in the
classroom, his work ethic, and his willingness to have discussions concerning:
chemistry, education, chemical education, graduate life, life after graduate life, and
anything else at all that might be on one's mind.
Next, I am extraordinarily fortunate to have Prof. Ron Wikholm 'in my corner.' A
few years back, he took a bit of a gamble in allowing a first-semester graduate student
act as a substitute lecturer on his behalf. While his decision has likely benefited us both,
I feel like I got the better end of the deal. Without a doubt, the opportunity to deliver
numerous general and organic chemistry lectures has been of great benefit to my
development as a lecturer and instructor, and has likely given me a leg up on my peers
in terms of experience.
I have been fortunate to have a number of fantastic teachers and mentors over
the duration of my time here. To Profs. Peczuh, Birge, Sotzing, Bailey, Neth, and
Bruckner: thank you for your time, advice, knowledge, and support. It is all too common
to hear students/former students bemoan their time as a graduate student, but thanks in
large part to the fantastic mentors that have surrounded me, I can easily say the past
four years have been the most exciting and enlightening thus far in my life. Again,
thanks!
Having a solid support system in any endeavor is of utmost importance. Though
there are countless names that could be listed here, I'd be remiss in not including the
vi
following. Dr. Martha Morton runs a tight ship down in the NMR laboratory and without
the smooth operation of something so fundamental to my day-to-day discoveries,
significantly less would likely been accomplished on my part. Thanks, Martha! In a
similar vein, Charlene Fuller needs to be commended for her excellent job quite literally
keeping this building in operation. In my opinion, it is unlikely that there is any single
individual in the entire department that has as much impact on the overall productivity
and scientific contributions as Charlene. Thank you, Charlene!
I have had the opportunity to collaborate with a number of scientists during my
time at UCONN. First, a big thanks to Dr. Matthew Bowman (Doc) for pushing harder
and demanding more than I had been accustomed to: a key factor in my scientific
growth. Next, thank you to all of the undergraduate researchers for all of their hard work
and contributions to my research, including: Jen Holcomb, Kristen Amore, Will Devine,
Samantha Gibson, Joe Podgriel, and Neha Grewal. You may not realize it, but I have
often learned as much from you guys as you have (hopefully) learned from me. It has
been a joy working alongside you as a mentor!
Though I owe a big thank you to all members and former members of the
Leadbeater lab, I owe a very special thank-you to my friend, colleague, roommate,
brewer-in-kind, and scientist Chad M. Kormos. Without hesitation, I can say that no
single individual has had as much impact on my development as a chemist as Chad. It
has been a great joy to volley ideas with him: thinking, creating, and occasionally
succeeding!
Finally, thanks Mom!
VII
TABLE OF CONTENTS
Prologue
List of Publications
List of Abbreviations
Chapter 1 Introduction to Microwave Heating
1.1 Microwave Theory
1.2 "Microwave Effects"
1.3 Microwave Assisted Organic Synthesis
1.4 Small Scale Equipment
1.5 Conclusions & Outlook
Chapter 2 Raman Spectroscopy Monitoring of Microwave Assisted Reactions
2.1 Introduction
2.2 Raman Spectroscopy Fundamentals
2.3 Raman Spectroscopy, Pros and Cons
2.4 Overcoming Disadvantages of Raman Spectroscopy
2.5 Raman-Microwave Interface
2.6 Qualitative Reaction Monitoring
2.7 Quantitative Reaction Monitoring: Coumarin
2.8 Mechanistic Insight via Raman Spectroscopy & Computational Modeling
2.9 Mechanistic Insight: Kinetic Studies of a Non-Cyclizable Analogue
2.10 Reaction Monitoring: Biginelli Reaction
2.11 Quantitative Reaction Monitoring: Chalcones
2.12 Conclusions & Future Outlook
Chapter 3 Investigations Into "Non-Thermal" Microwave Effects
3.1 Background & Rationale
3.2 Specific Microwave Effects
3.3 Microwave Absorptivity
3.4 Inverted Temperature Gradients
3.5 Macroscopic Superheating
3.6 Selective Heating
3.7 Non-Thermal Microwave Effects
3.8 Raman Spectroscopic Investigations Into "Microwave Effects."
3.9 Conclusions & Future Outlook
viii
Chapter 4 Scale up of Microwave-Mediated Transformations
4.1 Introduction to Scale-Up Chemistry
4.2 Typical Large Scale Equipment
4.3 Continuous Flow Approach
4.4 Stop-flow Approach
4.5 Multi-Vessel Rotor Approach
4.6 Batch Approach (Open Vessel)
4.7 Batch Approach (Sealed Vessel)
4.8 Microwave-Mediated Transformations: Scale Up Equipment
4.9 Rationale of Transformations Investigated
4.10 Large-Scale Suzuki Reactions
4.11 Condensation Reactions: 3-acetylcoumarin and Biginelli Reactions
4.12 Williamson Etherification
4.13 Aza-Michael Synthesis & Claisen Rearrangement as Heating Efficiency Protocol
4.14 Four-Step Reaction Sequence
4.15 Large-Scale Batch Solvent Heating in AccelBeam Prototype
4.16 Conclusions & Outlook
Chapter 5 Pd-Catalyzed Methodology Development: Synthesis of Diarylmethanes
5.1 Diarylmethane Background & Significance
5.2 Palladium-Enolate Coupling Background & Significance
5.3 Colloidal Palladium Background
5.4 Diarylmethane Synthesis: Initial Discovery & Screening
5.5 Putative Reaction Mechanism (Qualitative Observations).
5.6 Initial Diarylmethane Kinetic Studies
5.7 Conclusions & Future Outlook
Appendix 1. Detailed Raman Signal Strength-Concentration Correlations
Appendix 2. Interface of Raman Spectrograph & Microwave; Pertinent Equipment Data
Appendix 3. Insertion of d 10 Metals into TPP
Appendix 4. Aza-Michael Addition of Anilines
Appendix 5. Synthesis of Benzophenones, Initial Studies
Appendix 6. Investigations into the Synthesis of Tamoxifen
Appendix 7. Detailed Experimental Protocol
Appendix 8. Relevant Spectra and Spectral Data
About the Author
ix
Prologue
To say the least, this thesis covers a broad spectrum of science huddled together
under the unifying umbrella of "microwave chemistry." I put 'microwave chemistry' within
quotes, as I do not believe there is such a field as "microwave chemistry" any more than
there are fields of "hot plate chemistry" or "magnetic stir bar chemistry." The scientific
microwave apparatus certainly is an enabling tool that can increase productivity and
allows the organic chemist to explore possibilities otherwise not easily accessible, but it
is only a convenient piece of equipment, much like a separatory funnel.
In the broadest strokes, this thesis focuses on ways to make the scientific
microwave apparatus even more utilitarian for the organic chemist. As evidenced by the
title, I have broken my work into three sub-categories:
• In-situ reaction monitoring
• Scale-up of microwave heated organic reactions
• New methodology development
For me, the PhD thesis holds a number of important, yet highly varied purposes.
First and foremost, the thesis should be a document that exhibits to my scientific peers
that I am able to think logically, approach problems in a scientific manner, and present
my findings in a clear and concise fashion. Second, while much of the work contained
within has already been published, the thesis allows me to report on some findings that
otherwise likely would not see the light of day. Next, the thesis should be a valuable
future reference both for myself as well as for researchers within the Leadbeater
research group. The last—though certainly not the least important—purpose of this
thesis is to provide Prof. Leadbeater and his future group members with leads on new
projects and obvious extensions of current work. At some point as a scientist, you realize
that every question that you are able to answer leads to three new questions, so there is
never a lack of research to do and no project is ever, truly finished.
As I began the writing process and thought back to my time in the lab, I realized
that I have been extraordinarily fortunate. I have been exposed to a significant amount of
analytical chemistry, engineering and troubleshooting, physical chemistry, computational
chemistry, and of course organic chemistry in my four years here. Most of the work
presented in this thesis has been reported in a number of publications. Page 4 lists
these publications, grouping them by the relevant chapter where these projects are
discussed. The most difficult part of my task will be to tie this wide range of topics into a
single, coherent story line.
' The chapters generally begin with some background and an indication as to why
this chemistry was explored. After presenting my findings, I end with a "looking forward"
section which focuses on where I think this chemistry should logically move next, overall
scope an limitations to the given topic, and generally contains a bit more opinion and/or
qualitative observations.
The Appendices are utilized in a couple of different situations. In some cases, I
used an appendix when there was a need to go into great detail, but feared that this
minutia likely would interrupt the flow of the discussion (Appendices 1 & 2).
Appendix 3 contains a bit of work that was done in collaboration with Michelle
Dean and Prof. Christian Bruckner. I would by no means classify it as "my" chemistry,
though I learned a bit, and therefore thought it would be a good idea to pen a very brief
summary of this collaborative work, especially regarding my contributions.
Appendix 4 discusses some chemistry that I developed in my first few months in
the Leadbeater research group. It was a short project and not thoroughly investigated by
any means, therefore I didn't think it warranted an entire chapter within the thesis. That
said, this project was my first foray into research chemistry, I obtained some interesting
results which led to my first published work, therefore I thought it should be presented in
the thesis and submitted to the committee (and to posterity!).
2
Appendix 5 details a project that is closely related to the diarylmethane project,
though it had to be put on hold it in order to continue the kinetic studies portion of the
diarylmethane chapter. It is quite an interesting bit, a logical extension of the Pd-enolate
chemistry, and is a methodology project that I feel is "this close" to getting over the hump
in terms of development! This would be a great project for an incoming graduate student
or an advanced undergraduate researcher to continue with in the future.
Appendix 6 outlines a bit of a "thought project," investigated in conjunction with
Chad M. Kormos after a particularly stimulating group meeting: a proposed synthesis of
Tamoxifen, the most widely prescribed breast cancer medication. While it was ultimately
doomed, a great deal of insight was gleaned from this project, a deeper understanding of
palladium catalysis was fostered, and ultimately it led me to the Pd-enolate chemistry,
including the synthesis of diarylmethanes. This project also introduced me to the joys of
sifting through chemistry patent literature and I gained considerable experience trying to
extract the pertinent bits of chemistry flotsam out of the vast oceans of text. All in all,
though this particular project didn't pan out, it taught me quite a bit of chemistry along
the way, and thus warranted a few pages of discussion.
Finally, Appendices 7 & 8 are collections of detailed experimental protocol (7)
and relevant spectral information (8) for structures throughout the thesis.
Well, that is about it. I hope you find the overall format logically organized,
straightforward, easy to read, and enjoyable. Most of all, however, I hope you learn
something! I will leave the first-person active voice in favor of the passive and leave you
to what I hope to be an enlightening read.
3
List of Publications by Chapter
Chapter 1
Schmink, J. R.; Leadbeater, N. E. "Microwave heating as a tool for sustainable
chemistry: An introduction" in Microwave Heating as a Tool for Sustainable Chemistry,
N. E. Leadbeater, Editor, CRC Press, Boca Raton, Florida, 2010.
Chapter 2
Schmink, J. R.; Holcomb, J. L.; Leadbeater, N. E. Testing the validity of microwaveinterfaced, in situ Raman spectroscopy as a tool for kinetic studies, Organic Letters,
2009,11,365-368.
Schmink, J. R.; Holcomb, J. L.; Leadbeater, N. E. Raman spectroscopy as an in-situ tool
to obtain kinetic data for organic transformations, Chemistry, A European Journal, 2008,
14, 9943-9950.
Leadbeater, N. E.; Schmink, J. R. Use of Raman spectroscopy as a tool for in situ
monitoring of microwave-promoted reactions, Nature Protocols, 2008, 1-7.
Chapter 3
Schmink, J. R.; Leadbeater, N. E. Probing "microwave effects" using
spectroscopy, Organic and Biomolecular Chemistry, 2009, 7, 3842-3846.
Raman
Chapter 4
Schmink, J. R.; Kormos, C. M.; Devine, W. G.; Leadbeater, N. E. Exploring the scope of
scale-up of organic chemistry using a large batch microwave reactor, Organic Process
Research & Development, 2010, 14, 205-214.
Bowman, M. D.; Kormos, C. M.; Leadbeater, N. E.; McGowan, C. M.; Schmink, J, R.
Scale-up of microwave-promoted reactions to the multigram level using a sealed-vessel
microwave apparatus, Organic Process Research & Development, 2008, 12, 1078-1088.
Chapter 5
Schmink, J. R.; Leadbeater, N. E. Palladium-catalyzed synthesis of diarylmethanes:
Exploitation of carbanionic leaving groups, Organic Letters, 2009, 11, 2575-2578.
Appendices
Bruckner, C ; Dean, M. L.; Leadbeater, N. E.; Schmink, J. R. Microwave-promoted
insertion of group 10 metals into free base porphyrins and chlorins: scope and
limitations, Dalton Transactions, 2008, 1341-1345
Leadbeater, N. E.; Schmink, J. R. Use of a scientific microwave apparatus for rapid
optimization of reaction conditions in a monomode function and then substrate screening
in a multimode function, Tetrahedron, 2007, 63, 6764-6773.
Amore, K. M.; Leadbeater, N. E.; Miller, T. A.; Schmink, J. R. Fast, easy, solvent-free,
microwave-promoted Michael addition of anilines to a,(3-unsaturated alkenes: synthesis
of A/-aryl functionalized beta-amino esters and acids, Tetrahedron Letters, 2006, 47,
8583-8586.
4
List of Abbreviations
Ac
acac
Ar
BBN
BHT
BINAP
Bn
bp
Bu
CCD
CD
conv.
Cp
d
DABCO
dba
DBU
DCM
DIBAL-H
DMF
DMP
DMSO
DPPF
DVD
dppe
z'
t"
Ea
EDG
EM
Et
EWG
FDA
FTIR
GC
HPLC
HRMS
IR
ISM
JOC
KHDMS
m
Me
MEK
mp
MW
NA
NBS
NHC
acetyl
acetylacetonyl
aryl
borabicyclo[3.3.1]nonane
butylated hydroxytoluene (2,6-di-f-butyl-4-methylphenol)
2,2'-bis(diphenylphosphino)-1,1 '-binaphthyl
benzyl
boiling point
n-butyl
charge coupled device
compact disk
Conversion
cyclopentadienyl
doublet
1,4-diazabicyclo[2.2.2]octane
dibenzylidenacetone
1,8-diazabicyclo[5.4.0]undec-7-ene
dichloromethane
diisobutylaluminum hydride
A/,A/-dimethylformamide
Dess-Martin periodinane
dimethylsulfoxide
1,1'-bis(diphenylphosphino)ferrocene
digital versatile (video) disk
1,2-bis(diphenylphoshphino)ethane
dielectric constant
dielectric loss
activation energy
electron donating group
electromagnetic
ethyl
electron withdrawing group
United States Food & Drug Administration
Fourier transform infrared spectroscopy/spectrum
gas chromatography
high performance liquid chromatography
high resolution mass spectrometry
infrared
industrial, scientific, medical
Journal of Organic Chemistry
potassium hexamethyldisilazide
multiplet
methyl
methyl ethyl ketone, 2-butanone
melting point
microwave
nominal aperture
A/-bromosuccinimide
A/-heterocylclic carbene
NIR
NMP
NMR
OL
OPRD
Ph
s
t
q
tan 5
TBAB
TET
Tf
TFA
TFAA
THF
TOF
TL
TLC
TPP
TsOH
near infrared
A/-methyl-2-pyrrolidinone
nuclear magnetic resonance
Organic Letters
Organic Process Research & Development
phenyl
singlet
triplet
quartet
loss tangent, loss angle
tetrabutylammonium bromide
Tetrahedron
trifluoromethansulfonyl
trifluoroacetic acid
trifluoroacetic anhydride
tetrahydrofuran
time of flight
Tetrahedron Letters
thin layer chromatography
tetraphenylporphyrin
p-toluenesulfonic acid
6
Chapter 1B
Introduction to Microwave Heating
1.1 Microwave Theory. The microwave region of the electromagnetic spectrum
(Figure 1-1) is broadly defined as the wavelength region between 0.1-100 cm,
corresponding to a frequency between 300-0.3 GHz. As this frequency range includes
applications such as wireless devices (2.4 and 5.0 GHz, US), satellite radio (2.3 GHz),
and air traffic control among others, regulatory agencies limit the frequencies for
industrial, scientific, and medical (ISM) use of microwaves to five specific regions:
25.125, 5.80, 2.45, 0.915, and 0.4339 GHz. Household domestic microwave ovens
operate only at the 2.45 GHz frequency (12.25 cm wavelength).
Radiation
Radio
Infrared
Microwave
Long-wave Short-wave
(AM)
(FM)
TV
Visible
UV
X-rays and
Gamma Rays
Radar
^^H
Approximate
Scale
buildings
Wavelength
Frequency (Hz)
10 m
105
unicellular
organisms
human palm
(2.45 GHz, 12.25 cm)
trees
I
r
i
atoms
iym
100 cm 12.25 cm
3X108M5X109
molecules
3x10"
i
10">
i
Bond
breaking
Chemical
Implications
#
Molecular Rotations
Molecular
Vibrations
Valence shell
electron excitation
(e.g. 7t*7i*)
Molecular ionization
Figure 1-1. Regions of the electromagnetic spectrum with approximate scale as
well as chemical implications for selected wavelength regions.
The 2.45 GHz frequency largely has been adopted by the scientific microwave
community, and examples utilizing microwave irradiation other than 2.45 GHz to heat
reactions are few.1 For the most part, previous research utilizing frequencies besides
1
(a) Gedye, R. N.; Wei, J. B. Can. J. Chem. 1998, 76, 525-532. (b) Moller, M.; Linn, H. Key Eng. Mater.
2004, 264-268, 735-739. (c) Takizawa, H.; Uheda, K.; Endo, T. J. Am. Ceram. Soc. 2000, 83, 2321-2323.
8
2.45 GHz has been directed toward examining those that may be more appropriate and
energy efficient especially with regard to solvent absorptivity. For instance, Serpone and
co-workers recently reported on a 5.8 GHz reactor that does a better job heating nonpolar solvents such as pentane or benzene than traditional 2.45 GHz frequency
microwaves.2
Microwave heating is based upon the ability of a particular substance such as a
solvent or substrate to absorb microwave energy and efficiently convert that energy into
heat (kinetic energy). It is largely due to the electric component of the electromagnetic
field of a microwave that leads to microwave heating. Molecules with a dipole moment
(permanent or induced) attempt to align themselves with the oscillating electric field of
the microwave, leading to rotations. In the gas phase, these molecular rotations are
energetically discrete events and can be observed using microwave spectroscopy.
However, in the liquid and solid phases, these once-quantized rotational events
coalesce into a broad continuum as rotations are rapidly quenched by collisions, leading
to translational movement.
Molecules in the liquid or gas phase begin to be sympathetic to incident
electromagnetic (EM) irradiation leading to rotations when the frequency approaches
~1Q6 Hz.3 The upper limit for EM-induced rotations is about 1012 Hz as molecules cannot
rotate an appreciable amount before the field changes direction at frequencies above
this. As rotation arc decreases with increasing frequency, so too does the molecule's
ability to turn incident irradiation into kinetic energy. The optimum frequency in which a
molecule turns the incident irradiation into rotations is a function of many things including
the molecule's permanent dipole moment, the size of the molecule, and temperature, but
(d) Malinger, A. K.; Ding, Y.-S.; Sithambaram, S.; Espinal, L; Gomez, S.; Suib, S. L. J. Catal. 2006, 239,
290-298.
2
Horikoshi, S.; lida, S.; Kajitani, M.; Sato, S.; Serpone, N. Org. Proc. Res. Dev. 2008, 12, 257-263.
3
For an excellent discussion on dielectric heating fundamentals, see: Gabriel, C; Gabriel, S.; Grant, E. H.;
Halstead, B. S. J.; Mingos, D. M. P. Chem. Soc. Rev. 1998, 27, 213-223.
9
for most small molecules the relaxation process is most efficient in the microwave region
(.3-300 GHz) of the electromagnetic spectrum.
An analogy can be made to American baseball (Figure 1-2). During the swing,
the batter can be said to be "rotationally excited," and can deliver some amount of
rotational force to the incoming pitch. At the point of impact the batter's rotational energy
is rapidly converted into translational energy of the ball. Similarly, one water molecule
excited rotationally by incident microwave irradiation can strike a second molecule of
water, converting rotational energy into translational energy. Under microwave heating, a
large number of molecules are being rotationally excited by the incident microwave
irradiation, striking other molecules and converting rotational energy into translational
energy, i.e. kinetic energy, and hence heat.
Figure 1-2. On the left, a molecule (a) that has been rotationally excited by
microwave irradiation is approached by a second molecule (b), panels 1-3. Upon
impact (panel 3), the rotational energy of (a) is converted to translational
movement of (b). In panel 4, notice the increase in translational vector
magnitude, the consequence of which leads in an increase in molecular collisions
(kinetic energy). This concept is not so unlike that of a University of Connecticut
baseball batter about to strike a baseball and impart his rotational energy onto a
baseball in the form of translational energy. Photo courtesy of UConn
Athletics/Stephen Slade Photography.
10
Because microwave heating is firstly dependent upon a molecule's dipole
moment, it stands to reason that more polar solvents such as ethanol, DMF, DMSO, or
water better convert microwave irradiation into heat than non-polar ones such as toluene
or hexane. Previous efforts have been undertaken to quantify this general trend and
relate relative microwave absorptivities4 to a substance's dielectric constant (e'),
dielectric loss (e"), or a combination of both, termed loss tangent or loss angle (tan 3 =
£"/£').
The dielectric constant describes the polarizability of a molecule in the
microwave field while the dielectric loss describes the efficiency with which a molecule
converts the incident electromagnetic irradiation into molecular rotation, and hence heat.
The loss angle (tan 3) is a measure of reactance (resistance in a capacitor)5 of a
molecule. The easiest way to understand this concept is to examine the extremes.
Something that has a tan 3 = 0 means that it is completely transparent to microwave
irradiation. For tan 3 = 0; 3 = 0, incident irradiation passes through the material with its
path unchanged. For a perfectly absorbing material, tan 3 = °°; 3 = TT/2 radians or 90
degrees. Here the material under irradiation shows complete resistance to the incident
irradiation and would be classified as a perfect absorber. Practically speaking, materials
with a tan 3 approaching 1.0 are very strong microwave absorbers, e.g. ethanol (tan 3 =
.941) or ethylene glycol (tan 3 = 1.350) are both exceptional absorbers of microwave
irradiation at 2.45 GHz. Table 1-1 lists the constants for dielectric constant, dielectric
loss and loss tangents for a range of solvents.
While the dielectric loss or tan 3 can be used to assess a substance's microwave
absorbance, the use of any single parameter drastically oversimplifies the issue of
"efficient" microwave heating. A number of other factors contribute to the overall heating
4
For Reviews, see: (a) Ref. 3 (b) Hayes, B. L. Microwave Synthesis: Chemistry at the Speed of Light, CEM
Publishing, Matthews, NC.
5
For an excellent discussion of microwave absorptivity and theory from first principles, see: Craig, D. Q. M.
Dielectric Analysis of Pharmaceutical Systems, Taylor and Francis: Bristol, PA, 1995.
11
efficiency by microwave irradiation. Attributes such as a solvent's specific heat capacity,
heat of vaporization, or the depth of field for the substance being irradiated can
sometimes have a larger impact upon heating rate than its respective dielectric loss or
loss tangent. Furthermore, dielectric loss and dielectric constant are functions of both
irradiation wavelength as well as temperature; specific heat changes as a function of
temperature; and heat of vaporization changes as a function of pressure. Room
temperature water, for instance, is most microwave absorbent at approximately 18 GHz
(Figure 1-2), but as temperature increases, so does the optimum frequency at which
water converts microwave irradiation to heat. Generally, however, when synthetic
microwave chemists speak of "good" or "bad" microwave absorbers, implied is a 2.45
GHz irradiation source, a small depth-of-field (1-10 cm) and synthetically relevant
temperatures (50-150 °C).
Solvent
Water
Ethanol
DMSO
DMF
Acetonitrile
Acetone
DCM
THF
Ethyl Acetate
Toluene
Hexane
Dielectric
Constant
(*')
80.4
24.3
45
37.7
37.5
20.7
9.1
7.4
6
2.4
1.9
(£")
Loss
Tangent
(tan 3)
9.89
22.9
37.1
6.07
2.32
1.11
0.382
0.348
0.354
0.096
0.038
0.123
0.941
0.825
0.161
0.062
0.054
0.042
0.047
0.059
0.040
0.020
Dielectric
Loss
Table 1-1. Dielectric constant (£'), dielectric loss (e"), and loss tangent (tan d) for
selected solvents at 2.45 GHz. Data from: Hayes, B. L. Microwave Synthesis:
Chemistry at the Speed of Light, CEM Publishing, Matthews, NC.
200
180 i
160
140
120
100
80
60 H
40
20
0
0.01
0.1
1
10
100
Frequency (GHz)
• dielectric constant (z')
•••• dielectric loss (E")
tan 8 (x 100)
Figure 1-2. Dielectric constant (E'), dielectric loss (E"), and loss angle (tan 8) all
are functions of irradiation frequency. Shown here are the plots for water at 25
°C, which heats most efficiently at approximately 18 GHz. Plot generated from
data from references 3 and 5. Tan d values are scaled (x 100) for clarity.
1.2 "Microwave Effects." "Microwave heating can enhance the rate of reactions
and in many cases improve product yields." This rhetoric typifies that found strewn
throughout literature extolling the virtues of utilizing microwave irradiation to "promote"
reactions. While that sentence is technically not false, it is every bit as true if one was to
remove the word, "microwave" altogether, left only with "Heating can enhance the rate of
reactions."
That said, microwave heating can be different than conventional, convectionbased heating. Numerous attempts have been made to evaluate differences in
microwave versus conventional heating, either real or perceived. These differences have
been divided into two categories: "specific" and "non-thermal" microwave effects. A
comprehensive discussion on these topics can be found in Chapter 3.
1.3 Microwave Assisted Organic Synthesis. With or without all the various
speculation into "microwave effects," there are a number of excellent reasons to have at
your disposal a scientific microwave apparatus. The use of microwave irradiation to heat
13
reactions has been most widely adopted by organic chemists and a number of books
and reviews have been published on this subject.6 It is certainly a useful tool that exhibits
a range of applications from the relatively routine lab work7 to affording the organic
bench chemist an opportunity to carry out exciting new chemistry.
JOC
OL
TET
TL
OPRD
Total
MW
%MW
2002
36/1465
28/1213
39/1334
61/2503
7/197
2003
52/1587
43/1305
47/1366
91/2396
7/198
2004
70/1473
56/1388
62/1480
132/2385
9/201
2005
98/1633
70/1502
105/1480
188/2207
12
2006
108/1510
66/1565
126/1522
226/2184
14/204
2007
134/1552
83/1438
167/1569
230/2175
19/211
2008
118/1524
86/1426
179/1525
213/1981
19/202
2009
146/1508
101/1470
173/1444
252/2057
12/239
171
240
329
473
540
633
615
684
2.55
3.50
4.77
6.75
7.73
9.11
9.24
10.10
Table 1-2. Percentage of published journal articles for 5 major organic chemistry
publications utilizing microwave irradiation. Article hits for keyword search
"microwave" in all fields/total articles published.
"]
•
12 1
E10
Q>
t
S e
2002
2003
2004
2005 2006 2007 2008 2009
Year
• Total •JOC " O L «TET ATL "OPRD
Figure 1-3. Plot of the relative percentage of publications utilizing microwave
heating in five leading journals dedicated to organic chemistry.
A number of relevant books reviewing microwave assisted organic synthesis have been published,
including: (a) Loupy, A. Ed., Microwaves in Organic Synthesis, 2nd Edition, Wiley-VCH, Weinheim, 2006. (b)
Kappe, C. O.; Stadler, A. Microwaves in Organic and Medicinal Chemistry, Wiley-VCH, Weinheim, 2005. (c)
Lidstrom, P.; Tierney, J. P. Eds., Microwave-Assisted Organic Synthesis, Blackwell, Oxford, 2005. (d)
Hayes, B. L. Microwave Synthesis: Chemistry at the Speed of Light, CEM Publishing, Matthews, NC
7
Performing a simple, fractional, or vacuum distillation from a scientific microwave allows an exquisite level
of control not easy to duplicate in an oil or sand bath. After all, reflux at 35 watts is significantly different than
reflux at 85 watts.
14
The seminal publications of Gedye 8 and Majetich9 in 1986 paved the way for the
use of microwaves in organic synthesis. The widespread application of microwave
irradiation to heat organic reactions can be appreciated by the increase in total number
of publications as well as the increased percentage of publications that utilize microwave
irradiation in five major organic chemistry journals from 2002-2009, illustrated in Table 12 and Figure 1-3.
Certainly, the most useful attribute to the scientific microwave is its ability to
facilitate the chemist when developing new synthetic methodology.10,11 Due to the sealed
vessel conditions, the microwave allows the scientist access to a range of conditions that
are otherwise difficult (though not impossible) to attain in other fashions. For example,
an organic chemist generally will select solvents in accordance with boiling point and
known or assumed activation energy barriers. A stubborn reaction may be carried out in
refluxing xylenes (bp 137-140 °C), 1,2-dichlorobenzene (bp 178-180 °C), or possibly Nmethyl pyrrolidinone (NMP, bp 202 °C). However, under sealed vessel conditions, nearly
any solvent the bench chemist selects becomes a viable option, regardless of desired
reaction temperature. Even dichloromethane (bp 40 °C @ 1 atm) can be heated to 160
°C within the typical pressure limitations of most commercially-available scientific
microwave reactors.12
Perhaps the most interesting and underutilized solvent in organic chemistry is
water. While there are certainly a number of reactions that generally do not tolerate the
presence of water, e.g. Grignard or alkyl lithium reactions, there are plenty of reactions
that not only tolerate the presence of water, but in some cases benefit from the addition
8
Gedye, R., Smith, K., Westaway, H. Tetrahedron Lett. 1986, 27, 279-282.
Giguere, R. J.; Bray, T. L; Duncan, S. M.; Majetich, G. Tetrahedron Lett. 1986, 27, 4945-4948.
10
For a recent review highlighting microwaves in organic synthesis from 2004-2008, see: Kappe, C. O.;
Dallinger, D. Mol. Divers. 2009, 13, 71-193.
11
Kappe, C. O. Chem. Soc. Rev. 2008, 37, 1127-1139
12
(a) Goodman, J. M.; Kirby, P. D.; Haustedt, L. O. Tetrahedron Lett. 2000, 41, 9879. (b) The embedded
applet described in (a) can be found at: http://www-jmg.ch.cam.ac.uk/tools/magnus/boil.html
9
15
of water to the solvent system, or even as the lone solvent. Furthermore, water is
especially suitable for high temperature organic reactions, and thus is great to pair with
microwave irradiation.13 The dielectric constant of water changes as a function of
temperature and while we characterize water as a very polar solvent at room
temperature, at elevated temperatures it becomes quite different. For example, water at
150 °C has a dielectric constant similar to DMSO at room temperature, at 175 °C the
dielectric constant becomes similar to DMF, at 200 °C it is similar to acetonitrile at 25 °C,
and water heated to 300 °C has a dielectric constant on par with room temperature
acetone.14 This Jekyll-and-Hyde attribute is quite useful and certainly can be taken
advantage of: water is able to solvate reagents at high temperatures and then upon
cooling the products become very insoluble and facilitate isolation of the newlysynthesized compounds.
100 -I
o
60
|
«»
50
40 H
30
i
20 i
V ^
10
0
100
200
Temperature (°C)
300
Figure 1-5. Plot of the dielectric constant of water as a function of temperature
illustrating how water becomes less polar with heating. Points generated from
data obtained from CRC Handbook of Chemistry and Physics.
Freedom in solvent selection not only allows the bench chemist a greater
flexibility in new methodology development and a reduction in work up time, but also
affords the potential for "greener" chemistry to be developed. Certainly, ethanol or ethyl
13
For recent reviews of water in microwave-assisted synthesis, see: (a) Strauss, C. A. Aust. J. Chem. 2009,
62, 3-15. (b) Polshettiwar, V.; Varma, R. S. Chem. Soc. Rev. 2008, 37, 1546-1557. (c) Dallinger, D.; Kappe,
C. O. Chem. Rev. 2007, 107, 2563-2591.
14
Values for organic solvents obtained from: Anslyn, E. V.; Dougherty, D. A. Modern Physical Organic
Chemistry, University Science Books, Sausalito, CA, 2006.
16
acetate could be considered "green" solvent choices as both can be derived from
biological sources.15 Additionally, solvents like ethanol and ethyl acetate represent less
toxic alternative and generally require less energy to remove at the end of a synthesis
due to their moderate boiling points. Indeed, both are found on Pfizer's "Green Solvent
List,"16 an in-house solvent selection guide that acts as a reminder to practicing chemists
to select less problematic solvents whenever feasible. Finally, the most celebrated
"green" solvent must surely be water. It is easy to extract from, inexpensive, non-toxic,
non-flammable, and widely available.
However, the true "greenness" of water is very often overstated. Intuitively,
something so ubiquitous as water and indeed so essential to life should automatically
qualify it as "green," but overlooked is the fact that it cannot be incinerated after use and
it takes a considerable amount of energy to distill water in order to purify it. Water
purification at treatment plants, too, is a costly and energy-intensive endeavor. Thus, the
pros of the use of water as a solvent are balanced by these cons and water is likely no
more or less green than solvents such as ethanol, ethyl acetate, or methyl ethyl ketone.
Indeed, when evaluated using a full complement of the most essential metrics, it has
been reported that, "water is only a truly green solvent if it can be directly discharged to
a biological effluent treatment plant."17 Obviously, dissolved heavy metal catalysts, ionic
liquids and other phase transfer reagents, trace amounts of newly-synthesized organic
compounds whose human or aquatic toxicology is likely unknown would render water
unfit to this type of disposal. This said, water still represents an attractive solvent and
15
That said, bio-based ethanol and ethyl acetate represent a tiny fraction of the volume used. The majority
of ethanol utilized in the lab has been synthesized from the hydrolysis of ethylene, which has been distilled
from crude oil. Similarly with ethyl acetate: the acetic acid portion was likely produced by the Monsanto
(Rhodium) or Cativa (Irridium) acetic acid processes whose feedstock begins with methanol, again
originating from fossil fuels.
16
Cue, B. W., oral presentation, 2009 American Chemical Society School for Green Chemistry and
Sustainable Energy, Golden, Colorado, July 24, 2009.
17
(a) Blackmond, D. G.; Armstrong, A.; Coombe, V.; Wells, A. Angew. Chem. Int. Ed. 2007, 46, 3798. (b)
see also, the Water Framework Directive, issued by the European Commission, http://www.euwfd.com.
17
likely a greener choice than most if appropriate pre-disposal treatments are employed.
Furthermore, the
ready
access
to
high temperatures
by
commercial
scientific
microwaves and the relatively efficient manner in which microwave irradiation heats
water, makes microwave assisted organic synthesis in water an attractive option in new
methodology development.
1.4 Small Scale Equipment. The first purpose-built, monomode 18 microwave
reactor was introduced in the early 1990's by Prolabo, a French company. Today, the
two largest manufacturers of small-scale, scientific microwave apparatus are Biotage
and CEM. These two manufacturers offer similar small-scale equipment. The reactors
are capable of heating reactions to 300 °C, reaching internal pressures of 300 psi and
utilize 300-watt (CEM) or 400-watt (Biotage) magnetrons. Temperature is generally
monitored via an infrared detector located below (CEM) or along side (Biotage) the
reaction vessel. Alternatively, CEM offers a fiber optic probe that can be immersed in the
reaction. The waveguide design and intellectual property of the two units is quite
different, though both are equally effective heating reactions.
Figure 1-6. Small-scale dedicated microwave units for scientific applications. A)
Biotage Initiator, B) CEM Discover, and C) Looking in to the CEM Discover
microwave cavity. C1) An infrared temperature detector is located beneath the
protective spill cup. C2) The small opening allows for the introduction of
compressed air at the end of the reaction time to rapidly cool the contents.
18
Monomode units utilize a waveguide that focuses the microwave irradiation into a small, dense volume.
Multi-mode reactors are similar to household units where the microwave irradiation is bounced from wall to
wall until it is absorbed.
18
CEM offers three vessel sizes (Figure 1-7, left) that can be used in conjunction
with their monomode microwave line under sealed-vessel conditions: a 10-ml tube with a
practical working volume of 2-3 ml, a 35-ml tube with a practical working volume of 1012 ml, and an 80-ml vessel that has a maximum practical working volume of 30 ml.
Biotage shares the 10-ml reaction tube and also utilizes a 30-ml tube with a practical
working volume of approximately 10 ml. In addition, Biotage offers smaller vessels that
can be used with working volumes of 0.5-2.0 ml or 0.2-0.5 ml (Figure 1-7, right). The
CEM microwave versions offer the flexibility to run under atmospheric conditions utilizing
standard 50 or 100 ml round bottom glassware.
Figure 1-7. Left panel: Discover tubes allow user to operate on three scales, left
to right: 80-ml, 35-ml, and 10-ml tubes. Right panel: Biotage reactor accepts four
reaction vessels, left to right: 20-ml, 10-ml, 5-ml, and 1-ml. US quarter dollar circa
1990 inserted to illustrate scale.
Pressure, temperature, applied microwave power, and stirring are monitored in
real time. Additionally, the
software
allows for on-the-fly
changes to
reaction
temperature, applied power, or pressure settings. Generally, only enough power should
be applied to afford a reasonable ramp time (1-5 °C/second). Too much power in the
ramp leads to inaccuracies, as the acquisition hardware and software are unable to keep
pace with the reaction dynamics. This situation often leads to temperature overshoot,
sometimes by 10-50 °C or more. Obviously, this is not desirable, as actual reaction
temperatures become nebulous leading to irreproducible science. It is generally best to
19
provide too little power initially and adjust the applied power, if necessary. Upon reaction
completion, both units utilized pressurized air to quickly cool the reaction back to
ambient conditions.
200
100
300
400
500
Time (s)
Temp(C)
Pressure(PSI)
Power(W)
Figure 1-8. Typical plot of microwave heating run on the 1 millimole/2-ml scale.
Power is modulated by the microwave to maintain the desired reaction
temperature of 180 °C.
Both the CEM and Biotage microwave apparatus can run as stand-alone units or
be interfaced to a PC. Use of a PC allows all pressure, temperature, and reaction time
data to be recorded, generating a data point per second (Figure 1-8). Additionally, CEM
has developed extensive peripherals to accompany the small-scale microwave unit.
These include an automated reaction-handling interface; a gas-loading kit that allows for
the introduction of a reactive gas such as hydrogen or carbon monoxide; and a
simultaneous jacketed cooling device that pumps a microwave-transparent cryogenic
fluid around a jacketed reaction tube to cool a reaction as it is being irradiated by the
microwave.
1.5 Conclusions & Outlook. Since 1986, the use of scientific microwave
apparatus has seen dramatic growth within the organic laboratory. The use of
microwave
equipment
commonplace.
in
both
Furthermore,
academia
continued
and
industry
refinements
to
continues
the
to
equipment
be
by
more
the
20
manufacturers have made the scientific microwave a safe, reliable, automated, and
incredibly convenient tool for the organic chemist.
The remainder of this thesis will focus mainly on three applications of the
scientific microwave. The first two applications, reaction monitoring (Chapters 2 & 3) and
scale up (Chapter 4), are burgeoning areas of microwave assisted chemistry and
significant advancements have been made in the past five years in these areas.
Chapters 5 & Appendices 3, 4, and 5 focus on some new methodology development
made possible in large part due to the facile access to high temperatures afforded by the
scientific microwave apparatus.
21
Chapter 2«
Raman Spectroscopy Monitoring of
Microwave Assisted Reactions
22
2.1 Introduction. The field of microwave assisted organic synthesis has
progressed tremendously over the past two decades. Reaction temperatures and
pressures can be accurately
monitored on a second-by-second
basis, applied
microwave power can be modulated by PC-interfaced software with a precision of +/- 0.1
watts, and most importantly, dedicated scientific microwave apparatus are built with the
scientist's safety in mind. In the event of an "unanticipated pressure release," the unit is
designed to automatically cease irradiation and to completely contain the reaction
contents. An unavoidable by-product to this peace of mind, however, is the organic
chemist's inability to monitor the progress of a reaction visually. Is the reaction stirring
adequately? Has a precipitate formed? Has there been a color change, etc? When
optimizing new protocol or monitoring the progress of reactions, the scientist generally is
required to stop it, allow the reaction mixture to cool, and then use standard analysis
techniques such as NMR spectroscopy or TLC. As a result, optimization of reaction
conditions such as time and temperature is often a matter of trial and error. For this
reason, recent investigations by the Leadbeater research group have investigated the
feasibility of utilizing either a digital camera19 or a Raman spectrometer20 to monitor
reactions while under microwave irradiation. The Raman spectroscopy work has built
upon the seminal publications of Myrick21 and co-workers and Pivonka and Empfield 22
who first utilized Raman spectroscopy to monitor microwave reactions. Specifically, this
work focused on utilizing Raman spectroscopy to monitor the progress of reaction both
qualitatively and later, quantitatively.
2.2 Raman Spectroscopy Fundamentals. Raman spectroscopy is a secondorder spectroscopic technique, requiring two discrete events in order to acquire Raman
19
(a) Leadbeater, N. E.; Shoemaker, K. M. Organometallics 2008, 27, 1254-1258. (b) Bowman, M. D.;
Leadbeater, N. E.; Barnard, T. M. Tetrahedron Lett. 2008, 49, 195-198.
20
(a) Leadbeater, N. E.; Smith, R. J. Org. Lett. 2006, 8, 4589-4591. (b) Barnard, T. M.; Leadbeater, N. E.
Chem. Commun. 2006, 3615-3616.
21
Stellman, C. M.; Aust, J. F.; Myrick, M. L. Appl. Spectrosc. 1995, 3, 392.
22
Pivonka, D. E.; Empfield, J. R. Appl. Spectrosc. 2004, 58, 4 1 .
23
data and generate a spectrum. 23 As illustrated in Figure 2-1, the first event is a
polarization of a molecule that is induced by the electric component of the incoming light.
This polarization has been termed a "virtual state" by physical chemists as it represents
a non-quantized excitation event. The molecule is higher in energy than it exists in the
ground state, outside the rotational and vibrational manifolds, but not so high that it has
reached the lowest electronically excited state. For the organic chemist, it may be easier
to think of this event as a momentary distortion of the electron cloud around the
molecule.
First f 2
Excited I 1
Electronic |
State L °
Raman Scattering
Excitation «> K* •
_
Anti-Stokes
CHfW
Scattering Stokes Shift
s/)/ft
Virtual
States
1
1
'
.. ...
Ground Vibrational State
E=hv
1
AE,
E=hv-AE,
E=/?v+4Ef
Figure 2-1. After excitation to a "virtual state," a molecule may encounter a
photon, leading to three possible phenomena: (a) Rayleigh scattering occurs in
an elastic collision between the incident photon and molecule and the photon
returns to the detector energetically unaltered and accounts for -99.9% of all
collisions with excited molecules, (b) A Stokes shift is detected when the incident
photon imparts some energy upon the excited molecule and returns to the
detector with a longer wavelength, and (c) The anti-Stokes shift is observed in
the very rare occurrence where a photon encounters a molecule in the virtual
state that is also vibrational^ excited and the photon absorbs some energy from
the molecule, returning to the detector with a shorter wavelength.
For a comprehensive overview of Raman spectroscopy fundamentals, theory, and applications, see: (a)
McCreery, R. L. Chemical Analysis, Vol 157, Ed. J. D. Winefordner, John Wiley and Sons, New York, 2000.
(b) Handbook of Raman Spectroscopy, From the Research Laboratory to the Process Line I. R. Lewis and
H. G. M. Edwards, Marcel Dekker, New York, 2001. (c) Applications of Vibrational Spectroscopy in
Pharmaceutical Research and Development Eds. Pivonka, D. E.; Chalmers, J. M.; Griffiths, P. R. John Wiley
and Sons, New York, 2007. (d) Characteristic Raman Frequencies of Organic Compounds Eds. Dollish, F.
R.; Fateley, W. G.; Bentley, F. F, John Wiley and Sons, New York, 1974.
24
In the second event, an incident photon may strike a molecule that is in the
virtual state. If the photon interacts with the molecule in an elastic manner, the incident
photon simply strikes the molecule and returns to the detector energetically unaltered.
This is termed Rayleigh scattering and accounts for greater than 99.9% of all
interactions between the incident photons and molecules in the virtual state. Because
Rayleigh scattering does not change the light, no information can be gleaned.
However, if a photon happens upon a molecule that is in a virtual state and
interacts with that molecule inelastically, an information-rich Raman event occurs. There
are two possibilities. The first is that a photon will strike a molecule in the virtual state,
impart some of its energy onto that molecule activating a resonant vibrational frequency,
and thus the photon will return to the detector lower in energy and with a longer
wavelength. This is termed the Stokes shift. Conversely, a photon can strike a molecule
already vibrationally-excited, absorb some energy quenching the vibration, and thus
return to the detector higher in energy and with a shorter wavelength. This is termed the
anti-Stokes shift.
In either event, the energies that are imparted or absorbed are identical and
quantized within the vibrational manifold. This means structural information can be
gleaned, similar to infrared spectroscopy. Indeed, IR and Raman are complementary
spectroscopic techniques and both report back similar information. The carbonyl stretch
at 1705 cm"1 for acetone is evident in both Raman as well as IR spectroscopy. However,
because IR signal strength is proportional to polarity of bonds (asymmetric stretching),
the carbonyl stretch will be pronounced. With Raman spectroscopy, signal strength is
proportional to bond polarizability (symmetric stretching), thus the carbonyl stretch that is
very strong in IR is quite weak in the Raman spectrum. Instead, non-polar bonds such
as carbon-carbon double bonds that are nearly transparent to IR spectroscopy give rise
to strong Raman signals.
25
2.3 Raman Spectroscopy, Pros and Cons. Besides the complementary nature
of the two spectroscopic techniques there are a number of both advantages and
disadvantages
to
utilizing
Raman
spectroscopy
as
a
reaction-monitoring
tool.
Disadvantages to Raman spectroscopy include the relatively rare occurrence of such
events when compared to absorption spectroscopies; the competing (and overwhelming)
fluorescence events exhibited by some molecules; and the necessity that solutions be
completely homogeneous and remain so throughout the course of a reaction, as Raman
spectroscopy relies upon light scattering. Advantages to Raman spectroscopy include a
"through the glass" acquisition of data, its complementary nature to IR, and borosilicate's
relative transparency to Raman spectroscopy.
In greater detail, the first disadvantage to utilizing Raman spectroscopy is that a
Raman scattering event is rare compared to an absorption event. In accordance with
Beer's Law, A = ebc, a typical absorption experiment (1 x 10"3 M concentration, 1 cm
path length, and a molar absorptivity, E = 1000), over 90% of the incident light is
absorbed. Conversely, only approximately 1 in 1010 incident photons will undergo a
Raman scattering event in a similar system, thus making a Raman event about ten
billion times less likely that an absorption event.24 Typical methods to counteract this
problem include examining solutions that are more concentrated or employing longer
acquisition times. While these are effective for static systems, neither represents an
ideal fix for monitoring dynamic systems, e.g. for monitoring reactions or carrying out
kinetic studies. Another option is to decrease the wavelength of the incident light source
as higher energy light will lead to a larger cross-section of molecules in the virtual state
and hence an increase in Raman signal. The utility of this is extremely limited, however,
due to competing fluorescence.
See Appendix 1 for detailed discussion of Raman signal intensity with respect to a number of factors.
26
Thus, a second disadvantage to Raman spectroscopy is that if molecules under
observation exhibit electronic excitation sympathetic with the excitation source (laser),
fluorescence becomes a major limitation to the use of Raman spectroscopy as a viable
technique. Because fluorescence is orders of magnitude more intense than Raman
spectroscopy, even limited excitation of molecules to their excited electronic states
quickly renders Raman spectroscopy worthless. In order to combat this effect, lower
energy light sources can eliminate electronic transitions to excited states, and
consequently limit fluorescence, though decreasing Raman signal strength. Finding the
most suitable energy for the excitation source becomes a bit of a balancing act.
Finally, in order to utilize Raman spectroscopy in dynamic systems like
monitoring the progress of reactions, the solution must remain completely transparent
and homogeneous throughout the reaction. Any precipitate or turbidity immediately
alters the path length of the system and hence Raman signal strength. Indeed, Raman
spectroscopy is utilized to monitor nucleation events during controlled crystallization in
the pharmaceutical industry as losses in Raman signal strength can be detected long
before evidence of crystallization is apparent to the human eye.25 Finally, substances
under investigation obviously need to generate structures with polarizable, Raman-active
functional groups.
2.4 Overcoming Disadvantages of Raman Spectroscopy. Despite the number
of inherent disadvantages, Raman spectroscopy is becoming more widely used for a
number of reasons. First, three technological advances have rendered a number of the
previous limitations obsolete: namely the advances in personal computer power,
development of charge-coupled devices (CCDs), and of the development of cheap diode
lasers that operate at 785 nm. CCDs can now routinely acquire Raman data at the rate
25
Applications of Vibrational Spectroscopy in Pharmaceutical Research and Development Eds. Pivonka, D.
E.; Chalmers, J. M.; Griffiths, P. R. John Wiley and Sons, New York, 2007.
27
of 50 integrations per second spanning the region from 250-3250 cm"1 or about 15,000
data points per second and above, limited only by the purchaser's pocketbook. The
commercialization of compact disc (CD) and digital versatile disk (DVD) players also has
had a profound effect upon the utility of Raman spectroscopy. Pre-1985 spectrometers
generally employed argon or krypton ion lasers, requiring either 208 or 480 V electrical
power and significant amounts of cooling water. Today, the ubiquitous 785 nm laser can
be purchased for pennies, available in every computer with an optical drive, every CD
and DVD player, laser pointers, etc. Granted, more powerful variants that are suitable for
Raman spectroscopy begin to cost more, but as a reference, the 500 watt, 785 nm
Raman spectrometer that has been utilized for all following studies is the size of a
shoebox, utilizes a standard 110 V power source, is air cooled, and is commercially
available for under $20,000. Some practical figures of merit can be gleaned from Figure
2-4, below illustrating the practical range of this unit.
28
1-
s
o>
CM
§
**
*-
s
^
•*
CM
o
.800
CO
CO
CM
*r
•*
in
in
d
CO
1^
CM
in
TT
-ao
d
in
iri
05
CO
Tf
• « •
in
00
CM
CM
oo
CO
d
CD
CO
CD
CM
CM
00
r^
CO
iri
CM
CM
CO
d
o
o
iri
CO
in
O)
o>
CD
oS
in
o
CO
CJ5
iri
*-*
00
CM
<3-
h~
C5
CM
f
CO
t
CO
co
CM
co
fCM
^
d
•<t
CO
CO
CM
CM
CO
CM
00
o
•»
in
• *
CO
CO
CO
iri
!
m
00
in
in
in
•*
in
o
iri
rCM
Tf
O)
r-i
o
o
o>
in
^
• *
CO
•<!•
,_
CO
CO
o
05
CD
h-!
• *
CO
r^
T_
CO
•t
d
00
CO
CM
o
iri
CO
CO
CM
•y
d
CM
• *
en
in
s
CO
CD
CO
CO
TC
to
CO
in
CD
CO
CO
• *
o
r>^
• *
CM
CO
i-•t
CM
•q-
CD
C\i
00
T_
CO
00
en
CO
CO
T
d
o
CM
f
^
CO
CD
T
^ in^
05
l>-
m
iri
t^
• *
p
CD
•<fr
•<t
00
T-
•»
O
OS
O
O)
O)
rit
O
a
co
d
%n
o
in
d
CD
•<r
•»
CM
l>r-
• *
OO
<*
CM
rt
N.
CO~ CO
K
^
^
•rCO-,
^t
rt
03'
CM,
CO
•5T
•t
oo
h-
CD
h-
CO
in
o
E
•»
in
co
co
CM
in
CM
o
CO
>
O
CO
•^
CM
d
ssin<x>< rCO
CO
-*
CM
CM
w
0>
m
o
CO
CO
(tj
co
f».
co
to" i>J
CO
CO
CO
CM
03
CM,
CD
CM
CO
T
~
o
o>
V *
<o
O)
CO
t»CM
CM
oi
c»
•*
oi
to
••a-
CO
'f
00
hCM
CM
CM
CM
Tt
CO
o
CM
CM
00
CM
CM
O
r~-
CO
iri
CM
CM
CO
• *
in
05
,J
CO
CM
O
CO
CM
CM
^_ in
CO
d
CO
r>r^
•>t
in
in
d
in
CD
CM
Tf
00
CO
00
CM
CM
••a-
,-:
CM
CM
CO
CO
r-
i^CM
CM
O
CO
^.
CO
iri
CO
CO
05
CO
o
o>
05
• *
in
o
o>
CM
<o
CM
CM
CO
TT
05
CM
CO
Tf
CM
CM
CO
CM
CM
CO
CO
CM
CO
CO
CO
Tf
05
o
o
O
05
^>
•tf
CO
O
CO
l~co
• *
CM
00
CD
CM
03
CO
05
d
iri
m
d
CM
00
CD
CD
O
05
d
•<t
y-^
r^
"fr
00
CO
CM
CO
o>
o
CO
in
iri
o>
CM
CM
CO
• *
CO
05
CO
CO
CO
CD
t
CO
• *
oi
CO
o
• *
CM
o
CM
,_
iri
O
a>
CO
05
T-^
CO
CO
CM
p
CO
CO
CM
CO
CO
00
o
CO
05
1^
f-
in
in
1^
CM
a>
t^
^CM
CO
r>CM
CM
CM
d
CO
CM
00
•t
CO
CO
CM
d
iri
in
O
in
CO
CO
CM
°
0
CO
o
_J
1
in
d
in
00
•5f
CM
§
o
O
s
*T
o
.400
.040
O
O
O
O
•^-
O
O
O
O
CO
O
O
O
O
CM
A);su3)u| jeuBjs
0}
CC
O)
J£
CD
CO
CO
<D
a)
c •co-o
g c
^
TO
c
CO
II
c
HI
9
E
=
w
«
co o £
=
O a) 0)
C r - Q
O • ^ >^-
i= CD
i_
CD
<D
E <°
2 2
I o Is
o
—
-C
C
o
CO
_
CD
.£
>r O
CD
c
X)
c
o
to
CD CD
CO T -
C ^CO
CD
c
c
P.Z °>
ro
c
Or
s s §.
oo O
°> CD
O o>
co «
» I" §
CO
_
c a>
co >
E "ra
CO
0)
x
a.
<U
I!
*-
a> 0
•a
CO
i.E
o
CO CO
§18
c
""
(0
C0 ^r
oi
co _£-
rax: ^
•*-'
c
CO . =
CD
>.
_C0 " •
CO <D
CO
Q
CO
!c
CD
3 2
Q.
CM CO
o
C
OD
J CC
i- o
D)
il O
05
CM
Besides
the
recent
technological
advances
that
have
allowed
Raman
spectroscopy to crawl out from under its own limitations, there are a number of inherent
advantages to utilizing Raman spectroscopy as a reaction-monitoring tool, the first of
which is its ability to record spectra literally "through the glass." A Raman spectrum can
be recorded simply by holding the laser up to a container containing the reagent in
question. This is facilitated by the fact that borosilicate glass is largely Raman
transparent, thus any typical laboratory glassware can be utilized without significant
interference. Furthermore, because Raman spectroscopy relies on light scattering, it
allows for a non-invasive reaction monitoring set up. This is especially useful and indeed
necessary when performing reactions under sealed vessel conditions in the microwave
as these conditions obviously limit the scientist's access to the contents of the reaction.
For instance, though ReactIR™ has been used extensively for in-situ reaction
monitoring,26 this monitoring tool necessitates the user to immerse a probe into the
reaction contents. This would limit the scientist to open-vessel reaction conditions and is
not ideal to monitor microwave reactions. Finally, Raman signal strength is proportional
to substrate concentration, thus it should be possible to monitor the formation of
products, and possibly of reaction intermediates, and then correlate Raman signal
strength to substrate concentration.
2.5 Raman-Microwave Interface. The interface27 of a CEM S-class scientific
microwave28 with a Raman spectrometer from Enwave Optronics29 is straightforward
with a bit of equipment modification (Figure 2-5). All S-Class microwave units are
currently available with a small access port to the microwave cavity. Figure 2-5 illustrates
26
For selected recent examples, see: (a) Poljangek, I; Likozar, B; Krajnc, M J. Appl. Polymer. Sci. 2007,
106, 878. (b) Denmark, S. E.; Pham, S. M.; Stavenger, R. A.; Su, X. P.; Wong, K. T.; Nishigaichi, Y. J. Org.
Chem. 2006, 71, 3904. (c) Grabarnick, M.; Zamir, S. Org. Proc. Res. Dev. 2003, 7, 237. (d) Pintar, A.;
Batista, J.; Levee, J. Analyst 2002, 127, 1535.
27
See appendix 2 for a detailed, stepwise procedure to interface the Raman spectrometer to the scientific
microwave apparatus as well as detailed spectrometer specifications.
28
CEM Corp., Matthews, NC. http://cem.com.
29
Enwave Optronics, San Diego, CA, http://enwaveopt.com.
30
the interface. For the Raman spectrometer, a microwave transparent extension (quartz
light pipe, Figure 2-5, panel B) was utilized to avoid introducing metal components into
the microwave field. The light pipe allowed for a near-lossless extension of both the
excitation laser and acquisition fiber optic. This whole apparatus slides into the
microwave cavity. The light pipe should be "focused" to achieve maximum sensitivity,
which simply consists of sliding back and forth until maximum Raman signal to noise is
achieved. Generally, the best Raman signal is realized by sliding the light pipe in the
cavity until it is approximately 0.5 millimeters from the reaction flask.
Figure 2-5. A) Interface of Raman spectrometer with the CEM Discover S-Class
microwave. B) Raman probe with quartz light pipe extension that allows access
into the microwave cavity. C) Looking down into the microwave cavity. The
quartz light pipe is placed right up to the microwave reaction vessel.
The Raman software is user-friendly. For single scans, a "dark" scan is collected
in the absence of any laser excitation, at which point the laser is turned on and the
Raman signal collected. Through the software interface (Figure 2-6), the scientist can
control collection parameters such as integration time, scale of the axes, and zoom to
areas of interest. Of great benefit to the bench chemist for performing kinetic studies, is
the software's integrated "Time Charter" application (Figure 2-7), allowing for up to 250
sequential spectra to be acquired at a set rate. Furthermore, a background scan can
now be taken while under laser irradiation. This scan is subtracted from all subsequent
31
scans, and will exclude Raman signal due to solvent and starting materials. This allows
the user to monitor the formation of new peaks corresponding to product formation.
Finally, the generated data can be exported to Microsoft Excel, adding to the utilitarian
nature and allowing the scientist a straightforward way to analyze the data.
&e rateO fiapUy Co*jr« # » O w t
giWE Spar ID )
i^jioj ^^gj^j ca^^j^iajjisai
[0*»<».Y>
j
i 1
Intensity
»™
D
' ;
...^„.^K^",„,JVr.^'L^"-\..._-•-
300
\ .
„__„„_.„
400 500 GOO TOO BOO 900 1000 UOO 1200 1300 1400 1500 1600 1700 1B0D 1900 Z000 »00 2300 2300
Raman Shift (cm-1)
OpanHto:
IC iProgrsmFi(ts£m»manV35» MASiSample ip<
919*M
& S**t| ,_aW8«*pS«tyg;
8/13/2007
jf^EWamanRi
Figure 2-6. Screenshot of the EZRaman acquisition software.
mm-iv-
Teaching Setup {cm-11
prim r~ 641
Load
Data
I
|
Peak |
Seat* j
Save 1
Data |
Save 1
Graj*. I
Read Spectra j
PItB3P
THIS Trend j
piiur-
Reset Ports J
lli
I
HI
/i a l l . 1
J j f t a j n f i W ' V * MGOftftaO "2030 2200 2400
if |jR«ei^»SMt(cm-1|V X
SeHm'|$5J[428
tup(o
1
Apt*'
|
Rata
j
CortnuoutScan
.
Slop Scan
Scan Pafoewieit
RetetPoW: I
Bei* |
1—
j
SngfeScwi
PTjttsp
.?
Da* Scan
DrcgiapUtt j
PUttf"
Save Parts
Loac PTsj
Ap* J
IMT
AVE
Cycle Tmeliec)
j .<
Appfc.|
Mo. Spec Mo
[ 250
59315J
• Sndte
T Slocked r Overlaid
sp^n j
.** Tme d Spat Ho
Spec Patasetot
LdCyefcM I
MiftPon! I tend
Slat*No ( " 5
Sooc Count | 75
c
3000
• /-.; . .
2000
S p e e t n i Pic* Mo
1000
10
1000
20
3D
40
50
60
70
80
200
300
Spectmn Nurtei
«X
5X
ftekttl
j
PeW2
!
P*kW
|
Pc*84 (
Specinf Nunfcfi
Saw!
SM|
Pom!(1| |45.218699t
* » I
W
Show Pelt: I
*$•/.
ColsdngSffia)
345AM
10717/2007 SpscbutnS 75
Figure 2-7. Screenshot of integrated "Time Charter" software while monitoring a
reaction. Note the bottom left window in panel records peak intensities of up to
four user-selected wavenumbers and displays the cumulative data.
32
2.6
Qualitative
Reaction
Monitoring.
Initial application
of the
Raman-
microwave interface monitored the progress of reactions qualitatively in order to quickly
determine optimal reaction conditions such as time, temperature, solvent selection, and
catalyst loading. Furthermore, by utilizing the real-time capabilities embedded in the
software, facile determination of reaction endpoints was possible.
The first reaction examined was the formation of coumarin derivatives (Figure 28). This reaction was selected as it fulfilled a number of the requirements essential for
Raman monitoring. First, the coumarin product exhibits strong Raman-active stretching
modes at 1608 cm"1 and 1563 cm"1 where the salicylaldehyde and ethyl acetoacetate
starting material exhibit minimal Raman activity in this area (Figure 2-9). Additionally, the
reaction was homogeneous throughout and neither the products nor the reactants
showed competitive fluorescence at 785 nm excitation.
CC°* J^K -^=- QQ^
Figure 2-8. Condensation between salicylaldehyde and a (3-keto ester to yield
coumarins. The reaction can be run in ethanol or ethyl acetate and is generally
catalyzed by a catalytic amount of an amine base, here piperidine.
In the formation of 3-substituted coumarin derivatives, the rate at which this
signal grew in was monitored as a function of reaction temperature and substitution of
the B-keto ester or diethylmalonate (Figure 2-10). One can appreciate the utility of the
Raman spectrometer as a qualitative reaction-monitoring tool from the following
example: Optimal conditions were developed for the formation of 3-acetylcoumarin (ethyl
acetoacetate plus salicylaldehyde) which proved to be about 8 minutes at 130 °C. Not
surprisingly, replacing the ethyl acetoacetate with diethyl malonate, the rate is slowed
somewhat as the activated methylene now is less acidic. However, the decrease in
33
reaction rate was identified during the course of the reaction (Figure 2-10, C), allowing
for an on-the-fly extension of reaction time for the formation of the ester variant 2.3.
50 -i
1050
1250
1450
1650
1850
2050
2250
Wavenumber (cm-1)
—1.0 M salicylaldehyde —1.0 M ethylacetoacetate —1.0 M product
Figure 2-9. Raman spectrum of 1.0 M solutions of 3-acetylcoumarin (red), ethyl
acetoacetate (green), and salicylaldehyde (blue) in ethyl acetate. The strong
signals at 1608 and 1563 cm'1 could be monitored in the formation of 3acetylcoumarin. The inset coumarin molecule illustrates the complex stretching
mode that leads to the calculated Raman signal at 1602 cm"1 (actual: 1608 cm" )
and was calculated using Gaussian 03 at the B3LYP-6/31G(d) level of theory.
2.1 R=CH3
2.2 R=Ph
2.3 R=OEt
conditions
EtO
OH
^ L '
J^
1565
1585
1605
Wavenumber (cm-1)
200
300
400
Time (seconds)
R = CH 3 (ethyl acetoacetate)
R = Ph (ethyl beruoylacetate)
R = OEt (diethylmalonate)
200
300
400
Time (seconds)
A
B
C
Figure 2-10. Qualitative monitoring of substituted coumarin derivatives. (A)
Peaks due to the formation of 3-acetylcoumarin over the first 20 scans. (B)
Impact of temperature upon reaction time. (C) Impact of substitution of -R upon
reaction rates. For panels B and C, note the inflection in peak intensity at the
beginning of the reaction as the microwave ramps to desired temperature.
Conditions: 0.5-1.0 M reagents in ethanol or ethyl acetate, 4-16 mol % loading of
piperidine catalyst.
34
2.7 Quantitative Reaction Monitoring: Coumarin. For the organic chemist,
much information can be obtained by investigating a reaction mechanism in great detail.
Reaction progress can be monitored quantitatively via a number of avenues, including
automated high-performance liquid chromatography (HPLC), infrared (IR, ReactIR)
spectroscopy, or heat flow calorimetry. While strides have been made in utilizing these
techniques in situ, 30 none are ideally suited to interface with a scientific microwave
reactor to generate real-time data. Furthermore, there are no reports of analogous
studies utilizing microwave-interfaced Raman spectroscopy as a quantitative reactionmonitoring tool. 31 As such, the feasibility of using in situ Raman spectroscopy to monitor
reactions quantitatively while under microwave irradiation was assessed, first using a
model reaction to determine kinetic parameters and later to corroborate this technique
as a robust and efficient method of generating kinetic data.
For quantitative studies, an open-vessel approach had to be utilized (Figure 211). This allowed for the last-second addition of the reaction catalyst, and specified a
quantitative start time. While this open-vessel technique precludes the ability to achieve
reaction temperatures above the boiling points of solvents utilized, data could be
extrapolated to higher temperatures. Furthermore, when carrying out kinetic studies, it
often is beneficial to slow the reaction in order to generate more robust and reliable data.
Typically, the reaction contents were brought to the desired temperature and a
background Raman scan would be taken. This scan would be subtracted from all
subsequent
scans. The
continuous
scan function was
used and the
Raman
30
For selected recent examples in organic chemistry employing in situ IR, see: (a) Poljangek, I; Likozar, B;
Krajnc, M J. Appl. Polymer. Sci. 2007, 106, 878. (b) Denmark, S. E.; Pham, S. M.; Stavenger, R. A.; Su, X.
P.; Wong, K. T.; Nishigaichi, Y. J. Org. Chem. 2006, 71, 3904. (c) Grabarnick, M.; Zamir, S. Org. Proc. Res.
Dev. 2003, 7, 237. (d) Pintar, A.; Batista, J.; Levee, J. Analyst 2002, 127, 1535. For recent examples utilizing
heat flow calorimetry, see: (e) Blackmond, D. G. Angew. Chem. Int. Ed. 2005, 44, 4302-4320. (!) Mathew, J.
S.; Klussmann, M.; Iwamura, H.; Valera, F.; Futran, A.; Emanuelsson, E. A. C; Blackmond, D. G. J. Org.
Chem. 2006, 71, 4711-4722. For a recent example utilizing automated HPLC, see: (g) Kedia, S. B.; Mitchell,
M. B. Org. Proc. Res. Dev. 2009, 13, 420-428.
31
(a) Waal, D. D.; and Heynes, A. M. J. Solid State Chem. 1989, 80, 170 (b) Ehly, M. E.; Gemperline, P. J.;
Nordon, A.; Littlejohn, D.; Basford, J. K.; De Cecco, M. Anal. Chim. Acta 2007, 595, 80.
35
spectrometer would acquire a spectrum at the requested interval, generally every 10
seconds.32 The first spectrum was generated before the introduction of catalyst.
Immediately after the acquisition of the first spectrum, the catalyst was injected, initiating
the reaction and set the reaction starting point (t=0). Signal growing in would then be
due to the formation of product, with loss of signal due to the consumption of starting
materials.
UX-2
S3 • „ •
}> M i x
Figure 2-11. Open vessel set up which allowed for last-second injection of
catalyst and a quantitative t=0 start time.
For the initial studies, the 3-acetylcoumarin (2.1) synthesis first used to monitor
reactions qualitatively was investigated (Figure 2-12).
O
+
OH
O
O
ii
ii
XX
E t C ^ ^ ^ ^
piperidine (cat.)
^_
ethyl acetate
2.1
Figure 2-12. Piperidine-catalyzed synthesis of 3-acetylcoumarin 2.1 was reaction
investigated quantitatively using the Raman spectrometer-interfaced microwave.
Raman signal was shown to be proportional to the concentration of the 3-acetylcoumarin
across a wide range of concentrations allowing translation of Raman signal intensity into
units of concentration for a given path length (reaction flask). Solutions of known
When utilizing the Raman spectrometer in a quantitative manner, it is important to account for the lag time
between scans. When requesting a scan every 10 seconds, a scan is actually generated approximately
every 12.2 seconds. After acquisition of the data, there is a lag during the 'memory dump' phase between
the spectrometer and the PC that accounts for this additional 2.2 seconds. This is not a problem, as the PC
also records the exact date and time that each memory dump is transferred, to the nearest second, and
includes this data in the Excel spreadsheet that is generated.
36
concentrations of 3-acetylcoumarin in ethyl acetate were prepared, brought to reflux (8384°C) and the Raman spectrum collected (Figure 2-13). After subtraction of signals due
to the solvent, a plot of signal intensity at 1608 cm"1 versus concentration was made
which yielded a straight line.
20 i
in
•a
5
15
y = 47393x-158.05
R2 = 0.99858
(0
3
O
*L 10
(A
I 5
0.1
0.2
0.3
0.4
Concentration (mol/L)
Figure 2-13. Plot of Raman intensity of the peak arising at 1608 cm'1 versus
concentration of 3-acetylcoumarin in ethyl acetate yielding a straight line, y = mx
+ b; m = Raman intensityM"1 coumarin.
When monitoring reactions using Raman spectroscopy at different temperatures
it is important to take into account that fact that the Stokes shift (which is being
monitored) is inversely proportional to temperature.33 This relationship is due to the
fundamental manner in which Raman spectroscopy probes a molecule; that is it excites
it in the lowest-energy electronic state. As temperature increases, there is a smaller
population of molecules in the ground state to be excited, thus the signal intensity drops.
Over small temperature ranges, the change in the Stokes shift is fairly negligible but for
analytical accuracy, an appropriate scaling factor was developed that would allow the
comparison peak intensities at different temperatures with confidence.
This was
achieved by measuring the intensity of the 1608 cm"1 signal of known concentrations of
3-acetylcoumarin at temperatures within the range of interest. The average of ten scans
at five temperatures ranging from 35 - 75 °C and at four different concentrations gave a
1
See Appendix 1 for an in-depth discussion of Raman signal strength.
37
total of 20 data points. From this, a reasonable scaling factor was determined (Figure 214).
1.18 -i
Rel;itiv e Inten sity
1.16 1.14
1.12 1.1 1.08 1.06
1.04 1.02 1 0.98 -I
30.0
1
40.0
1
1
50.0
60.0
Temp (°C)
1
1
70.0
80.0
Figure 2-14. A scaling factor was developed to account for the slight change in
Raman signal intensity that arises as a function of temperature.
With the appropriate calibration curve and scaling factor in hand, the order of the
reaction with respect to each reagent was determined. Using the kinetic method of initial
rates (varying the concentration of one reagent at a time and measuring the rate of
reaction at t = 0) in conjunction with the isolation method (performing experiments where
the concentration of one or more reagents is kept constant to determine rate
dependence as a function of the reagent being probed), reaction orders were
determined.34 Holding the concentrations of salicylaldehyde and ethyl acetoacetate
constant at 1.00 M, an initial range of piperidine concentrations were screened, ranging
from 0.0200 M to 0.400 M (Figure 2-15). At low catalyst loadings (0.0200 - 0.0800 M)
the reaction appears to be very close to first-order with respect to the piperidine
concentration. However, the order of the reaction with respect to piperidine is rather
complex and with increasing catalyst loading, it becomes apparent that the reaction rate
is approximately proportional to the square root of the concentration of piperidine used.
For an excellent reference of kinetic studies, see: \nvestigation of Rates and Mechanisms of Reactions,
Part 1. Bernasconi, C. F., Editor. John Wiley and Sons, New York, 1986.
38
0.000
0.100
0.200
0.300
0.400
0.500
[piperidine] (mol/L)
Figure 2-15. Plot of rate of formation of 3-acetylcoumarin (2.1) as monitored by
the peak forming at 1608 cm"1 versus piperidine concentration. Increasing
concentrations illustrate an approximate 1/4-order dependence on catalyst
loading, though a first-order approximation can be made at low catalyst loadings.
The complex nature of the rate dependence upon the catalyst concentration may
not be unexpected. The pKa values in DMSO for the piperdinium cation, the phenolic
proton of the salicylaldehyde, and the acidic proton of ethyl acetoacetate are 10.9, 14.8,
and
14.1, respectively.
Thus, the
development
of
numerous
competitive
and
mechanistically non-productive equilibria would be expected, leading to a complex
reaction order with respect to the piperidine catalyst.
Determination
of reaction order with respect to ethyl acetoacetate
and
salicylaldehyde was carried out at a piperidine concentration of 0.0800 M. In the first
series
of experiments,
initial
rates were
measured
as
a function
of
various
concentrations of ethyl acetoacetate (0.0312 M - 2.00 M) holding the salicylaldehyde
concentration constant at 1.00 M and the piperidine concentration constant at 0.0800 M.
The experiments were then repeated, holding the initial concentration of ethyl
acetoacetate at a constant 1.00 M and similarly varying the concentrations of
salicylaldehyde. From Figure 2-16 the overall rate equation is shown to be:
rate = k [salicylaldehyde]1 [ethyl acetoacetate]1 f([piperidine])
where f([piperidine]) = [piperidine]1 for concentrations at or below 0.080 M, and
f([piperidine]) = [piperidine]1'2 for concentrations above 0.080 M
39
[aldehyde]
AlU'sec' 1
rate0bs (M«min' 1 )
[EAA]
AlU'sec" 1
2.00
1.00
314
0.398
2.00
304
0.385
167
0.211
1.00
167
0.211
0.50
81.7
0.103
0.50
85.1
0.108
0.25
44.4
0.056
0.25
46.1
0.058
0.125
24.3
0.031
25.1
0.032
0.063
17.1
0.022
0.12E>
0.06C!
15.2
0.019
0.031
9.41
0.012
0.031
10.8
0.014
6
U"
6
•
_
5 •
in
•o
ni
o>
3
4
•
£
3 •
5
2•
CO
CA
7
.c
*
en
g
1 - <f=*0 j
'
-1 '
•0.125 M
-i
5
H
i 3
/^^
<A
g
rate ob s (M«min 1 )
2
1 •<
C
20
40
Time (seconds)
+ 0.250 M
* 0.500 M
Variation of [salicylaldehyde]
60
*1.00M
~°!
-1
£
I
• 0.125 M
20
40
Time (seconds)
+0.250 M
A0.500 M
60
*1.00M
Variation of [ethyl acetoacetate]
Figure 2-16. The method of initial rates was used in conjunction with the isolation
method to determine the order of the reaction with respect to salicylaldehyde and
ethyl acetoacetate (EAA). Converting units of Raman intensity to concentration
units (Figure 2-10) allowed calculation of the rate in conventional terms (M»min1)
and the rate constant, k. IU=Raman intensity units.
Next, the activation energy (Ea) was determined for the reaction by running the
reaction at a range of temperatures between 25 and 80 °C (Figure 2-17). The reactions
were performed using 1.00 M concentrations of ethyl acetoacetate and salicylaldehyde
and 0.08 M piperidine and the product signal at 1609 cm"1 was monitored. Raman signal
intensity was converted to units of concentration as above. The Arrhenius plot of In kobs
versus 1/T gave a straight line, the slope of which corresponds to -Ea/R, (R=8.314 J'tmol"
1
*K"1). From this, the activation energy for the reaction was calculated to be 38.3 kJ/mol.
Similarly, the Eyring plot of In (kobsIJ) versus 1/T yielded a straight line from which the
activation enthalpy (AH*) was determined to be 35.5 kJ/mol.
40
-7.0
-1
y = -4602.1x +11.39
R2 = 0.999
E =38.3 kJ/mol
-1.5 H
-2
y = -4273x + 4.5939
R2 = 0.998
AH" = 35.5 kJ/mol
-7.5
p-8.0
•*-2.5
£-8.5
-3
-9.0
-3.5 i
-9.5
0.0028 0.0029 0.0030 0.0031 0.0032 0.0033
1/T
Temp (K)
1/T
309.65
320.3
330.25
340.8
350.15
0.003229
0.003122
0.003028
0.002934
0.002856
slope
scaled slope
(IU«sec 1 ) (Figs. 2-12 & 13)
24
27.8
40.8
45.8
68.4
63.2
96.7
100.2
132
133.5
0.0028 0.0029 0.0030 0.0031 0.0032 0.0033
1/T
k
(M"1«min"1)
0.0304
0.0517
0.0801
0.1224
0.1690
+/-
Ink
ln(k/T)
0.0018
0.0063
0.0009
0.0048
0.0050
-3.494
-2.962
-2.525
-2.100
-1.778
-9.230
-8.731
-8.325
-7.932
-7.636
Figure 2-17. Top left: plot of In k vs. 1/T yields a straight line from which the
activation energy is calculated to be 38.3 kJ/mol. Top right: plot of In (kU) vs. 1/T
yields a straight line from which the activation enthalpy (AH*) is calculated to be
35.5 kJ/mol. Bottom: table of data used to generate the plot. Slope in Raman
intensity units per second can be scaled using the temperature-scaling factor
using Figure 2-14. Units of Raman intensity are converted to units of
concentration using Figure 2-13. The reactions were carried out in quadruplicate
and the error is reported as +/- one standard deviation.
The quantitative data obtained w a s then used to extrapolate and calculate a rate
constant, kobs, for other reaction temperatures. W e wanted to compare our results here
with those w e obtained for our qualitative studies previously w h e n using the R a m a n
apparatus simply as a tool for monitoring how long the reaction took to reach completion.
Performing the reaction at 130 °C w e found that it took approximately 8 min to reach
completion. Extrapolating our quantitative rate data to 130 °C w e found that k= 0.972 M"
1
«min" 1 or 0.0162 M"1«sec"1. Using the half-life equation for a second-order reaction (ti / 2 =
1/k0bs) the half-life for the reaction at 130°C was calculated to be 62 seconds. Thus,
reaction completion ( - s e v e n half-lives) would take approximately seven minutes, fitting
well with previous qualitative results (Figure 2-10, panel B).
41
2.8
Mechanistic
Insight
via
Raman
Spectroscopy
&
Computational
Modeling. A three-dimensional heat map plot of Raman intensity versus time across the
region of 1500 cm"1 to 1680 cm"1 was generated (Figure 2-18). From this plot, it can be
seen that a peak rapidly appears at 1630 cm"1 before disappearing as the reaction
progresses (Figure 2-19). This plot led to speculation into why this peak arose, and
whether or not it could be used to investigate potential mechanistic pathways for the
formation of 3-acetylcoumarin. From this data, it was hypothesized that there were two
likely possibilities leading to this signal, either: (a) there is a rapid formation of some
Raman-active reaction intermediate that is then directly converted to 3-acetylcoumarin,
or (b) there exists some competitive equilibrium between a low-energy, non-productive
reaction intermediate that over time re-equilibrates and is consumed to yield the product.
wavenumber (cm-1)
Figure 2-18. Three-dimensional surface plot of Raman intensity versus time for
the Raman spectral region 1500-1680 cm"1. The fleeting signal at 1630 cm"1
clearly can be seen before disappearing as the reaction proceeds to completion.
This plot led to further investigations in an attempt to discern the mechanistic
pathway and propose plausible intermediates.
42
3-acetylcoumarin
0 *
500
1000
1500
2000
Time (seconds)
2500
3000
Figure 2-19. Normalized two-dimensional plot generated from the data from
Figure 2-18 for the rapid formation then disappearance of the Raman signal at
1630 cm"1 (Intermediate) and the signal arising at 1608 cm"1 (3-acetylcoumarin).
The maximum intensity for each respective signal is normalized to 1 for clarity.
The formation of 3-acetylcoumarin can nominally be described as a Knoevenagel
condensation in conjunction with a trans-esterification. Disregarding proton transfers, the
Knoevenagel condensation can be divided into two sequential steps: bimolecular35
addition (A) and then elimination of water (B) to afford the olefin. Again disregarding
proton transfer steps, the transesterification also can be split into two sequential steps:
bimolecular addition (A') followed by elimination of ethanol (B') to afford the aryl ester.
To order these possibilities, only two rules need be followed: B must follow A and B'
must follow A', i.e. A-B-A'-B', A-A'-B-B', A'-A-B-B', and A'-B'-A-B are the four
possibilities. Figure 2-20 below illustrates various intermediates that would arise in the
four plausible pathways that satisfy this description.
Computational modeling was used as a tool both to predict relative energies as
well as vibrational stretching frequencies (i.e. Raman and IR) for the 9 unique proposed
intermediates (Figure 2-20) and the data generated were then used to determine the
Technically, 'bimolecular' here can mean either true bimolecular addition or an intramolecular addition,
depending on the sequence of events.
43
most likely mechanistic pathway in the formation of 3-acetylcoumarin. The molecular
modeling program Gaussian 03 36 was utilized for all calculations. In order to maximize
the cost to benefit ratio of the calculations,37 density functional theory was utilized
applying 6-31 G(d) basis set of the Becke38 exchange functional in conjunction with the
correlation functional developed by Lee, Yang, and Parr.39
OH
,
^ B O ^ O
AG = +0.7 kJ/mol
OH ,
—
•
^%^*°
^ O H
CtM °.
Mrto^o
AG = +0.7 kJ/mol
+
o
EtO
AG = +64.8 kJ/mol
T I
Ciui
^A°HB
OH
AG = +1.3kJ/mol
AG = +32.2 kJ/mol
OH 1
-c6£
o
—AG = +35.5 kJ/mol
B
—
I
Pi
w^*u§AG = +35.5 kJ/mol
a
AG = +74.3 kJ/mol
O X )
^
- O ' ^O
AG = -52.8 kJ/mol
AG = +1.3kJ/mol
II
L
Cr
r°
^cAo
AG = +43.8 kJ/mol
b
Figure 2-20. Four plausible reaction pathways in the formation of 3acetylcoumarin and their calculated energies in vacuum relative to starting
materials.
Gaussian 03, Revision B.01, Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.
Cheeseman, J. R.; Montgomery, Jr., J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C ; Millam, J. M.; Iyengar, S
S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.
Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.
Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Bakken, V.; Adamo, C; Jaramillo, J.
Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C ; Ochterski, J. W.; Ayala
P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels
A. D.; Strain, M. C; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.
Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.
Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.
Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C ; and Pople, J. A.
Gaussian, Inc., Wallingford CT, 2004.
37
As a point of reference, the time required to calculate relative energy in vacuum and Raman stretching
frequencies for Intermediate 1b in Figure 2-19 takes 20 hours, 0 minutes, 54 seconds on an Intel Pentium 4
processor (3.2 GHz).
38
Becke, A. D. Phys. Rev. A 1988, 38, 3098.
39
Lee, C; Yang, W.; Parr, R. G. J. Phys. Chem. 1996, 100, 16502.
44
The activation energy for this reaction had been determined experimentally to be
38.3 kJ/mol. As such, mechanistic pathway 1 (Figure 2-20) likely can be ruled out.
Completion of the Knoevenagel condensation (addition and elimination) before the transesterification step would necessitate two very high-energy intermediates: 1b (AG = +64.8
kJ/mol) and 1c (AG = +74.3 kJ/mol). Based on the relative energy level of 1a, an
argument could be made that a reasonable population of this species would be apparent
before continuing through the remainder of the reaction to afford 3-acetylcoumarin.
However, the predicted Raman spectrum for intermediate 1a includes no significant
stretching events sympathetic to the energy range of 1600-1750 cm"1.
OH
,
o o
XX
9
o
cc_o^--6T-ccS-cc^
O
N
Eto
O
II
AG = +0.7 kJ/mol
^ ^ O H
AG = +64.8 kJ/mol
AG = +74.3 kJ/mol
1b
1c
1a
AG = -52.8 kJ/mol
Figure 2-21. Pathway 1 generates two relatively high-energy intermediates, 1b
(+64.8 kJ/mol with respect to starting materials) and 1c (+74.3 kJ/mol), as
determined using computational techniques. As such, this mechanistic pathway
likely can be eliminated.
Similarly, the relatively high-energy intermediates 4a (AG = +35.5 kJ/mol) and 4b
(AG = +43.8 kJ/mol) can be used to reasonably exclude reaction pathway 4 (figure 222). While these calculated energies are significantly closer to the experimentallydetermined activation energy (38.3 kJ/mol), because each of the energies calculated for
4a, 4b, and 4c is high, there is no logical reason that a significant population of any of
these intermediates would accumulate in the reaction and hence be visible by Raman
spectroscopy. For these reasons, pathway 4 was preliminarily eliminated as a possible
mechanism. Additionally, none of these intermediates was calculated to have significant
Raman active stretching modes in the energetic range in question, further indicating the
unlikely possibility that pathway 4 is valid.
45
c0
^
Q
X
Q
X
•
—
-
[i :i
,
LoS—
0'^0H
Q
r
^
,
^°
OH
^<>~°-
- ^ ^ O ^ O
^ - ^ O ^ O
AG = +35 5 kj/mol
AG = +43.8 kJ/mol
AG = +32.2 kJ/mol
4a
4b
4c
-^--O—O
AG = -52.8 kj/mol
Figure 2-22. Pathway 4 can be eliminated, as there is no logical reason that any
of these relatively high-energy intermediates would accumulate in the reaction
flask. Furthermore, none of these intermediates is predicted by computational
methods to have a significant Raman stretching event in the energy range 16001750 cm"1.
Of the four proposed possibilities, then, pathways 2 and 3 remain. These two
pathways are quite similar with the elimination of water as the final step of the reaction,
but differ in the initial bimolecular addition step. In terms of relative calculated energies
for the intermediates, both pathways are plausible. Furthermore, in both mechanistic
pathways, it can be reasoned that there would be some build-up of a reaction
intermediate before continuing on to the 3-acetylcoumarin product. In pathway 2, both
intermediates 2a and 2b are calculated to be relatively low-energy intermediates, and a
rapid accumulation of intermediate 2b (AG = +1.3 kJ/mol vs. starting materials) would be
expected. In pathway 3, the accumulation of intermediate 3b (AG = +1.3 kJ/mol vs.
starting materials) could be postulated as it resides in a local minimum energetically.
Both pathways 2 and 3 are supported by the Raman-active
stretching
frequencies calculated computationally. For intermediate 2b/3b, the optimized structure
is calculated to have a very low energy (AG = +1.3 kJ/mol vs. starting materials) and a
calculated Raman frequency slightly higher than 3-acetylcoumarin (1605 cm"1 vs. 1602
cm"1), fitting with the observed transient Raman signal at 1630 cm"1. The only other
intermediate calculated to have a significant Raman stretching event above 1600 cm"1
and higher energy than the 3-acetylcoumarin
is intermediate
1b (1611 cm"1).
Unfortunately, however, the Raman signal calculated for 2b/3b is very weak40 where the
signal for intermediate 1b is calculated to be very strong. The observed Raman signal for
40
Besides the ability to calculate Raman-active frequencies, relative intensities of these vibrations can be
calculated by Gaussian 03.
46
the unknown intermediate is moderate in strength. As such, this Raman data would
support an argument for either intermediate 2b/3b or 1b accounting for the observed
intermediate. Either there is a very small population of the Raman-active 1b, or there is a
very large population of the less Raman-active 2b/3b. Either would likely lead to the
medium-intensity signal observed at 1630 cm"1.
OH
cc
o
OH
O
-OHl
'EKT^O
AG = +0.7 kJ/mol
OH
OH
OEt
O'^O
•O- NOH
AG = +1.3 kJ/mol
AG = +32.2 kJ/mol
2b
2a
2c
OH
O
OH
O
,
^O
AG = -52.8 kJ/mol
EtO
OEt
Q'NOH
AG = +35.5 kJ/mol
OEt
•O- NOH
AG = +1.3 kJ/mol
AG = +32.2 kJ/mol
3b
3a
3c
Figure 2-23. Mechanistic pathways 2 & 3 differ in only the first, bi-molecular
addition step. The relatively low-energy intermediate 2b/3b is calculated to have
a Raman-active stretching frequency at 1605 cm"1.
Though
computational
modeling
predicts
a slightly
higher
Raman-active
vibrational frequency for intermediate 2c/3c as well, the relative energy calculated is
such that there likely would not be a sufficient population to be observable by Raman
spectroscopy under these conditions. Thus, the intermediate that is observed is may be
2b/3b, but definitive characterization via isolation was not possible. In terms of choosing
either pathway 2 versus pathway 3, there is likely little difference. The existence of
intermediates 2a and 3a would be fleeting, and progression to the shared intermediate
2b/3b would be for all intents and purposes, instantaneous.
2.9 Mechanistic Insight: Kinetic Studies of a Non-Cyclizable Analogue.
Though most of the computational data generated pointed to mechanistic pathways 2 or
3 as being the most plausible, it was deemed that further investigation into pathway 1
was warranted for a few reasons. First, intermediate 1b is predicted by computational
methods to have a very strong Raman-active stretching mode just above the predicted
47
stretching mode calculated for the 3-acetylcoumarin. Second, though the intermediate
energies are all calculated to be significantly higher than the experimentally determined
activation energy for the reaction, it is important to note that these energies are
calculated in vacuum. As such, these data should only be used as qualitative
guideposts, and the in silico values likely do not represent actual relative in situ energy
levels due to other phenomena such as solvation energies and intermolecular
interactions. Finally, it was important to give further consideration to this possible
mechanistic pathway, as intermediate 1b (complete Knoevenagel condensation) has
been postulated as a reaction intermediate by a number of other research groups.41
Cyclization not Possible
O
OH
+
O
Knoevenagle HO, '
o
^
"Et
O
XX
EtO
o
Michael
Addition
Figure 2-24. An equilibrium possibility arises under reaction conditions.
Maximum theoretical yield might initially be suspected to be about 50% because
the E-isomer of the intermediate 1b cannot undergo cyclization to afford the
coumarin. However, under these conditions, the potential exists that the azaMichael addition of piperidine to 1b (£) would afford the ability of the
unproductive isomer to re-equilibrate, eventually leading to complete conversion
to 3-acetylcoumarin. This possibility might also explain the complex reaction
order with respect to piperidine that was observed.
At first blush, reaction pathway 1 should lead to only 50% theoretical yield if there
is no preference in yielding the E-isomer versus the Z-isomer, as only the Z-isomer
would lead to cyclization to afford 3-acetylcoumarin. However, under these conditions,
(a) Fringuelli, F.; Piermatti, O.; Pizzo, F. Synthesis 2003, 2331-2334. (b) Chizhov, D. L ; Sosnovskikh, V.
Y.; Pryadeina, M. V.; Burgart, Y. V.; Saloutin, V. I.; Charushin, V. N. Synlett 2008, 281-285. (c) Pivonka, D.
E.; Empfield, J. R. Appl. Spectrosc. 2004, 58, 4 1 .
48
intermediate 1b would be a molecule that is electronically very susceptible to Michaeltype addition, as the B-carbon is activated both by the ester and ketone electronwithdrawing groups flanking the a-carbon. Upon addition of the piperidine nucleophile42
to this intermediate, free rotation about the former olefin would be possible, leading to a
re-equilibration of the E-Z population of intermediate 1b, and eventually leading to
complete consumption of starting materials to afford a theoretical 100% yield. Figure 224 illustrates this possibility.
In an attempt to reconcile these competing data, a comparative study was
undertaken where 2-methoxybenzaldehyde was substituted for salicylaldehyde in the
'coumarin' synthesis. The methyl ether will not allow for cyclization, in effect halting the
reaction at the intermediate 1 b stage to yield the (E/Z)-2.4 (Figure 2-25).
In the same manner that had been used to determine the activation energy for 3acetylcoumarin (Figure 2-17), this reaction was examined at a range of temperatures.
The Raman signal was again converted to units of concentration of (E/Z)-2.4 and
variations in Raman signal as a function of temperature were accounted for by
developing appropriate calibration curves. A plot of In k versus 1/T again yielded a
straight line from which the activation energy was calculated to be 51.3 kJ/mol. This
significant jump in calculated activation energy was appreciated even qualitatively while
monitoring the reaction: unlike the synthesis of 3-acetylcoumarin, this reaction was quite
sluggish under these reaction conditions.
For this reaction, a strong Raman peak was observed at 1601 cm"1, significantly
lower than the intermediate Raman signal (1630 cm"1) and lower than the peak due to
the coumarin (1608 cm"1) observed in the synthesis of 3-acetylcoumarin (Figure 2-18).
This seems to refute the existence of 1b (Figure 2-20) as the potential intermediate.
42
Incorporation of the piperidine into a second mechanistic step, too, would explain the complex order with
respect to the piperidine that was observed (Figure 2-15).
49
o
I,
Et
OCH3
I?
||
I
^
N
^
A
° ^ ^ ^^
ethyl acetate
piperidine (cat.)
(f^f
E a =51.3 kj/mol
OCH3
(E/Z)-2A
Figure 2-25 The calculated activation energy (Ea) for the Knoevenagel
condensation of ethyl acetoacetate with 2-methoxybenzldehyde to afford (E/Z)2.4 was determined to be 51.3 kJ/mol. This in conjunction with the fact that the
Raman signal for (E/Z)-2.4 arises at a lower energy of 1601 cm"1 indicates that
the synthesis of 3-acetylcoumarin does not proceed through pathway 1.
As so much of the data in this study contradicts one another, to conclusively
assert one reaction pathway over another would be scientifically unsound. The relative
energy data seem to implicate reaction pathways 2 or 3 more so than pathways 1 or 4
(Figure 2-20), though the Raman data suggest that pathway 1, too, may be valid.
However, without isolation of the reaction intermediate with full spectral analysis or
observation of the intermediate with at least one other spectroscopic technique,43 exact
determination of this intermediate is not possible.
2.10 Reaction Monitoring: Biginelli Reaction. A second, more complex,
reaction was examined to further assess the scope of the Raman spectrometer as a
reaction-monitoring tool. Initially, it was hoped that monitoring the formation of various
products afforded by the Biginelli condensation 44 would be
dihydropyrimidinone
amenable to quantitative reaction studies, as the wide range of electronic and steric
impacts could be assessed for each of the three components of the reaction.
R
^°
X
H N
o o+ *
+
2
H+
NH
*
"3
R2SLR,
r T
y
Figure 2-26. The Biginelli reaction is a three-component condensation among an
aldehyde, a 1,3-dicarbonyl compound, and a urea.
43
44
In situ IR studies of this reaction are planned for the future.
Biginelli, P. Gazz. Chim. Ital. 1893, 23, 360.
50
While the formation of the products exhibited a reasonably strong Raman signal
at 1642 cm"1, a number of difficulties kept this reaction from being an ideal candidate to
thoroughly explore it in a quantitative manner. However, an appreciation of the reaction
kinetics could be developed qualitatively regarding the impact of: (1) acid catalyst type
and loading, (2) nearly first order kinetics with respect to ethyl acetoacetate, (3) more
complex order with respect to both the urea and benzaldehyde reagents, and finally (4)
general reactivity trend based upon the electronic and steric attributes of the
benzaldehyde component. Specifically, the Biginelli condensation among substituted
benzaldehydes, ethyl acetoacetate, and urea was examined (Figure 2-27).
^
K^
^
o
+
O
O +
XA«
HaN
NH2
*"
Eto
—°
HN
YNH
O 2.5-2.14
Figure 2-27. The Biginelli condensation among a range of substituted
benzaldehydes, ethyl acetoacetate and urea was examined qualitatively using
Raman spectroscopy.
First, the Raman spectrometer was used to assess the impact of various acid catalysts
upon the rate of reaction (Figure 2-28). While the organic acids, p-toluenesulfonic acid
(TsOH) and trifluoroacetic acid (TFA), were able to catalyze the reaction, both of the
mineral acids examined (hydrochloric acid and sulfuric acid) were more effective
promoters of this reaction. Though HCI (prepared from 37% aqueous HCI solution) was
slightly better catalyzing the reaction, H 2 S0 4 was utilized to mitigate any impact that
water might have upon the dynamics of the reaction and further complication of the
kinetics.
51
1400
1200
1000
&
w
800
0)
600
-
400
A±AA-"
•
200
0
100
-200 *
• 12%TsOH
200
300
Time (s)
" 1 2 % sulfuric A12%TFA
400
A12%HCI
Figure 2-28. Monitoring the formation of dihydropyrimidinone 2.9 in ethanol with
a range of acid catalysts.
Next, an attempt to determine the reaction order with respect to each of the three
reagents was undertaken. Again, the method of initial rates in conjunction with the
isolation method was utilized.
1.5
S
a>
>
1
0.5
0 M
0
0.5
1
1.5
Concentration (M)
—•— benzaldehyde —•—ethyl acetoacetate —*— urea
2
-
first order
Figure 2-29. Plot of relative reaction rate with respect to reagent loading. While
varying the concentration of the reagent in question, the other two were held to
1.0 M. The dotted line indicates expected rate versus concentration for first-order
dependence. The reaction appears to be nearly first order with respect to ethyl
acetoacetate.
While the reaction is nearly first order with respect to the concentration of the
ethyl acetoacetate, the order becomes much more complex with respect to both the urea
and benzaldehyde components of the reaction (Figure 2-29). The complex order is not
52
unexpected, as the postulated mechanism for the Biginelli condensation contains
numerous pre-rate-determining-step equilibria among the aldehyde, urea, water, and
acid catalyst.45 Furthermore, the rate-determining step is reported to be the addition of
the 1,3-dicarbonyl
into a reactive acyl-iminium species (Figure 2-30). It is not
unreasonable then, that only the order with respect to the ethyl acetoacetate is at all
straightforward.
active electrophile
R=EDG stabilize
B
Figure 2-30. Truncated reaction mechanism proposed by Kappe where a
number of equilibrium steps precede the formation of the acyl iminium ion (A).
Likely, the rate-determining step (r.d.s.) of the reaction transforms intermediate A
to B, where intermediate A is the active electrophile. The subsequent cyclization
and dehydration steps are relatively rapid. For these reasons, there exists a
complex reaction order with respect to the aldehyde and urea components, but
the reaction exhibits nearly fist-order dependence upon the concentration of ethyl
acetoacetate.
Finally, relative reaction rates were examined as a function of benzaldehyde
substitution (Figure 2-31). For this evaluation, one assumption was made: the relative
Raman intensity for the olefin of the ct,f3-unsaturated ester in the product was assumed
to be identical regardless of the substitution
on the phenyl ring. As a first-order
approximation, this should be valid as the phenyl ring is electronically isolated, i.e. it is
not in conjugation with the olefin. Therefore, there should be minimal impact on the
Kappe, C. O. J. Org. Chem. 1997, 62, 7201-7204.
53
polarizability of the double bond and resulting Raman intensity from substrate to
substrate.
°
+
Et(X ^ O
O
o
o + H*N
AA
X)Et
X
12%H 2 S0 4
H3C.
NH2
refluxing ethanol
HN
NH
0
2.5-2.14
compound aldehyde rel. rate
2500 i
I T ° 1.63
Time (s)
Figure 2-31. The relative rates of the Biginelli condensation with respect to the
benzaldehyde substitution could be appreciated qualitatively. In general, more
electron-rich systems reacted faster while electron-deficient ones were slower to
form the product.
Even as a qualitative tool, trends can be spotted from the data generated. It
appears there are two separate qualities that can impact the rate of the reaction: sterics
and electronics. First, the most straightforward, easily spotted and explained, and with
greater impact is that benzaldehydes are able to stabilize the positive charge of the acyliminium intermediate (Figure 2-30, Intermediate A) with resonance donating groups, e.g.
54
4-OCH 3 , and react faster. The stabilization of this reactive intermediate hypothesized by
Kappe would lead to an effective increase of the concentration of this species, thus
accelerating the reaction. While it is evident that sterics do hinder the reaction
somewhat, this impact upon the rate of reaction appears less pronounced than
electronic effects (Figure 2-31, 2.9 vs. 2.11 and 2.13 vs. 2.14).
The quantitative investigation of the Biginelli was abandoned, though with an
appropriate protocol, this could prove to be a rewarding project. A wide number of
attributes could systematically be explored: substitution of the benzaldehyde, utilizing
thiourea in place of urea, and experimenting with a range of 1,3-dicarbonyl compounds
in place of ethyl acetoacetate, for example.46 The main reason this project was orphaned
was due to its overly complex nature. At the time, what was desired was a very
straightforward reaction that had been widely explored in order to compare the kinetic
data generated utilizing Raman spectroscopy with other, well-established methods.
2.11 Quantitative Reaction Monitoring: Chalcones. The aldol condensation
has been widely studied and the formation of polarizable a,P-unsaturated carbon-carbon
double bonds of the chalcone system is highly Raman active and gives a characteristic
signal47 at approximately 1600 cm"1. Following these initial qualitative and quantitative
studies, the Claisen-Schmidt condensation was examined in order to corroborate of
results obtained using Raman spectroscopy with results from other laboratories that
utilized other spectroscopic techniques.
However, for a quantitative approach, discrete calibration curves must be set up for each product.
Furthermore, because the Raman signal for this compound is less intense than when monitoring the
formation of 3-acetylcoumarin, and because there is some overlap of signal for the ethyl acetoacetate,
benzaldehyde, urea, and the product, it would be essential that calibration curves be established with the
appropriate ratio of product to starting material. For example, if the reaction is to be investigated at 1.0 M
concentration, a calibration curve should be built that examines solutions that are 0.1 M in product AND 0.9
M in starting materials, 0.2 M product and 0.8 M in starting materials, etc.
47
Dollish, F.R.; Fateley, W.G.; Bently, F.F. Characteristic Raman Frequencies of Organic Compounds, John
Wiley and Sons, New York, 1974.
55
Wavenumber
2250
Figure 2-32. Top: Claisen-Schmidt condensation to afford substituted chalcones.
Bottom: Three-dimensional waterfall plot of Raman intensity over time in the
range of 250-2250 cm"1 generated while monitoring the formation of transchalcone from the sodium hydroxide-catalyzed condensation of benzaldehyde
with acetophenone in ethanol.
It was hypothesized that the condensation between an aromatic aldehyde and an
aromatic ketone (Claisen-Schmidt reaction) would show a strong correlation on the
acidity of the a-proton of the acetophenone derivative. The pKa's for substituted
acetophenones48 are well known and have been studied extensively, thus the formation
of substituted chalcones seemed to be an ideal reaction with which to compare these
results with established techniques and kinetic data.
Again, a calibration curve is used to transform Raman signal intensity per time to
the standard kinetic parameters of concentration per time. Here, however, separate
calibration curves were obtained for each discrete temperature at which the reaction was
performed instead of deriving a temperature scaling factor as had been done when
monitoring the formation of the 3-acetylcoumarin 2.1. A typical plot can be seen for the
calibration curve for 4'-chlorochalcone 2.16 in Figure 2-33 below.
Bordwell, F. G.; Comforth, F. J. J. Org. Chem. 1978, 43, 1763-1768.
56
Calibration C u r v e for 4'-chlorochalcone
12,000.00
10,000.00
o>
(0
c
o>
55
c
•40 C
y = 45826x - 227.71
R2 = 0.9996
8,000.00
• 50 C
y = 44411x-209.46
R2 = 0.99976
6,000.00
60 C
y = 43063x - 205.44
R2 = 0.99977
4,000.00
• 70 C
y = 41736x-202.72
R2 = 0.99972
2,000.00
• 80 C
y = 40819x-220.03
R2 = 0.99944
(0
E
ra
cc
0.00
0
0.05
0.1
;0.15
0.2
0.25
Concentration (mol/L)
Figure 2-33. Calibration curve used to convert units of Raman intensity at 1602
cm"1 to units of concentration (mol/L) for 4'-chlorochalcone at a range of
temperatures from 40-80 °C.
This was simple to do: a solution of a known concentration could be prepared in
ethanol, transferred to the reaction flask, placed in the microwave and brought to the
desired temperature. A dozen scans could be taken, the temperature ramped to the next
desired temperature and a dozen more scans could be taken. Once the complete
temperature range was examined, a second known concentration of the coumarin in
ethanol could be examined. As such, a scientist could generate a calibration curve
converting Raman signal intensity into units of chalcone concentration while accounting
for signal strength variations at a range of temperatures in less than thirty minutes.
For the kinetic studies, enough stock solution containing the appropriate
benzaldehyde and acetophenone in ethanol was prepared from which twenty trials could
be performed; four at each temperature at five different temperatures. Twenty-five
milliliters of this stock solution is placed in a standard 50-ml round-bottom flask with a
Teflon-coated stir bar, placed in the microwave cavity and brought to the desired
57
temperature at which point a dark scan is taken. The appropriate amount of catalyst
dissolved in a small amount of ethanol is injected and a Raman spectrum is collected
aproxmately every seven seconds. The initial rate of reaction can thus be determined
using the first few scans of the reaction; generally under one minute is all that is needed.
All twenty data points can be generated in less than two hours of the chemist's time.
Using the calibration curve, the raw spectral data is converted from units of
Raman intensitysec" 1 to standard units of rate (mol»L"1,mirf1). With this kinetic data in
hand, the Arrhenius plot of In k vs. 1/T yields a straight line (y=mx +b) from which the
activation energies can be calculated (m=-Ea/R, R=8.314 J'M"1«K"1). Alternatively, the
Eyring plot of In (/c/T) vs. 1/T yields a straight line from which the activation enthalpy,
AH*, can be determined. The condensation between 4'-methoxyacetophenone and
benzaldehyde to yield 4'-methoxychalcone, for example, is calculated to have an
activation enthalpy of 59.6 kJ/mol (Figure 2-34).
-2.5 <X.
^ \ ~
-3.5 1
^
y =-7167x+17.53
R2 = 0.998
AH* = 59.6 kJ/mol
\
P
5- -4.5 1
—
^N.
\ ^
^^\
-5.5
-6.5 -T
0.0028
^
t
v
—
0.0029
0.003
i
0.0031
1/T
i
i
0.0032
0.0033
Figure 2-34. Eyring plot of In (/c/T) vs. 1/T for the formation of 4'methoxychalcone 2.21 yields a straight line whose slope is equal to AH'VR.
Activation enthalpies were calculated for a wide range of substituted chalcones
(Figure 2-35). After a few initial studies, it was noted that the substitution on the
58
acetophenone played a larger role in deviations in activation enthalpies when compared
to the unsubstituted chalcone.
Figure 2-35. Calcuated activation enthalpies (AH*) for the formation of
substituted chalcones.
The correlation to the acidity of the acetophenone is not unexpected, as the ratedetermining step is likely tied to the pKa of the a-proton. Therefore, it was surmised that
holding the aldehyde component of the reaction constant while varying the ketone under
investigation should show strong correlation to known pKa data for the acetophenones.
Indeed, this was the case. Plotting the calculated activation enthalpies versus
known pKa values for seven substituted acetophenones (Figure 2-36, open circles)
showed strong correlation (R2=0.985). This data was then used to predict the pKa's for
four previously-unreported acetophenones (Figure 2-36, darkened squares), including:
2'-chloroacetophenone 2.17 (pKa=23.0), 3'4'-dimethoxyacetophenone 2.25 (pKa=25.1),
3'-bromo-4'-methoxyacetophenone
2.26
(pKa=25.1), and
2'-chloro-5'-methoxy-1',3'-
dimethylaceto-phenone 2.27 (pKa=24.3).
59
26.5
y = 0.1035x + 19.509
R2 = 0.98506 •••
26
£ 25.5
3-Br-4-OCH,
O
§
sz
a.
o
25 -j
J0'
3,4-(OCH3)2
o..--"o
24.5
<D
o
to
(0
..©•
24
2-CI-5-OCH3-1,3-(CH3)2
Q- 23.5
2-CI
23
22.5
30
35
40
45
50
55
60
65
activation enthalpy (kJ/mol)
Figure 2-36. Plot of known acetophenone pKa values versus calculated
activation enthalpies for the formation of the chalcone derivatives when
condensed with benzaldehyde (open circles). A strong correlation was noted
(R2=0.985) which allowed the extrapolation of pKa values for unreported
acetophenones (filled squares), including: 2'-chloroacetophenone (23.0), 2'chloro-5'-methoxy-1 ',3'-dimethylacetophenone
(24.3),
3'4'-dimethoxyacetophenone (25.1), and 3'-bromo-4'-methoxyacetophenone (25.1).
2.12 Conclusions & Future Outlook. The use of Raman spectroscopy to rapidly
generate
both qualitative and quantitative
reaction data shows great potential.
Furthermore, Raman spectroscopy couples especially well with microwave assisted
organic synthesis as it is a "through-the-glass" technique and sealed vessel reactions
can be monitored effectively. Previous drawbacks of real-time Raman spectroscopic
techniques largely have been overcome in recent years with a number of advancements.
Currently, research in the Leadbeater research group is limited somewhat by the
necessity to monitor very Raman active molecules. For continued advancements by this
lab, a Raman spectrometer with faster acquisition rates is essential. More sophisticated
equipment, though obviously more expensive, will facilitate the ability to monitor a
broader range of transformations, and ultimately make greater strides in this burgeoning
field.
60
Chapter 3«
Investigations Into "Non-Thermal" Microwave Effects
3.1 Background & Rationale.
versatile
approach
to
synthetic
While the use of microwave heating offers a
organic
chemistry
-
a key
advantage
over
"conventional heating" is that the chemist has access to temperatures well above
normal boiling points of solvents misinterpreted
data
and speculative
there have been a number of reports of
claims
regarding
so-called
"non-thermal"
microwave effects. There have been reports in the literature suggesting that different
product distributions and yields can be obtained if certain chemical transformations
are carried out at the same measured reaction temperature using microwave as
opposed to conventional heating. The focus of this chapter is threefold: an
introductory disscussion of both "specific" and "non-thermal" microwave effects, the
critical examination of a few representitive reports of these effects, and finally, the
use of Raman spectroscopy (and common sense) to refute a number of these
dubious claims.
3.2 Specific Microwave Effects. "Specific" microwave effects are conceptually
straightforward, grounded in sound theory, and generally backed up by well-executed
experiments. Specific microwave effects encompass macroscopic heating events that
occur slightly differently under microwave irradiation than when utilizing conventional
(convection) heating methods. Additionally, specific microwave effects are often difficult,
though not impossible, to reproduce without the use of microwave irradiation. Such
examples would include: (1) observed heating differences based on microwave
absorptivity, (2) inverted temperature gradients, (3) macroscopic superheating, and (4)
selective heating of substances in heterogeneous reaction mixtures.
Wolfgang Pauli coined the phrase, 'not even wrong' to describe an apparent scientific argument that is
based on assumptions known to be incorrect. 'Guilty by association,' though a Red Herring, often is used to
judge something whether a person/place/thing/idea. Both of these phrases pop to mind when discussing the
topic of Chapter 3. It is a dangerous path for a scientist to venture down in order to refute these upcoming
claims, as the best possible outcome for the scientist will be to state the obvious.
62
3.3 Microwave Absorptivity. The first specific microwave effect has already
been touched upon in the opening chapter: substrates that better convert incident
microwave irradiation into heat, heat the bulk faster. Thus, heating 2 ml of water to 100
°C from room temperature will take considerably less time and microwave energy than
heating 2 ml of toluene across the same temperature range and utilizing the same
applied microwave power at 2.45 GHz. While other attributes certainly impact the rate of
heating, because the differences in dielectric loss factors (water: e"=9.89; toluene:
e"=0.096) are so profound, any variations in heat capacities or heats of vaporization will
have negligible impact. However, it is important to note that there would also be
differences in heating rates if heated conventionally, but that any differences would likely
show the highest correlation to specific heat capacities. Indeed, it takes a calculated 170
J to heat 2 ml water (cp = 4.185 >g" 1, IC 1 @ 25 °C) by 80 °C but only about 60 J to heat
the same 2 ml of toluene (cp =1.13 J*g"1*K"1@ 25 °C), ignoring changes to specific heat
across that temperature range.
Luckily, differences in microwave absorptivity are generally of little impact:
commercial m'onomode units are able to heat effectively just about any pure solvent.
Furthermore, as reactions generally have multiple components such as acid or base
catalysts, metal catalysts, and one or more reactants, reactions will often heat much
more efficiently than the solvent alone. Finally, in the extreme cases where microwave
units are unable to heat reactions due to poor substrate and solvent absorptivities,
additives can be utilized which allow the bench chemist access to any solvent system.
Most commonly, ionic liquids50 or reusable inserts such as silicon carbide 51 or Weflon™ 52
For a review, see: Leadbeater, N. E.; Torenius, H. M.; Tye, H. Comb. Chem. High Throughput Screen.
2004, 7,511.
51
Kremsner, J. M.; Stadler, A.; Kappe, C. O. J. Comb. Chem. 2007, 9, 285.
52
Niichter, M.; Ondruschka, B.; Tied, A.; Lautenschlager, W.; Borowski, K. J. American Genomic/Proteomic
Technology2001, 34.
63
have been utilized when microwave transparent solvents such as toluene or hexane
must be employed for a particular reaction.
3.4 Inverted Temperature Gradients. The next highly-touted specific microwave
effect is the phenomenon of inverted temperature gradients when utilizing microwave
irradiation. Conventional heating heats reactions from the outside in, and the glass walls
of the reaction vessel are generally the warmest part of the reaction, especially during
the initial ramp to the desired temperature. Microwave heating, on the other hand, can
flip this gradient as heat is generated across the entire reaction volume and a larger
cross-section of the reaction may reach the ideal reaction temperature sooner than it
would have with conventional heating.
225
£175
I
185
165
J146
I
1105
|
125
a.
£
75
25
Figure 3-1. Infrared image of temperature gradients across an unstirred reaction
heated for 60 seconds with microwave irradiation (left) and conventionally (right).
Adapted J.-S. Schanche, Mol. Diversity 2003, 7, 293 - 300; Reproduced with
permission from Springer.
However, it must be noted that efficient stirring and controlled heating can
generally mitigate temperature gradients, especially extreme gradients, in both
microwave and conventionally heated reactions. Furthermore, it is important to note that
the side-by-side thermal images first published by Schanche in 200353 and reproduced
by Kappe in his review in 2004,M illustrate unstirred reactions that are heated for only 60
seconds either by microwave irradiation or by a conventional oil bath. This image should
1
(a) Schanche, J.-S. Mol. Diversity 2003,7, 293 - 300. (b) Biotage AB; www.biotage.com.
Kappe, C. O. Angewandte Chemie International Edition 2004, 43, 6250-6284.
64
be used as a warning to chemists comparing conventionally heated reactions to those
heated under microwave irradiation, especially when comparing reactions carried out at
very high temperatures for short reaction times. Indeed, this phenomenon likely has
caused more problems than benefits, resulted in more confusion than understanding,
and led to tenuous speculation.
3.5 Macroscopic Superheating. A third example of specific microwave effects is
the phenomenon of macroscopic superheating. 55,56 Solvents will only boil when they are
in contact with their own vapor. Therefore, a solvent that is not in contact with its own
vapor can be heated to above its normal (atmospheric) boiling point without the onset of
boiling.57 This phenomenon can be appreciated when microwaving a degassed solvent
in a pristine reaction vessel. Imperfections in glassware or on boiling stones have areas
that cannot be wetted by the solvents. These areas create small pockets of the solvent's
vapor termed nucleation sites. Without nucleation sites, solvents are only in contact with
their own vapor at the top of the vessel, thus boiling (and hence release of heat) is
limited to this relatively small interface. Chemat and Esveld58 were able to exploit this
phenomenon, and using microwave irradiation, hold solvents well above their boiling
points for extended periods of time. For example, in unstirred vessels without the
nucleation additives, they could hold ethanol at 88 °C (normal bp 78°C) in the microwave
reactor for "many hours." While superheating can occur in conventionally heated
reactions, Chemat and Esveld claim that superheating is more likely in microwave
reactions due to the inverted temperature gradient. The most likely sites for nucleation in
the absence of boiling stones are the pits and scratches on glassware. Under microwave
55
Saillard, R.; Poux, M.; Berlan, J. Tetrahedron 1995, 51, 4033-4042.
Baghurst, D. R.; Mingos, D. M. P. J. Chem. Soc. Chem. Commun. 1992, 674-677.
57
Lienhard, J. H. IV; Lienhard, J. H. V. Heat transfer in boiling and other phase-change configurations: A
Heat Transfer Textbook, Third Edition; Phlogiston Press: Cambridge MA, 2008, 457-463.
58
Chemat, F.; Esveld, E. Chem. Eng. Technol. 2001, 24, 735-744.
56
65
irradiation, however, they claim that reactor walls are likely the coolest part of the
system, thus making a nucleation event much less likely.
Chemat and Esveld go on to exploit this phenomenon by examining two
reactions: the acid-catalyzed esterification of various alcohols with benzoic acid and the
cyclization of citronellal (ene reaction) presumably to afford the four diastereomers of
isopulegol, though the authors do not indicate the observed product. The authors were
able to heat the esterification reaction 38 °C above the normal boiling point of methanol
and carry out the ene reaction 35 °C above the normal boiling point of citronellal.
Accordingly, rate enhancements were observed at these higher temperatures when
compared to conventionally heated—and more importantly stirred—reactions.
O
CH3OH
men
iueq.
(a) k'3' 6 5 ^
^ ^ A
+
n H
U H
O
H2S04(cat.) ^ 5 s ^ X n P u
•
\^fi
n
^T
Temp t1/2 Method
70 4.2 hr MW~
OOH "a
[i^j
7' 0
«
a
'l
1()8
at
4H.0111
3 h r
c^uuv
onv
55mjn
MW
atmospheric pressure
MW, 215 °C
c^
O
(b)
HO
1 atm
b.p. 180 °C
4 diastereomers
Figure 3-2. (a) Acid-catalyzed esterification of benzoic acid in methanol. Using
microwave irradiation, Chemat and Esvald we able to heat the reaction to 108
°C, over 40 °C higher than the normal boiling point of methanol. Accordingly, a
dramatic increase in reaction rate can be obtained, (b) Ene- reaction of citronellal
yielding the four diastereomers of isopulegol. Again, the authors are able to heat
the reaction well above the normal boiling point of the citronellal, even at
atmospheric pressure.
The first three examples of specific microwave effects (observed heating
differences based on microwave absorptivity, inverted temperature gradients, and
macroscopic superheating) are very real, observable phenomenon. Additionally, it has
been shown that these phenomenon can occasionally be exploited and have an impact
on observed reaction rates. However, it is important to note that as a synthetic chemist,
66
the application of these phenomena holds rather little synthetic utility. For example, the
inverted temperature gradient is likely only significant during the ramp portion of the
reaction. Equilibrium will quickly be reached, and under microwave irradiation, the glass
will be only a few degrees cooler than the reaction itself. Certainly, there is not a
massive 60-80 °C difference in wall temperature versus reaction temperature as is
implied by the thermal image first published by Schanche. Furthermore, these wall
effects, as well as the potential for superheating, are both virtually eliminated with
effective stirring. There are likely few synthetic chemists willing to give up stirring for a
few degrees in reaction temperature, as the number of reactions that proceed smoothly
without the aid of stirring is limited. Indeed, an esterification where the solvent is one of
the substrates and neat, unimolecular rearrangements may represent the majority of
such examples.
3.6 Selective Heating. The final example of specific microwave effects is the
ability to selectively heat very microwave absorbent substrates and/or catalysts under
heterogeneous reaction conditions. In a recent example, Strouse and co-workers claim
to have recently demonstrated this specific microwave effect when utilizing microwave
irradiation to synthesize CdSe and CdTe nanomaterials in the non-polar hydrocarbon
solvents heptane, octane, and decane.59 They speculate that the precursor substrates
are able to selectively absorb the microwave irradiation, leading to more uniform
morphologies of the resulting nanomaterials compared to conventional heating methods.
Other similar results have been reported whereby the authors speculate that a rate
enhancement due to the selective heating of highly-absorbing solids. In one example,
Bogdal and co-workers were able to selectively heat the insoluble oxidant C r 0 2
A. L. Washington and G. F. Strouse, J. Am. Chem. Soc. 2008, 130, 8916.
67
(Magtrieve
) in toluene.
Using an IR thermo vision camera, they were able to show
the Cr0 2 was being heated to approximately 140 °C, with no onset of boiling of the
toluene (bp 110 °C).
Specific microwave effects claim to have been demonstrated in enzymecatalyzed transformations as well. Copty and co-workers61 invoke selective heating of
green fluorescent protein by microwave irradiation, leading to denaturing of the enzyme
and hence an increase in fluorescence that is not consistent with the observed changes
in bulk temperature. Similarly, Deiters and co-workers62 claim to observe an increase in
reactivity in three of four hyperthermophilic
enzymes they investigated at bulk
temperatures far below their optimal activity window when the reactions were carried out
under microwave irradiation. It is important to note, however, that this phenomenon may
likely be dependent upon the particular enzyme as other studies have found no
difference in enzymatic activity whether heated with microwave irradiation or when using
conventional heating.63 Indeed, the microwave mediated selective heating at the point of
reaction seems to be the exception rather than the rule, existing in only very specific
instances or highly manipulated protocol, if it exists at all.
3.7 Non-Thermal
Microwave Effects. Unlike specific microwave effects,
venturing into the world of "non-thermal microwave effects"64 puts the scientist on rather
tenuous ground. The three most commonly reported non-thermal effects include: 1)
perceived enhancements in reaction rates that could not be explained according to
(a) Bogdal, D.; Lukasiewicz, M.; Pielichowski, J.; Miciak, A.; Bedaarz, S. Z. Tetrahedron 2003, 59, 649653. (b) Lukasiewicz, M.; Bogdal, D.Adv. Synth. Catal. 2003, 345, 1269-1272.
61
A. Copty, F. Sakran, O. Popov, R. Ziblat, T. Danieli, M. Golosovsky, and D. Davidov, Synthetic Metals
2005, 155, 422.
62
D. D. Young, J. Nichols, R. M. Kelly and A. Deiters, J. Am. Chem. Soc, 2008, 130, 10048.
63
N. E. Leadbeater, L. M. Stencel and E. C. Wood, Org. Biomol. Chem. 2007, 5, 1052.
64
For reviews, see: (a) Perreux, L.; Loupy, A. in Microwaves in Organic Synthesis, 2nd ed. (Ed.: A. Loupy),
Wiley-VCH, Wemheim, 2006, Chapter 4, 134 - 218. (b) De La Hoz, A.;Diaz-Ortiz, A. Moreno, Chem. Soc.
Rev. 2005, 34, 164.
68
typical thermal models, e.g. the Arrhenius equation,
2) claimed decreases in the
entropy of systems under microwave irradiation, and 3) modification of chemoselectivity
or regioselectivity for a reaction that is carried out under microwave irradiation compared
to those heated conventionally.
When scientists report perceived enhancements in reaction rates due to
microwave irradiation, reactions are often performed side-by-side, one in a microwave
and one in an oil bath. There are many examples where increased yields and/or
decreased reaction times are obtained or observed when irradiating with the microwave
versus "conventional" heating and according to the authors, these reactions are carried
out at identical temperatures.66 However, in all cases, significant misinterpretation of
data, systematic error in scientific technique, poor experimental technique, or a
combination thereof lead to poor conclusions, irreproducible results, and generally poor
science.67
In one early example that examines non-thermal microwave effects, Lewis
and co-workers report significant differences in activation energy when the reaction is
heated conventionally (AG* = 105 ± 14 kJ/mol) versus when heated by microwave
irradiation (AG* = 57 ± 5 kJ/mol). 68
condensation
The reaction investigated is a unimolecular
reaction between an amine and carboxylic acid in NMP on a
polystyrene resin. It is also (most) important to note that the microwave portion of the
For a review, see: Loupy, A. Ed., Microwaves in Organic Synthesis, 2nd Edition, Wiley-VCH, Weinheim,
2006, Chapter 4, and references therein.
66
For a comprehensive review, see: (a) Loupy, A. Ed., Microwaves in Organic Synthesis, 2nd Edition, WileyVCH, Weinheim, 2006. For selected examples that report differences based on side-by-side comparison of
conventional vs. microwave heating, see: (b) Louerat, F.; Bourgrin, K.; Loupy, A.; Ochoa de Retana, A. M.;
Pagalday, J. Heterocycles 1998, 48, 161-169. (c) Loupy, A.; Chatti, S.; Delamare, S.; Lee, D. Y.; Chung, J.
H.; Jun, C. H. J. Chem. Soc. Perkin Trans. 1 2002, 1280-1285. (d) Jeselnik, M.; Varma, R. S.; Kocevar, M.
Chem. Commun. 2001, 1716-1717. (e) de la Hoz, A.; Diaz-Ortiz, A.; Moreno, A. Chem. Soc. Rev., 2005, 34,
164. (f) Loupy, A.; Perreux, L. Tetrahedron,2001, 57, 9199.
67
The world of "magical microwave effects" is (unfortunately) expansive. The most comprehensive collection
(though not exhaustive) of these reports can be found in the 133-page compilation of Chapters 4 & 5 of
Perreux, L.; Loupy, A. in Microwaves in Organic Synthesis, 2nd ed. (Ed.: A. Loupy), Wiley-VCH, Weinheim,
2006, pages 1 3 4 - 2 7 7 .
68
Lewis, D. A.; Summers, J. D.; Ward, T. C; McGrath, J. E. J. Polym. Sci. 1992, 38, 1647-1653.
69
investigation was carried out in a domestic microwave without stirring or real-time
temperature mesurements. In these cases, the reacton is heated for a given amount
of time, the reaction is then stopped, removed from the microwave and a temperature
probe is inserted into the reaction. Too much time elapses between cessation of
microwave heating and actual temperature measurement. A more scientifically sound
interpretation of the data simply indicates that there likely are errors in temperature
measurement.
AG* = 57 ± 5 kJ/mol (MW heating)
AG* = 105 ± 14 kJ/mol (conventional heating)
Figure 3-3. Claimed differences in activation energies of an intramolecular
amidation reported by Lewis and co-workers. Though the authors claim to run
the reactions isothermally, the use of a domestic microwave without real-time
temperature monitoring likely leads to inaccurate temperature measurements.
Similarly, Loupy and coworkers have reported a number of studies that they
claim proceed at different rates when under microwave irradiation. They reason that
when the relative polarity of a transition state increases with respect to the reactants,
the microwave irradiation can couple selectively with that polar transition state
leading to increase in microwave absorbance. 69
In one particularly troubling example 70 (though not atypical), Loupy and
coworkers report on the Leuckart reductive amination reaction of aryl ketones in the
presense of formic acid and formamide. They run side-by-side comparisons with
three substrates, isothermally, using microwave heating versus a sand bath.
69
For selected examples, see: (a) Loupy, A.; Perreux, L.; Liagre, M.; Burle K.; Moneuse, M. Pure Appl.
Chem. 2001, 73, 161. (b)
70
Loupy, A.; Monteux, D.; Petit, A.; Aizpurua, J. M.; Dominguez, E.; Palomo, C. Tetrahedron Lett. -1996, 37,
8177-8180.
70
Interestingly, they report conversions to three desired products of 95, 95, and 98%
when utilizing microwave irradiation, but only 2, 3, and 12% using conventional
heating. Though the article is rather ambiguous regarding reaction set-up, thorough
examination of the report leaves little doubt that the conventional reactions were run
at 1 atmosphere (i.e. open vessel) while the microwave reactions were carried out in
a commercial microwave apparatus that automatically seals the vessel. Thus,
alternative (and significantly more plausible) rationale for yield discrepancies emerge
when one considers the reaction temperatures (202, 193, and 210 °C), and the
boiling point of the formic acid (100 °C). Furthermore, it is well documented that
formamide
disproportionates
to
carbon
monoxide
and
ammonia
at
these
temperatures.
3 eq. formamide
202 °C, MW, >98%
^-
202 °C, conv., 2%
3 eq. formic acid
H,CO
OCH;
3 eq. formamide
193°C,MW, 95%
3 eq. formic acid
193 °C, conv., 3%
H3CO
3 eq. formamide
3 eq. formic acid
OCH,
210 °C, MW, 95%
210 °C, conv., 12%
Figure 3-4. Demonstration of "non-thermal" microwave effects by Loupy and
coworkers. Reactions were carried out isothermally, though the microwave
reactions were in sealed vessels while conventionally-heated reactions were
run open to the atmosphere (bp formic acid 100-101 °C, thermal
disproportionation of formamide to carbon monoxide and ammonia).
Indeed, it is precisely for this reason that the Leukart and Leukart-Wallace
reactions are run under sealed vessel conditions. 71 The initial decomposition of
formamide to the volatile gases quickly reaches equilibrium when sealed, allowing for
Webers, V. J.; Bruce, W. F. J. Am. Chem. Soc. 1948, 70, 1422-1424.
71
alternative
pathways
of
reactivity
(i.e.
the
reductive
amination).
Thus,
the
discrepancy of yields is not a "miraculous MW effect, which can be, finally, easily
justified," as put forth by Loupy. Rather, the descrepancy of reaction outcomes can
better be explained by poor scientific technique and the lack of a true experimental
control.
For one final, particularly egregious example of the existence of non-thermal
microwave effects, Hong and co-workers invoke "Unprecedented microwave effects"
when they are able to afford differentiated products in the [4+2] cycloaddition
between 6,6-dimethylfulvene and benzoquinone. 72 Interestingly, no mention is made
of the fact that the microwave-mediated protocol are run in DMSO at 120 °C while
the conventional
method uses benzene at 80 °C. While the authors claim,
"Importantly, these reactions do not occur under conventional thermolytic conditions,"
and that "These results suggest that microwave irradiation can in fact alter the
reaction pathway," no control experiments are run. Indeed, one might suspect that if
there had been a systematic investigation into reaction parameters, the only variable
to NOT matter would be the use of an oil bath versus the use of a microwave to heat
the reaction.
benzene
80 °C
\\
"conventional
heating"
DMSO
//
120°C
"microwave
heating"
Figure 3-5. Demonstration of "unprecedented microwave effects" by Hong
and co-workers.
A number of recent reports have re-visited such claims. In these works,
various techniques are utilized to examine the impact of microwave power on
reaction rates and also to dissect where errors may have previously arose. For
Hong, B. C; Shr, Y. J.; Liao, J. H. Org. Lett. 2002, 4, 663-666.
72
instance, Kappe and co-workers utilize multiple fiber-optic probes,
to get a clearer
picture of temperature gradients and hence inaccuracies in measured and reported
microwave
reaction
conditions.
They
found
significant
variation
in
reaction
temperature based on fiber optic location, especially under heterogeneous reaction
conditions. This effect was most apparent when high initial microwave power is
applied, as detection software cannot acquire data at a sufficient rate to be accurate.
In these cases temperature overshoot is common. Additionally, Kappe and coworkers
utilize silicon carbide heating inserts 74 or simultaneous cooling 75 to investigate the
impact
of
microwave
power
at
a constant
temperature.
Similarly,
a
kinetic
investigation into the intramolecular Newman-Kwart rearrangement by Moseley and
co-workers illustrated identical rates of reaction and hence activation energies and
rate constants regardless of whether the reaction was carried out under microwave
irradiation or heated conventionally. 76 The Leadbeater research group, too, has made
valiant efforts to refute dubious claims, using simultaneous cooling finding that
applied microwave power has no impact on rates of enzyme-catalyzed reactions, and
only that the reaction temperature affects overall rates. 77 As these results continue to
come forward and as previous claims are systematically debunked, one thing
becomes ever more clear: heating is heating.
3.8 Raman Spectroscopic Investigations Into "Microwave Effects." Raman
spectroscopy was utilized in order to investigate the impact of microwave power input on
spectroscopic signatures of molecules. Specifically, the oft-touted phenomenon of
Herrero, M. A.; Kremsner, J. M.; Kappe, C. O. J. Org. Chem. 2008, 73, 36.
(a) Kremsner, J. M.; Kappe, C. O. J. Org. Chem. 2006, 71, 4651-4658. (b) Razzaq, T.; Kremsner, J. M.;
Kappe, C. O. J. Org. Chem. 2008, 73, 6321-6329.
75
Hosseini, M.; Stiasni, N.; Barbieri, V.; Kappe, C. O. J. Org. Chem. 2007, 72, 1417-1424.
76
J. P. Gilday, P. Lenden, J. D. Moseley and B. G. Cox, J. Org. Chem., 2008, 73, 3130.
77
Leadbeater, N. E.; Stencel, L. M.; Wood, E. C. Org. Biomol. Chem. 2007, 1052-1055.
74
73
"localized superheating" was investigated. The remainder of this chapter will present this
work and show that no evidence of any such phenomenon was discovered. 78
The theory of Raman spectroscopy was covered in Chapter 2, but one
phenomenon not fully addressed is the impact of temperature upon a Raman scattering
event. While Raman spectroscopy is not temperature dependent as a whole, when
isolating the respective scattering events, (Stokes and anti-Stokes shifts), a temperature
dependence arises. Namely, the intensity for the Stokes shift is inversely related to
temperature, but the intensity for the anti-Stokes is directly dependent. The substituted
Boltzmann distribution illustrates this point.
anti-Stokes _
Stokes
v
0
]'
kT
O o - V /
Where /=signal intensity, v0= laser frequency, and v}= Raman shift
Because the information delivered by both the Stokes and anti-Stokes shift is
identical, manufacturers of Raman spectroscopy equipment often will report only the
Stokes shift as it is the more likely scattering event and thus will afford a stronger signal
and stronger signal to noise ratio. The spectrometer utilized for these studies is no
different, reporting only the Stokes side of the Raman spectrum. The ramifications of this
means that as temperature increases there are fewer molecules in the ground vibrational
state leading to a concomitant loss of signal intensity, and Raman signal therefore can
act as a monitor for temperature. A schematic of this phenomenon is illustrated in Figure
3-6, below. The practical ramifications of this can be seen in Figure 3-7, where a solution
of benzaldehyde is heated to various temperatures and the Raman intensity for the peak
arising at 1599 cm"1 is plotted versus temperature.
Schmink, J. R.; Leadbeater, N. E. Org. Biomol. Chem. 2009, 7, 3842-3846.
74
Excitation
Virtual
States
Stokes Shift
Anti-Stokes
,.
i ,
*
it
Ground
Electronic State . n
i r.,
^r
Figure 3-6. Schematic of Stokes and anti-Stokes Raman scattering events as a
function of temperature. At low temperatures, a larger cross-section of molecules
is in the ground vibrational state and Stokes shift is very strong. As temperature
increases, however, fewer molecules exist in the grouhd state thus leading to
loss of Stokes signal. Conversely, the anti-Stokes Raman shift exhibits the
opposite phenomenon, increasing in intensity as temperature rises. Arrow widths
depict the relative number of molecules at the given energy level.
4.8
? 46 H
w 4.4
I
• 0.993
4.2 H
f"
•£
3.8
3.6
20
40
60
Temperature (°C)
80
Figure 3-7. A plot of intensity as a function of temperature for a signal arising at
1599 cm"1 from a solution of benzaldehyde in hexane.
When carrying out quantitative reactions studies (Chapter 2), this phenomenon
had to be accounted for by setting up calibration curves to correlate Stokes shift Raman
signal intensity to substrate concentration at each temperature at which the reaction was
run. Over the course of these calibrations, it was hypothesized that this phenomenon
could potentially be exploited as an effective tool with which to probe the hypothesized,
vet never proved, concept of "localized superheating." Could the Raman spectrometer
75
be used as a thermometer of sorts and would (a) changes in Raman signal be
proportional only to temperature, or (b) would there be changes in the Raman spectrum
proportional to applied
microwave
power that were
independent
of
monitored
temperature, thus indicating the existence of "localized superheating?"
A working hypothesis was developed postulating that the concept of localized
superheating did not actually exist and that previous claims of such phenomena were
again the result of data misinterpretation. The primary rationale ironically, being that
"localized superheating" most certainly exists! Indeed, is this not exactly what the
Maxwell-Boltzmann distribution predicts: at a given temperature, there are some
molecules that have a higher kinetic energy than others?
olecules
1 A1
•
i
/
••
Number
\
1 \
i
•
\
hypothetical
activation energy
'
•J
1!
•1 -
A.
Energy
Low Temp
\
\
!
I •
\
\V
Energy
Figure 3-8. Left: Maxwell-Boltzmann energy distribution as defined by the
Maxwell-Boltzmann equation. The total number of molecules must always remain
the same. Right: When a hypothetical activation energy is inserted it can be seen
that there are significantly more molecules with sufficient energy to undergo a
hypothetical reaction at higher temperatures than lower ones. Microwave
chemists often claim reactions that proceed faster under microwave irradiation
then under conventional heating at the same measured temperature! Thus when
researchers claim "localized superheating" effects, either the system disobeys
the Maxwell distribution or temperature measurements are imprecise. Figure
adapted from http://en.wikipedia.Org/wiki/File:Maxwell-Boltzmann_distribution_1,
downloaded February 2010.
Thus, when microwave chemists invoke "rate enhancements due to localized
superheating that are not observable using conventional heating techniques," two
possibilities exist. The first possibility is that under microwave irradiation, the irradiation
76
distorts the Maxwell-Boltzmann distribution such that some number of molecules have a
higher energy, the consequence of which means that a number of molecules must have
a lower-than-expected kinetic energy to have a bulk temperature without change. This is,
in a word, unlikely, as it also implies that the Boltzmann distribution is invalid under
microwave
irradiation.
Alternatively,
data
is
being
misinterpreted,
temperature
inaccuracies are encountered, or non-identical protocol compared. While "localized
superheating" represents a catchphrase designed to excite readers of articles,
conference attendees, or reviewers of grant proposals, what it really implies is a "nonBoltzmann distribution" and thus should be viewed with the highest degree of suspicion
and skepticism! Additionally, since Raman spectroscopy
has its origins
in the
polarizability of a molecule, if microwaves can distort the electron cloud of a molecule or
selectively couple to a more polar region of a molecule as has been championed,
Raman spectroscopy may bear this out.
To begin the investigation into selective coupling of microwave irradiation to one
portion of a molecule, the effect of various solvents upon the Raman signal of
benzaldehyde was examined (Figure 3-10). Benzaldehyde seemed to be an ideal
substrate to investigate: it has a number of Raman active frequencies, is fully miscible in
a wide range of solvents, it has a polar carbonyl bond and non-polar C=C double bonds
of the benzene ring. Thus, if microwave irradiation would selectively couple to the polar
C=0 bond but less so to the non-polar C=C bonds of the benzene ring, or if microwave
irradiation would selectively heat the benzaldehyde when it is in a very non-microwave
absorbing solvent such as hexane, the Raman signal should bear this out. To illustrate
the Raman spectrometer's ability to detect perturbations a molecule's physical attributes,
the Raman signal for benzaldehyde was recorded in a number of solvent (0.4 M
benzaldehyde). As can be seen in Figure 3-10 there exists a significant variation in the
Raman signal intensity for benzaldehyde that is dependent upon the solvent.
77
7
2r 3
(A
e
250
750
-1
1750
1250
2250
1
Wavenumber (cm* )
Figure 3-9. Raman spectrum for benzaldehyde in the region 250-2250 cm"1. The
signal at -1700 cm"1 arises from the stretching mode depicted by the
benzaldehyde molecule on the right, and is largely due to the C=0 stretch.
Similarly, the signal arising at -1600 cm"1 arises by the complex stretching mode
of the benzaldehyde molecule on the left, largely due to the C=C stretch.
Stretching frequencies were calculated using Gaussian '03 at the B3LYP/6-31G
level of theory.
3 i
3
c
V
1*750
Wavenumber (cm*1)
•toluene
Solvent
toluene
MeCN
hexane
ecu
CH2CI2
diethyl ether
—"MeCN
—chloroform
1600:1700
0.739
0.788
0.808
0.892
0.908
0.919
- -MeOH
Solvent
diethyl ether
chloroform
DMF
methanol
ethanol
acetic acid
•AcOH
1600:1700
0.919
0.986
1.070
1.341
1.421
1.775
Figure 3-10. Raman spectrum (1550-1750 cm"1) of benzaldehyde in selected
solvents. A more comprehensive solvent selection, listed with the ratio of peak
intensities for the peak arising -1600 cm"1 to -1700 cm"1 can be seen in the
table. Note how the ratio increases as solvents become more polar.
78
Non-polar bonds such as the C=C bonds of the benzene ring when in a polar
environment are more easily polarized by the incident irradiation of the Raman laser
than when in a non-polar solvent. Because these bonds are more easily polarized,
Raman signal strength increases. Conversely, bonds that exhibit a significant dipole
moment (e.g. C=0 bonds) exhibit the opposite phenomenon. These polar bonds are
more polarized by the polar solvent, thus becoming less polarizable by the incident light
of the laser, ultimately resulting in a loss of signal strength for these bonds. An overlay of
the spectra of 0.40 M solutions of benzaldehyde in a number of solvents is shown in
Figure 3-10 together with relative ratios of the peak heights for the signals arising at
-1600 cm"1 (C=C double bond stretch) and -1700 cm"1 (C=0 double bond stretch) for a
more extensive range of solvents. As the polarity of the solvent increases the signal at
-1600 cm"1 generally becomes more intense and the signal at -1700 cm"1 generally
becomes less intense.
Seeing the variation of intensity in the Raman spectra of benzaldehyde in various
solvents, the impact of microwave irradiation upon the polarizability of a molecule and
whether the microwave irradiation could selectively heat one portion of a molecule at the
expense of another was explored. We chose a solution of benzaldehyde in hexane as
our solvent since hexane has a very low dielectric loss and is almost microwave
transparent (tan 5 = 0.020 @ 2.45 GHz, 20°C). As a result any effects from selective
heating of benzaldehyde molecules would be exacerbated. In order to ensure isothermal
conditions, trials were performed in an open-vessel set-up at reflux temperature and
stirring of the solution constantly with nucleation additives. This allowed us to maintain a
constant bulk temperature while varying the applied microwave power from 0 - 300 W.
Looking at the spectra, there is no detectable variation in peak heights or peak
ratios as a function of microwave power. At all times the temperature was measured to
be 72 °C using an external IR probe. The slight decrease in signal strength at 0 W is due
79
the cessation of reflux, and hence slightly more dilute concentration of the benzaldehyde
under no irradiation. More telling than absolute intensities, however, is the observation
that the relative intensities of the peaks do not change as a function of input microwave
power. Our data indicate that: (a) the benzaldehyde molecules are always at the same
temperature as the bulk solution, (b) there are no signs that microwave irradiation is able
to polarize benzaldehyde molecules, and (c) there is no selective coupling of the
microwave irradiation with a polar region of benzaldehyde over a non-polar region. As a
result, the data indicate that the benzaldehyde molecules are not at a temperature
greater than that of the bulk; i.e. there is not any localized superheating.
= 3
c
S 2
E
2
1550
—0W
1650
Wavenumber (cm1)
W - - 1 0 0 W —200W
— 200 W
- - •50
50W
1750
—-300W
Figure 3-11. Overlay of Raman spectra of a solution of benzaldehyde in hexane
in the region 1550-1750 cm"1. No difference in Raman signal intensity is noticed
as a function of microwave irradiation. The very slight loss of signal intensity
under no irradiation (0 W) is due to the cessation of reflux. As the solvent returns
to the flask from the condenser, the benzaldehyde becomes less concentrated
S4
a>
*•#
£ 3
c
n
E 2
n
CC
1
250
-1
450
—300 W
650
850 1050 1250 1450 1650
Wavenumber (cm1)
200 W —100 W — 50 W
—0W
Figure 3-12. Overlay of spectra for a solution of chlorobenzene in hexane being
heated at reflux (72 °C) using a range of applied microwave power shows no
effect of applied microwave power on the Raman signal of chlorobenzene.
80
To confirm these results, the experiments were repeated, this time monitoring the
Raman spectrum of solutions of chlorobenzene in refluxing hexane. Figure 3-12 shows
the overlay of Raman spectra for chlorobenzene while being subject to microwave
irradiation through a range of intensities from 0-300 W. In all cases the temperature was
measured to be 72 °C. Again, no variation in Raman signal intensity for chlorobenzene
can be detected.
Finally, because there has been so much speculation into localized superheating
of metal catalysts even under homogeneous conditions, one last example was
investigated. This last case proved rather difficult due to the wide range of requirements
that had to be met. Obviously, the molecule needed to have a transition metal
component to probe for selective heating of metal-centered molecules. Secondly, the
molecule needed to be soluble in a non-polar solvent so that the potential for localized
superheating could be exacerbated. Next, the molecule needed to have a Raman-active
stretching mode that could be monitored. Finally, the molecule could not exhibit any
interfering fluorescence.
After screening several possibilities, the system selected as the best candidate
was the chromium carbonyl complex with anisole, Cr(CO)3(r)6-C6H5OCH3) 3.1. As a 0.04
M solution in diisopropyl ether, this system just met all of the above requirements.
PCH3
OC
oc' C /
3.1
Again, the Raman spectrum of the solution was monitored isothermally at reflux
(69-71 °C) through a range of applied microwave power, with constant stirring and the
addition of nucleation chips. Here, the maximum applied microwave power that could be
81
delivered was only 100 Watts owing to the more absorbent system compared to the
benzaldehyde and chlorobenzene in hexane solutions above.
Due to the low concentration of 3.1 in diisopropyl ether, the chromium complex
did not elicit a particularly strong Raman spectrum. However, there were two sufficiently
prominent peaks that could be monitored at 990 and 1904 cm"1. Here, no dark scan of
the solvent was subtracted, allowing for the spectrum of diisopropyl ether to be
monitored for changes as a function of applied power as well.
3800 H
^'^^•Jf-^+W****^
-200250
450
650
"850
1050
1250
1450
1650
"1850
2050
2250
Waven umber (cm-1)
—0W
10 W
=-25 W
50 W
75 W
100 W
di-isopropyl ether
Figure 3-13. Overlay of Raman spectra in the region 250-2250 cm"1 for a solution
of 3.1 in diisopropyl ether. The dotted line is the spectrum for pure diisopropyl
ether for reference (offset along the y-axis for clarity). The peaks at 650, 990 and
1904 cm"1 are due to 3.1 but the rest are due to diisopropyl ether. The most
striking feature of these spectra is that applied microwave power seems to have
no impact on the intensity of the signals due to 3.1 but does have a pronounced
effect on the intensity of the signals due to solvent. The sharp apparent signal at
~1830 cm"1 is an artifact due to a 'stuck pixel' on the CCD detector that arises
sporadically.
82
-ow
10W
-25 W
-50 W
75 W
-100W
Figure 3-14. Close-up examination of the Raman signals at 990 and 1904 cm"
due to 3.1 illustrating that there is no impact of applied microwave power to these
signals.
The Raman spectra for the solution of 3.1 in diisopropyl ether provided
somewhat unexpected results. For the first time, an apparent trend in Raman signal
strength as a function of applied microwave power was noticed, though only in the signal
due to solvent, not in the signal due to the compound itself: the Raman signal intensities
for the peaks at 990 and 1904 cm"1 due to 3.1 remain unchanged regardless of applied
power. A few explanations can be developed to rationalize these observations, though
none are wholly satisfactory.
First, it is possible that as power is increased, loss of signal due to the diisopropyl
ether is due to the more vigorous reflux, which results in less of the solvent in the
reaction flask and a solution that is more concentrated in 3.1 but less concentrated in
diisopropyl ether. However, if this were true, one would also expect a concomitant
increase in the signal due to 3.1 as it would be more concentrated.
The second possibility is similar to the first, but makes an attempt to account for
the lack of change to the Raman signals of 3.1. It is possible that the solution was at the
saturation point with respect to 3.1. Indeed, the attempt to prepare a 0.05 M solution was
met with some difficulty. While the majority of compound 3.1 dissolved, the solution
remained somewhat turbid even after extensive heating at reflux. This was remedied by
83
passing the solution (while still hot) through a short plug of silica gel to yield a clear
yellow solution of 3.1 in diisopropyl ether. It is possible then that as more and more of
the diisopropyl ether resided in the reflux condenser that some of 3.1 would precipitate
out of solution and the Raman signal would remain constant. However, it is unlikely that
this is the case, as any turbidity in the solution (which would be expected as 3.1
precipitates) would lead to significant interference to the Raman spectrum.
The final explanation again concerns the rate of reflux. It is likely that as applied
power increases, the reflux becomes more vigorous, leading to a vessel cross section
that is compose of more diisopropyl ether vapor, i.e. there are more bubbles in the cross
section. However, while this would explain the loss of signal strength of the diisopropyl
ether, a loss of signal due to 3.1 would also be expected.
A number of other metal-organic complexes were examined in order to acquire
more data for homogeneous solutions of metal complexes. However, none were found
that met all of the requirements: strong Raman signals, not fluorescent, and that were
soluble in non-polar solvents. Unfortunately, then, these unusual results could not be
further investigated using similar systems.
3.9 Conclusions & Future Outlook. In situ Raman spectroscopy was utilized as
a tool for probing the "non-thermal" effects of microwave irradiation on various
molecules. For the most part, the results suggest that the local temperature at a
molecular level is no higher than the bulk temperature of the reaction mixture. Certainly,
in completely homogeneous systems (e.g. benzaldehyde or chlorobenzene in hexane)
there is no detected variation of Raman signal intensity as a function of microwave
power when the investigations are carried out isothermally. While the Raman data when
heating homogeneous solutions of soluble organometallic compounds is inconclusive,
they still point toward the conclusions drawn above: namely that there are no "localized
84
superheating" events, and there is no impact of microwave heating on the vibrational
energy manifold of a molecule.
While the microwave energy may interact with polar molecules more so than with
non-polar ones, a rotationally excited molecule is no more prone to reaction than a nonrotationally-excited one. Only once the rotational energy is converted into translational
kinetic (thermal) energy is a higher degree of reactivity associated with the molecule. As
a result, the more polar molecules are not at a temperature greater than that of the bulk
any more than the statistical distribution predicted by Maxwell-Boltzmann; i.e. there is
not any "localized superheating." Instead, the polar molecules effectively heat the
reaction mixture around them.
85
Chapter 4«
Scale-Up
86
4.1 Introduction to Scale-Up Chemistry. Microwave heating is a versatile and
widely used tool for discovery chemistry and is used to perform reactions on small
scales and facilitate initial drug discovery and development processes. The medicinal
chemist appreciates the scientific microwave apparatus due to its ease of use, the fact
that it has been designed around a scale most beneficial to the medicinal chemist (~1
millimole), and lastly because reaction times can be dramatically shortened due to the
ready access to elevated temperatures in sealed vessels.
While the use of microwave heating for performing reactions on the millimole
scale in sealed vessels is straightforward, a number of research groups within academia
and throughout the pharmaceutical industry have been actively addressing the issues
associated with scale-up.79 Possible approaches include continuous-flow reactors,80
small-scale batch stop-flow protocol,81 or large scale, single batch reactors.82 Recent
work within the Leadbeater research group has been focused at exploring all three
possibilities.83
4.2 Typical Large Scale Equipment. For large scale synthesis,84 microwave
manufacturers have for the most part developed four main approaches: 1) continuous-
For a discussion of scale-up of microwave-assisted organic synthesis see: Roberts, B. A.; Strauss, C. R.
in Lidstrom, P.; Tierney, J. P. Eds., Microwave-Assisted Organic Synthesis, Blackwell, Oxford, 2005.
80
(a) Moseiey, J. D.; Lawton, S. J. Chem. Today 2007, 25, 16. (b) Khadlikar, B. M.; Madyar, V. R. Org.
Process Res. Dev. 2001, 5, 452. (c) Kazba, K.; Chapados, B. R.; Gestwicki, J. E.; McGrath, J. L. J. Org.
Chem. 2000, 65, 1210. (d) Esveld, E.; Chemat, F.; van Haveren, J. Chem. Eng. Technol. 2000, 23, 429.
81
(a) Moseiey, J. D.; Woodman, E. K. Org. Process Res. Dev. 2008, 12, 967. (b) Pitts, M. R.; McCormack,
P.; Whittall, J. Tetrahedron, 2006, 62, 4705. (c) Loones, K. T. J.; Maes, B. U. W.; Rombouts, G.; Hostyn, S.;
Diels, G. Tetrahedron, 2005, 61, 10338 (d) Arvela, R. K.; Leadbeater, N. E.; Collins, M. J. Jr. Tetrahedron
2005, 61, 9349.
82
(a) Raner, K. D.; Strauss, C. R.; Trainor, R.W.; Thorn, J. S. J. Org. Chem. 1995, 60, 2456. (b)
Shackelford, S. A; Anderson, M. B.; Christie, L. C ; Goetzen, T.; Guzman, M. C ; Hananel, M. A.; Kornreich,
W. D.; Li, H.; Pathak, V. P.; Rabinovich, A. K.; Rajapakse, R. J.; Truesdale, L. K.; Tsank, S. M.; Vazir, H. N.
J. Org. Chem. 2003, 68, 267. (c) Khadilkar, B. M.; Rebeiro, G. L ; Org. Process Res. Dev. 2002, 6, 826. (d)
Fraga-Dubreuil, J.; Famelart, M. H.; Bazureau, J. P. Org. Process Res. Dev. 2002, 6, 374. (e) Cleophax, J.;
Liagre, M.; Loupy, A.; Petit, A. Org. Process Res. Dev. 2000, 4, 498. (f) Perio, B.; Dozias, M.-J.; Hamelin, J.
Org. Process Res. Dev. 1998, 2, 428.
83
(a) Bowman, M. D.; Holcomb, J. L.; Kormos, C. M.; Leadbeater, N. E.; Williams, V. A. Org. Proc. Res.
Dev. 2008, 12, 4 1 . (b) lannelli, M.; Bergamelli, F.; Kormos, C M . ; Paravisi, S; Leadbeater, N.E. Org. Proc.
Res. Dev. 2009, 13, 634.
84
For a recent evaluation of a wide range of equipment, see: (a) Moseiey, J. D.; Lenden, P.; Lockwood, M.;
Ruda, K.; Sherlock, J.-P.; Thomson, A. D.; Gilday, J. P. Org. Proc. Res. Dev. 2008, 12, 30. (b) Bowman, M.
D.; Holcomb, J. L.; Kormos, C. M.; Leadbeater, N. E.; Williams, V. A. Org. Proc. Res. Dev. 2008, 12, 41-57.
87
flow or the closely-related stop-flow, 2) open-vessel multi-mode reactors, 3) multi-vessel
rotors, and 4) a large-scale, sealed vessel batch approach.
Figure 4-1. Representative scale-up microwave units. A) A rotor unit from AntonPaar. B) The Biotage Advancer, a sealed-vessel batch unit capable of processing
300 ml per run. C) A continuous flow unit from Milestone. D) The CEM Mars,
designed primarily for open-vessel processing. Both a rotor and a autoclave can
be purchased separately to allow for sealed-vessel applications. E) The
Milestone Multi-Synth, a rotor or open vessel unit. F) An open vessel batch
microwave from Milestone.
4.3 Continuous Flow Approach. There are a number of advantages to
continuous-flow chemistry. It limits the amount of material in the microwave cavity at any
given time. As a result, the possibility of catastrophic loss of an entire reaction batch is
greatly reduced. Furthermore, the overall scale becomes essentially limitless and
reactions can be "scaled-out" not "scaled-up." However, continuous-flow processing has
some drawbacks. Many reactions mixtures are heterogeneous, biphasic, or require long
reaction times (e.g. 30-60 min) even at elevated temperatures. Continuous flow
88
technology is generally not amenable in these cases and extensive re-optimization must
be undertaken in order to develop appropriate homogeneous reaction conditions and
suitable residence times. This may require additional solvent and catalyst screening, and
a possible re-evaluation of optimum reaction concentration may need to be undertaken.
A change to either of these precipitates the need to re-evaluate optimum reaction
temperature and time.
4.4 Stop-flow Approach. A stop-flow approach to scale-up has similar
limitations to continuous flow: homogeneous conditions must be maintained throughout
the cycle to avoid clogging issues. Except in the most fortunate of cases, a complete reoptimization of reaction conditions generally is required when moving from batch to flow
chemistry. For better or worse, the fact of the matter is that the majority of small-scale
reactions are optimized under batch conditions. Ideally, the scaling of a protocol from the
milligram scale to the kilogram scale should be straightforward with little need for reoptimization. For the most part, flow chemistry does not meet this criterion and the
development of a batch microwave reactor that could perform reactions on the kilogram
scale would be highly desirable.
4.5 Multi-Vessel Rotor Approach. The multi-vessel rotor approach has a couple
advantages, though likely represents the leash attractive approach overall to the scale
up of microwave chemistry. Three major microwave manufacturers offer rotor equipment
for scale-up: CEM, Anton-Parr, and Milestone. For scale-up, a typical rotor can
accommodate 8-12 vessels each containing on the order of 20-80 ml. Thus, depending
on the manufacturer, 150-500 ml can be processed in a single run. Because each vessel
mimics small-scale conditions, there is a limited need for re-optimization, which is
advantageous. Furthermore, because reactions are split among a number of reaction
vessels, in the event of vessel failure, generally only one reaction portion is lost. Finally,
89
a rotor approach offers the capability of parallel processing at a moderate bump in scale,
i.e. a screen of 8 different reactions in one "batch."
That said, the applicability of a rotor approach is very limited.85 Often, measures
will need to be taken to ensure vessel-to-vessel homogeneity in terms of both
temperature as well as stirring efficiency among the various vessels. Furthermore, from
a real-world standpoint, the individual preparation of individual vessels followed by the
isolation at the completion of the reaction becomes extremely tedious for the laboratory
scientist. Finally, even with multiple vessels, overall scale is still limited to well below one
liter.
4.6 Batch Approach (Open Vessel). The batch approach to scale up has the
distinct advantage that little (if any) re-optimization needs to be carried out when moving
from the small scale. An open vessel batch approach can be utilized if the removal of a
reagent (e.g. water) over the course of a reaction is necessary. Indeed, in terms of realworld usage at very large scales, the open vessel "batch" approach is the only currently
applied microwave approach at very large scales where tons of materials are processed
each day: large microwaves are used to dry foodstuffs, textiles, and other materials by
heating the materials to drive off water.86
4.7 Batch Approach (Sealed Vessel). While a sealed vessel batch microwave
reactor may not be the solution when tons or hundreds of tons of a desired compound
must be synthesized yearly, a batch microwave reactor capable of processing material
at the kilogram scale would help bridge the gap between a small scale protocol and
larger kilo- or pilot-plant scale, adding a much needed tool to the process chemist's
toolbox.87 However, as Moseley and co-workers recently pointed out, "there is at present
no single commercially available scale-up reactor capable of meeting the needs of the
85
See Appendix 4 for further discussion.
See, for example: Decareau, R. V. Microwaves in the Food Processing Industry Academic Press, Inc.
1985, London.
87
Strauss, C. R. Org. Proc. Res. Dev. 2009, 13, 915-923.
86
90
pharmaceutical industry for the wide range of reactions typically required on >1 kg
scale."88
For an illustration of the importance of this realm of scale in organic synthesis,
consider the following scenario. In the development of new chemical entities (NCEs), the
transition from the medicinal chemistry route to a scale where enough material can be
prepared to carry out initial in-vivo toxicology studies is often the most difficult.89 Besides
the fact that the initial route only had to be efficient enough to obtain a few milligrams of
the NCE, medicinal chemists often use chemistry or methodologies (e.g. microwave
mediated transformations) that are currently difficult to scale to the size necessary to
obtain 1-5 kg of the NCE. Due to the prevalence of scientific microwave apparatus in
medicinal chemistry laboratories, process chemists often are delivered protocols that
used microwave heating in one or multiple steps. Thus, those charged with the duty of
the initial scale-up are faced with the difficult choice between running multiple reactions
for the microwave-mediated steps to achieve desired throughput or having to develop
modified conditions to avoid microwave heating. This second option may mean lowering
temperatures, lengthening reaction times and employing different solvents and/or higher
catalyst loadings. In the worst-case scenario, process chemists have to re-design the
entire route to the target compound in order to avoid the microwave-mediated steps
within a reaction sequence. Importantly, as the likelihood of any one of a given 20-50
promising NCEs becoming the active pharmaceutical ingredient in a new drug is low,
companies have a vested interest in pushing any re-optimization of a route to the target
molecule to the latest possible stage, ideally only after screening and toxicology tests
have thinned the field to a few promising candidates.
Moseley, J. D.; Lenden, P.; Lockwood, M.; Ruda, K.; Sherlock, J.-P.; Thomson, A. D.; Gilday, J. P. Org.
Proc. Res. Dev. 2008, 12, 30.
89
Scenario adapted from: Federsel, H-J. Ace. Chem. Res. 2009, 42, 671.
91
Scale Up Method
Continuous Flow
Advantages
• essentially limitless scale
• avoids catastrophic material loss
Stop-Flow
• essentially limitless scale
• avoids catastrophic material loss
• may be able to use small-scale
batch conditions w/o reoptimization
• Direct scale of small-scale
reaction conditions
• Screening multiple substrates
simultaneously a possibility
• an option when a product must
be removed during the reaction
(e.g. water or evolution of gas)
Multi-Vessel Rotor
Open Vessel Batch
Closed Vessel Batch
• most straightforward scale-up
• little/no need for re-optimization
• most attractive option for 0.5-5
kg scale
Disadvantages
• must be homogeneous
throughout reaction
• requires re-optimization from
small-scale batch conditions
• must be homogeneous
throughout reaction
• lower throughput than
continuous flow
• tedious reaction prep
• overall low throughput
• homogeneity (temperature &
agitation) an issue
• limited to atmospheric
conditions
• re-optimization of small scale
conditions (if above atm. bp of
solvent)
• likely not limitless scale
• complexity of microwave
technology means that cost
becomes exponential with
respect to scale
Table 4-1. Breakdown of advantages and disadvantages of various approaches
to scale up of microwave chemistry.
4.8 Microwave-Mediated Transformations: Scale Up Equipment.
Due to
these aforementioned reasons, a study on the development of large batch microwavemediated was undertaken. Research was undertaken primarily on two pieces of
equipment designed for the purpose of the scale up of microwave chemistry: the Biotage
Advancer and the prototype designed and built by AccelBeam Synthesis.
The Biotage Advancer (Figure 4-2) became commercially available in 2005. The
Advancer has a 350-ml Teflon vessel that has a practical operating volume between
100-300 ml. The reaction contents are stirred mechanically with an integrated paddle.
Temperature is monitored via an immersed fiber-optic probe and the maximum operating
temperature is 250 °C. Twenty atm is the maximum allowed pressure. If at the end of the
reaction, the system has generated sufficient pressure, the system can utilize this
pressure to eject the reaction contents into a receiving flask with an approximate volume
92
of four liters. This leads to a rapid, adiabatic cooling of the reaction contents.
Alternatively, compressed air can be passed over the vessel upon reaction completion
similar to small-scale microwave apparatus. However, due to the significantly larger
volumes (100-300 ml vs. 1-5 ml) this cooling method is rather inefficient and lengthy and
therefore less desirable than the rapid, adiabatic cooling. For the Advancer, there is an
integrated PC with software provided by Biotage to control all reaction parameters.
Additionally, all data including reaction temperature, pressure, stirring speed, and
applied microwave power are monitored and recorded on a second-by-second basis.
Figure 4-2. The Biotage Advancer can accommodate reaction volumes up to
300 ml under sealed vessel microwave conditions.
The second piece of equipment that was evaluated for the scale was the
prototype unit built by AccelBeam Synthesis (Figure 4-3). This unit has pushed the scale
of microwave assisted organic synthesis into kilogram scale reactions, filling this muchneeded niche. The microwave unit allows for reactions to be performed on scales from
2-12 L. There are three interchangeable reaction vessels; 5-L, 9-L and 13-L glass
vessels with working volumes of 2-4 L, 4-8 L and 7-12 L, respectively (Figure 4-4, top). A
universal cover (Figure 4-5, bottom left) acts as the interface for peripherals including a
stirring paddle that operates at speeds from 0-125 rpm, a fiber optic temperature probe,
93
the reaction ejection tube and a port for interfacing spectroscopic tools. In addition, a
small port on the cover allows for last minute addition of catalyst, reagents or solvent.
The desired reaction vessel, equipped with the cover is placed in a mechanically sealed
stainless steel reaction chamber capable of operating at pressures up to 350 psi. To run
a reaction, the chamber is pre-pressurized to 250-300 psi using nitrogen gas via a highpressure cylinder. This allows access to reaction temperatures above the normal boiling
points of solvents at atmospheric pressure. The reactor employs three 2.45 GHz watercooled magnetrons rated at 3kW each, with an accessible power of 2.5 kW each for a
total maximum allowed output of 7.5 kW.
Figure 4-3. (A) AccelBeam Synthesis' prototype microwave unit. (B) Bank of
three power supply units capable of delivering up to 2500 Watts each. Upon
completion, reaction contents can be cooled via: (C) ejection through a counter
flow cooling apparatus or (D) directly into a receiving vessel at atmospheric
pressure.
94
^CHHnSLRSS
^CHtrrwiftss
ST
'— • *~L
A
Figure 4-4. Top: Three interchangeable flasks (5, 9, and 13 liters) afford access
to a range of scales. Bottom left: the AccelBeam unit immediately before
reaction. The interchangeable Teflon lid allows for the interface of (A) an
immersed fiber-optic probe, (B) stirring paddle shaft which couples to lid upon
closing, (C) the reaction ejection tubing, as well as a port for the last-minute
addition of reagent, catalyst, or solvent. Bottom right: example of a reaction that
precipitates during the course of a reaction. In this case, the solvent can be
ejected, leaving behind a sponge of the substrate, which is collected via filtration.
Reaction parameters such as time, temperature, pressure, and magnetron power
are monitored and collected using software provided by the manufacturer. Additionally,
this software allows for operation of all pressure inlet and release valves. Microwave
power input was operator-controlled via the built-in analog control dials of each
magnetron's power source. Upon completion of a reaction, a valve in the reaction
ejection line is opened to allow the nitrogen pressure of the reactor to force the contents
out. This flow can be directed through a water-cooled counter-flow heat exchanger, or
directly into a receiving chamber at ambient pressure.
4.9 Rationale of Transformations Investigated. Researching the scale-up of
microwave assisted organic transformations mainly was guided by three aspects. First,
95
wherever possible, the transformation should be relevant industrially. This would include
the actual synthesis of a compound of industrial importance or might mean the general
transformation is one that is widely employed, e.g. condensation reactions. Secondly, in
some cases chemistry was examined which would explore the upper limitations of the
microwave equipment. For example, transformations were selected that were poorlymicrowave absorbent and which would require significant microwave input. Others were
selected to probe the efficacy of the integrated stirring equipment. The third aspect was
strictly adhered to and required that these investigations scaled directly from the small,
medicinal chemistry development stage into larger scales. Reactions were not to be reoptimized. Rather, where problems were met, the cause of these problems was probed.
Rarely were such cases encountered, but where they were, the problem was generally
associated with mass transfer, i.e. stirring. This last aspect is of utmost importance and
is where batch-type scale up shines in terms of practicality. Especially when scaling up
continuous flow or stop-flow microwave protocol, extensive re-optimization is undertaken
as the first step in moving from small to large scale;90 a situation that generally can be
avoided under batch conditions.
Furthermore, the scale-up work that follows can be broken into two main
categories based on project goals at the time. The chemistry carried out in the Advancer
microwave took place throughout 2008 and was really the first foray into sealed-vessel
batch "scale-up" of microwave reactions for the Leadbeater group. Though the increase
in scale (from 1-10 mmol to about 300 mmol) was certainly modest, the small steps
taken here paved the way for the significant strides that were made utilizing the
AccelBeam microwave reactor in 2009. With the AccelBeam, reaction scale could now
surpass the 1-kilogram scale on a per batch basis. As this was a prototype, however, the
chemistry that was developed here played more of a supporting role, and refinement of
90
Damm, M.; Glasnov. T. N.; Kappe, C. O. Org. Proc. Res. Dev. 2010, 14, 215-224.
96
the engineering and utility of this prototype unit was the principle objective. Reactions
were examined in an attempt to find the weaknesses of the unit so that they might be
remedied.
4.10 Large-Scale Suzuki Reactions. Palladium catalyzed C-C bond formation
has seen increasing utility within the pharmaceutical industry over the past 20 years, in
large part replacing traditional C-C bond formation requiring stoichiometric catalyst
loading (e.g. Friedel-Crafts).91 The Leadbeater research group has had extensive
success carrying out the Suzuki coupling at high temperatures with extremely low levels
of palladium to catalyze the reaction. The ability to employ such low palladium levels has
a couple significant advantages. First, palladium is expensive as well as being toxic.
Thus, the pharmaceutical industry often is wary of any protocol utilizing palladium, as
extensive steps must then be taken to recover the precious metal or at least to remove it
from the target compound. The second significant advantage that low palladium loading
in conjunction with high temperatures affords is the ability to use the precious metal in
the form of simple salts like PdCI2 or Pd(OAc)2 and avoid the use of expensive (and
often patented) organic ligands.
Furthermore, the ethanol/water solvent system
facilitates isolation of the biphenyl. A simple filtration and wash with water affords the
biphenyl with high yields and purity.
For a relevant example, the palladium catalyzed coupling of 4-tolylboronic acid
with 2-bromobenzonitrile was scaled in the Biotage Advancer (Figure 4-5). The biaryl
core in Losartan (Cozaar) first developed by Merck is present in a number of angiotensin
II receptor blockers, the sales of which exceeded $10 billion in 2008. 92 As so many of
these compounds contain this structural motif, differing only in the imidazole or
Dugger, R. W.; Ragan, J. A.; Ripin, D. H. B. Org. Proc. Res. Dev. 2005, 9, 253.
Generated from data complied by: Matthew Brichacek, Nicholas McGrath, Erik Rogers, Jason Morton,
Lindsay Batory, Renato Bauer, Jacqueline A. Wurst, Jon T. NjarSarson of Cornell University downloaded
from http://www.chem.cornell.edu/jn96/outreach.html, March 2010.
92
97
imidazole-like head groups, it was hypothesized that an efficient synthesis of the biaryl
core would be worthwhile to illustrate the microwave methodology at larger scales.
Borrowing from previous studies, a high-yielding protocol was developed for the
synthesis of 4'methyl-2-biphenylcarbonitrile 4.1. In 200 ml ethanol-water (50:50 by
volume) and utilizing 3 equivalents of sodium carbonate, 0.01 mol % of PdCI2 would
affect the Suzuki coupling in 97% isolated yield (average of 3 reactions agreeing to
within 3%) when heated to 140 °C for 10 minutes.
Cozaar/Hyzaar
(Merck, $3.71 B)
Diovan
(Novartis, $3.7 B)
Avapro
(BMS, $660 M)
Benicar
(Daiichi-Sankyo, $1.2 B)
Atacand
(AZ, $991 M)
Figure 4-5. The palladium-catalyzed Suzuki coupling between 4-tolylboronic acid
and 2-bromobenzonitrile affords the biaryl core (shown in red). This common
structural motif is shared by a number of angiotensin II receptor blockers whose
2008 sales exceeded $10 billion.
Further studies were undertaken on the Suzuki coupling at various scales. For
cost practicality, 4-bromoanisole was coupled with phenylboronic acid to yield 4methoxybiphenyl 4.2 (Figure 4-6). Optimum conditions were developed that employed 2
equivalents of sodium hydroxide, again in a 50:50 water-ethanol solvent, with simple
PdCI2 as the palladium source It was found that the catalyst loading could be reduced
further to 0.0004 mol % (4 ppm). This reaction first was scaled from 1 millimole in a
standard monomode microwave to 50 millimoles in the Biotage Advancer. The
conditions transferred directly with similar results (93% on the 1 millimole scale and 90%
98
on the 50 millimole scale). In the AccelBeam prototype, the transformation was scaled
further to 4.00 mo I, again with effectively identical results: an isolated yield of 91.8%.
The power and utility of the Suzuki coupling at this scale becomes increasingly evident,
especially utilizing these conditions. The use of inexpensive solvents (ethanol and
water), the elimination of the need for expensive phosphine ligands, and the very low
palladium loadings makes this coupling feasible economically. Indeed, the synthesis of
669 grams of 4-methoxybiphenyl required a mere 1.6 mg of palladium.
C^V
I
H m ^ ^
H3UJ
K^^SV--
+ f| T
I J
\z-
B OH
(
)2
0.0004 mol % PdCI2
2 eq. NaOH
., -,. .
;—»H2U/etnanol
_ _
MW, 150°C, 5min H3CO
\
|j
^ ^ ^ ^ ^
Mono-mode: 1 mmol, 93%
\\ T
Biotage Advancer: 50 mmol, 90%
X ^
AccelBeam: 4.00 mol, 91.8%
^-^
4.2
Figure 4-6. The palladium-catalyzed Suzuki coupling between 4-bromoanisole
and phenylboronic acid was scaled directly from 1 millimole to 50 millimoles in
the Advancer and to 4.00 moles in the AccelBeam without adjusting reaction
parameters.
4.11 Condensation Reactions: 3-acetylcoumarin and Biginelli Reactions.
Condensation reactions generally are excellent candidates for high temperature (e.g.
microwave-assisted)
conditions, especially when the product is stabilized by a
secondary effect like aromaticity. For instance, the formation of 3-acetylcoumarin, an
equivalent of water and an equivalent of ethanol from salicylaldehyde and ethyl
acetoacetate is calculated to be 52.8 kJ*mol"1 (12.6 kcal*mol"1) more stable.93 As such,
this considerable energetic sink allows for aggressive reaction conditions with little need
to worry about unwanted side products. Obviously, elevated reaction temperatures
afford significantly reduced reaction times, but often times also allow for reduction of
catalyst loading.
The synthesis of 3-acetylcoumarin 4.3 was examined on a wide variety of scales
(Figure 4-7). This transformation proved quite valuable for a number of reasons. First,
93
Energies in vacuum calculated using Spartan '08 at the B3LYP/6-31G(d) level of theory.
99
the Leadbeater research group has investigated this transformation in a number of
microwave units on a variety of scales. Next, extensive investigations have been carried
out into mechanistic and kinetic studies of this transformation. Finally, the reactants,
solvent, and catalyst for this synthesis are all very inexpensive, allowing for multiple trials
and access to the relatively large scale (21.0 moles). Indeed, this was the first reaction
examined in the AccelBeam prototype as unanticipated events (if encountered) could be
taken in stride without significant monetary implications.
II
Mono-mode: 5 mmol, 72%
/ ^ s ^ - ^ / V
Biotage Advancer: 300 mmol, 71%
|
T
T
AccelBeam: 3.0 mol, 66.6%"
l l
-^N-An
AccelBeam: 12.0 mol, 74.0%
v
O O
AccelBeam: 21.0 mol, 80.7%
4.3
Figure 4-7. The synthesis of 3-acetylcoumarin was examined at a wide range of
scales including the 13-liter, 21.0-mole scale in one batch. "Artificially low yield
due to blockage of the heat exchanger and loss of product.
O
O
O
+ A
A
OH
' ^ ^
OB
1 mol % piperidine
I^ol
*~
« J * » 20min
M W . u o o.domin
As expected, this reaction scales linearly throughout a range of scales from 5
millimoles to 300 millimoles in the Biotage Advancer and on to 3.0, 12.0, and 21.0 moles
in the AccelBeam prototype. This reaction necessitated mechanical changes to the
AccelBeam, in turn leading to some interesting outcomes. The AccelBeam was originally
designed so that the contents of the reaction would be ejected through a water-cooled
counter-flow heat exchanger. However, in the first attempt at chemistry in the
AccelBeam (synthesis of 3-acetylcoumarin at the 3.0 mole scale, 2 liters) a rather
significant difficulty was encountered. The crystalline coumarin product, while soluble in
ethanol at 130 °C, would rapidly crystallize in the heat exchanger as the solution cooled
which led to complete blockage of the heat exchanger and loss of material (66.6%
isolated yield).
Due to this loss of yield in the initial synthesis of 3-acetylcoumarin, and knowing
that resolving potential clogging issues would be paramount to successful scale-up
protocol, a second cooling option was engineered into the AccelBeam prototype. Instead
100
of passing the reaction contents through a heat exchanger upon completion of a heating
cycle, a flash-cooling option was devised. A five-gallon receiving vessel was integrated
(Figure 4-3 D) where the pressurized contents could be directly ejected in a controlled
fashion. As the contents (280 psi, 130 °C) were ejected into the vessel, the ethanol
would rapidly vaporize, immediately cooling the reaction contents. Not only did this solve
the problem of crystallization in the heat exchanger, it lead to significantly decreased
filtration times in the isolation of 3-acetylcoumarin. Where it took over 8 hours to fully dry
the 3-acetylcoumarin via vacuum filtration when the solution was allowed to passively
cool on the 3.0-mole scale, the product that had been isolated by the flash cooling set-up
was completely filtered in under an hour, and this was at the 12.0-mole scale.
Figure 4-8. Comparison of 3-acetylcoumarin crystals: The left panel illustrates
the granular structure of the crystals that significantly reduced the required time
to filter. The right panel shows the plate-like morphology of the 3-acetylcoumarin
that is allowed to cool slowly which necessitates lengthy filtration times.
Further investigation showed that the crystal structure of the flash-cooled product
is significantly different than the 3-acetylcoumarin that had been isolated after slow
cooling. Figure 4-8 clearly illustrates the granular structure of the flash-cooled coumarin,
which led to significantly reduced filtration times. Conversely, the plate-like morphology
of the passively cooled coumarin crystals led to difficult isolation and lengthy filtration
times.
101
The Biginelli reaction, first discovered in the late 19th century94 has recently
received a renewed interest.95 A number of dihydropyrimidinones have demonstrated
significant promise as anti-cancer drugs.96 It is for this reason that the scale up of the
Biginelli condensation
among
benzaldehyde, ethyl acetoacetate, and urea
was
investigated (Figure 4-9). This protocol scaled directly with great success. When moving
from the small, 2-millimole scale, to the larger 4.0-mole scale, the resulting yields were
nearly identical (54% isolated yield on the 1 millimole scale, 55% on the 4.0 mole scale).
^ x
Etc
°
+
Q
9
II
M
/V/^QEt
ethano1
el lc, u
' " '
MW, 120°C, 20min
Y° f^ii
HlV.NH
HI\L ^
Y
°4.4
Figure 4-9. In a 2000-fold direct scale up in the AccelBeam prototype, the
Biginelli reaction scales directly with no changes in protocol made.
4.12 Williamson Etherification. The synthesis of allyl phenyl ether was
undertaken using a Williamson etherification protocol.97 Standard conditions were
employed, with allyl bromide and phenol as substrates, potassium carbonate as base
and acetone as the solvent. This reaction was used to determine the stirring efficacy in
the Advancer microwave reactor, as effective stirring is important under these
heterogeneous reaction conditions.98 The reaction conditions were first optimized on the
2 millimole scale. The reactants were loaded at 1.0 M in acetone, using two equivalents
of potassium carbonate, and heated to 120 °C for 20 minutes. Moving to the Advancer,
we performed the etherification of phenol with allyl bromide on the 0.2 mole level and
used the same reaction conditions.
94
Biginelli, P. Gazz. Chim. /fa/. 1893, 23, 360.
Kappe, C. O. Ace. Chem. Res. 2000, 33, 879. (b) Stadler, A.; Kappe, C. O. J. Comb. Chem. 2001, 3, 624.
(c) Dallinger, D.; Kappe, C. O. Nat. Protoc. 2007, 2, 317.
Mayer, T. U., Kapoor, T. M.; Haggarty, S. J.; King, W. R.; Schreiber, S. L; Mitchison, T. J. Science 1999,
286,971.
97
For other reports on microwave-promoted ether formation, see: (a) Yadav, G. D.; Desai, N. M. Catal.
Commun. 2006, 7, 325. (b) Park, K. K.; Jeong, J. S. Tetrahedron 2005, 61, 545. (c) Lloung, M.; Loupy, A.;
Marque, S.; Petit, A. Heterocycles 2004, 63, 297. (d) Raner, K. D.; Strauss, C. R.; Trainor, R. W.; Thorn, J.
S. J. Org. Chem. 1995, 60, 2456.
98
Moseley, J. D.; Lenden, P.; Thomson, A. D.; Gilday, J. P. Tetrahedron Lett. 2007, 48, 6084.
95
102
OH
Rr
•Br
/f~~~^
K2C03
' .
acetone
120 °C
^^/O.
rr ^ r
^
^
Mono-mode: 2 mmol, 88%
Biotage Advancer: 200 mmol, 89%
4_5
Figure 4-10. The Williamson etherification between phenol and allyl bromide in
acetone using heterogeneous carbonate as the base tests the capability of the
stirring in the Biotage Advancer.
Efficient agitation was possible using the paddle stirrer. At the end of the reaction
we tried using the flash cooling method but, due to the quantity of solid material in the
mixture, it did not prove efficient due to clogging of the exit line. We did however obtain a
72% isolated yield of the ether product. Adjusting slightly the small-scale conditions, the
reaction was repeated reducing the quantity of potassium carbonate to 1.1 equivalents
and also using passive cooling which, although taking longer (10 min), allowed us to
recover the product with an 89% yield.
Moving to the AccelBeam, a second etherification protocol was developed on the
small scale. The stirring efficacy of this unit had proved efficient for heterogeneous
systems and suspensions in other reactions." Thus, a truly biphasic system was
examined in order to evaluate the stirring efficacy in this scenario. On the small scale, a
set of conditions was developed that efficiently benzylated a range of phenols with
benzyl chloride. The system used a 70:30 solvent system that consisted of acetone and
an aqueous solution of potassium carbonate. On the small, 2-millimole scale, 4-chloro-3methylphenol was quantitatively benzylated after 60 minutes at 150 °C, and the ether
was isolated in 82% yield. However, direct transfer of these conditions to the AccelBeam
at the 2.5-mol, 4-liter scale was met with difficulty (Figure 4-11). Here, the design of the
stirring mechanism was such that the rate of stirring was limited to about 60 rpm. Higher
revolution speeds would cause the shaft to bind. Indeed, on the first reaction attempt,
the stirring shaft jammed after heating for only 10 minutes and had to be heated without
stirring for the final 50 minutes. As such, conversion to the desired ether was low (25%
99
Chad Kormos, Ph.D. Thesis, University of Connecticut, 2010.
103
by 1H NMR of the crude reaction mixture). A second attempt was made with a slower
stirring rate (-40 rpm) with marginally better success (70% conversion after 60 minutes).
However, these results were still well below the expected complete conversion that was
observed at the small scale in a monomode microwave. Thorough, high-rpm stirring is
essential.
CI acetone:water, 70:30
Cl^^p
^ - ^
CH
3
K 2 C0 3
150 °C, 60min
Cl^^-f^
IH
<-H3
4.6
Monomode: 2 mmol, 82% isolated
AccelBeam: 2.5 mol, 25% conv. (1H NMR)
2 nd Attempt: 3.4 mol, 70% conv. (1H NMR)
Figure 4-11. The first attempted benzylations of 4-chloro-3-methylphenol in the
AccelBeam prototype was met with some difficulty. The design of the stirring setup was such that the rate of stirring was severely limited due to the instability of
the stir paddle at high rpm.
Learning from this reaction, a simple modification was made that significantly
increased the stability of the stir paddle: a small divot was placed in the center of the
glass reaction vessels. This new vessel was utilized and allowed for rapid stirring in one
final attempt at the Williamson etherification. Here, 4-chlorophenol was utilized in place
of the 4-chloro-3-methylphenol used above.100 The stability of the stirring paddle was
much improved and this reaction was run with the stir speed set to 108 rpm.
Unfortunately, this precipitated a second problem. As the reaction progressed, a
significant rise in pressure was observed during this reaction that had not been seen
before. The rise in pressure necessitated modulation of microwave power to keep the
reaction below 130 °C, a full 20 °C below the target temperature. It is hypothesized that
the high temperatures and vigorous stirring facilitated the decomposition of the
bicarbonate by-product to C0 2 .
100
The newly modified vessel was the largest, 12-L reaction vessel with a minimum reaction volume of
about 8 liters. The larger scale required significantly more material and 4-chlorophenol was cheaper than 4chloro-3-methylphenol, though should be electronically similar enough for valid comparison.
104
OH
/^^y
+
Cl
/ ^ ^ ^
1|
M
If
^
n
,
acetone:water, 70:30
K
CO
177^
*•
i30°C,60min
C
l
J iI. ^ J
^ ^
AccelBeam: 7.78 mol, 82% conv,
1.075 kg isolated (63.2%)
4.7
Figure 4-12. With a modified reaction vessel, a higher stirring rate (108 vs. ~40
rpm) could be utilized. Unfortunately, a build up of reaction vessel pressure
required that power be modulated to keep the contents a full 20 °C below the
optimum 150 °C. However, the efficient stirring proved its worth as 82% of the
phenol was converted to the benzyl ether and the product was isolated in a
moderate 63.2% yield.
Surprisingly, even though the reaction was not at the necessary temperature, the
crude 1H NMR revealed that the reaction had reached 82% completion. 1.075 kg of the
desired ether 4.7 was isolated in 63.2% yield, illustrating the impact of efficient stirring.
Unfortunately, no further investigations into the etherification reactions were carried out
due to time restrictions and other mechanical issues that cut time short with the
prototype. However, the modified vessel was able to support much higher stir speeds
and was even demonstrated in the lab to effectively emulsify a mixture of vegetable oil
and methanol. Hopefully, this orphaned (but nearly complete) etherification project will
be finished with the second-generation AccelBeam microwave reactor when it becomes
available.
4.13
Aza-Michael
Synthesis
and
Claisen
Rearrangement
as
Heating
Efficiency Protocols. The Leadbeater research group has developed a microwavepromoted synthesis of /V-aryl functionalized (3-amino esters using a Michael addition
protocol.101 These reactions are performed using anilines and methyl acrylate as
substrates and are catalyzed by acetic acid and the reaction requires significant
microwave input to reach the target temperature of 200 °C owing to the relatively
microwave-transparent reagents. Therefore, it was used to qualitatively assess the
ability of the Advancer to apply sufficient power to heat the reaction to 200 °C. Indeed, it
did so with ease and using aniline and methyl acrylate as reagents and performing the
101
See Appendix 4.
105
reaction on the 1.5 mole scale, a 76% conversion to the desired A/-aryl functionalized pamino ester 4.8 was obtained. This was comparable with the results obtained on the
development scale (15 mmol, 81%).
0
MU
NH
2
K ^
H
II
II
200 °C
^s^N^^\^OMe
lOmoiyoAcOH1^^
O
4.8
\#^
11 mol% TBAB \ ^ \ ^ ^ .
4
-5
Vs^
4.9
83%
17%
Figure 4-13. The aza-Michael addition of aniline to methyl acrylate to afford 4.9
and the Claisen rearrangement of ally phenyl ether to afford 4.10 were used to
asses the ability of the Advancer microwave to heat reactions that are not
particularly microwave absorbent. The Advancer was able to heat the azaMichael reaction effectively, but a small amount of TBAB had to be added to the
Claisen rearrangement in order for it to heat the reaction to the desired 245 °C.
Unfortunately, this led to some competitive formation of phenol (17%).
Similarly, the Claisen rearrangement of allyl phenyl ether was examined as this
neat reaction is extremely microwave transparent and even at the small scale, the
scientific microwave apparatus has difficulty bringing the reaction to the necessary 245
°C. Indeed, when using the Biotage Advancer, tetrabutylammonium bromide was added
to increase the microwave absorbance of the reaction in order to reach the desired
temperature. The transformation was performed using different quantities of TBAB and it
was found that 11 mol % was the minimum required in order to be able to heat the
reaction mixture to the target temperature of 245 °C. Holding at this temperature for 30
min allowed us to obtain an 83% conversion to the desired rearranged product 4.9.
However, this was accompanied with a 17% conversion of the allyl phenyl ether to
phenol that was not apparent without the TBAB additive on the small scale.
4.14 Four-Step Reaction Sequence. In order to simulate a situation where
multiple sequential microwave steps were employed in order to reach a desired target
compound, a sequence of reactions was developed as a medicinal chemist might at the
106
~1 millimole scale in order to synthesize a drug-like molecule.102 A base-catalyzed
condensation
between thiourea
and ethyl acetoacetate
in ethanol afforded
6-
methylthiouracil 4.10 using an adaptation of a literature procedure.103 This reaction was
scaled from 2 millimoles in a monomode microwave unit to 2.0, 4.0, and 5.5 moles in the
AccelBeam prototype reactor.
O
O
1 1
/V/^-oEt
1
S
A
+
H2N
NH 2
eq-K0H
ethanol
M W 125 °n
' .
25 mm
°<y^Y'CH
J, H
HJ,
n N
Y
»
4 / | Q S
3
Mono-mode: 2 mmol, 70%
AccelBeam: 2.0 mol, 90.0%
AccelBeam: 4.0 mol, 93.0%
AccelBeam: 5.5 mol, 75.0%
Figure 4-14. The first of the four-step sequence was the hydroxide catalyzed
condensation of thiourea with ethyl acetoacetate to afford 6-methylthiourail 4.10.
Here, the reaction performs better at the 2 and 4-mole scales as the
The 2.0 and 4.0 mole scales provided isolated yields of the intermediate of
90.0% and 94.6%, respectively. This reaction performed significantly better at scale
when compared to the 1-2 millimole optimization runs. This can be attributed to the fact
that over the course of the reaction, the potassium salt of the thiouracil product is
formed. Even at 125 °C in ethanol the product readily precipitates. On the small scale,
monomode microwave units utilize magnetic stirring placed to agitate the reaction
mixture. It is likely that magnetic stirring becomes ineffective as the precipitate forms,
thus hindering the reaction and resulting in lower yields. Here, the differences between
high-rpm, low-torque stirring and low-rpm, high-torque stirring can be appreciated. While
the stirring in the benzylation reactions described above was more efficient in the small,
monomode microwave, that biphasic reaction required rapid stirring (high rpm) but did
not generate much drag (torque). Conversely, for this condensation reaction, rapid
stirring is not essential, but the precipitation of the solid on the small scale halts the
magnetic stirring completely. On the larger scale with mechanical stirring, however, this
problem is mitigated and an appreciable increase in product yield is achieved, even
102
This sequence was developed in a collaborative effort and is also described elsewhere: Chad M.
Kormos, Ph.D. Thesis, University of Connecticut, 2010.
103
Mai, A.; Artico, M.; Sbardella, G.; Massa, S.; Novellino, E.; Greco, G.; Loi, A. G.; Tramontano, E.;
Marongiu, M. E.; La Colla, P. J. Med. Chem. 1999, 42, 619.
107
before the modification to the glassware was made (divot). However, mass transfer
again became problematic when attempting to scale-up to 5.5 moles (11 L volume). The
quantity of precipitate now proved too significant for even the mechanical stirring
mechanism, binding the motor, requiring the reaction to be halted prematurely, and
leading to a comparably low isolated yield of only 74.9%.
In the second step, the selective benzylation of the sulfur functionality on 6methylthiouracil to form benzylthiouracil 4.11 was performed. This protocol was adapted
from one reported by Botta and co-workers. 104 Using benzyl chloride in DMF, with an
equivalent of K 2 C0 3 and heating the reaction mixture to 100 °C for 25 minutes proved
optimal. Performing the reaction on the 8.4-millimole scale using a monomode
microwave unit, a 70% isolated yield of 4.11 was obtained. A comparable yield (64.0%)
was obtained upon scaling up to 3.4 moles using the batch reactor.
1 eq K 2 C 3
°^^Y'CH3
HN
NH
*Y
S
4.10
+
f l * T ^
O J
C I
'
°v^V^CH3
°
*~
MW, 100 °C
UM
M
H N - ^ N
25min
s
411
Mono-mode: 8.7 mmol, 70%
AccelBeam: 3.4 mol, 64%
Bn
Figure 4-15. The benzylation of 6-methylthiouracil scaled smoothly from 8.7
mmol to 3.4 moles.
Next, a POCI3/Et3N deoxychlorination protocol was developed on the small scale
(Figure 4-16). A 3.0 M suspension of 4.11 in POCI3 was prepared, to which one
equivalent of triethylamine was added. At this point, the reaction was heated to 120 °C
for 10 minutes in the microwave. A careful quench of the reaction with cold saturated
bicarbonate solution followed by extraction with ethyl acetate afforded 88% isolated yield
of the 3-chloropyrimidine 4.12 at the 3.4 millimole scale and 98% isolated yield at the
14.7 millimole scale, both with >95% purity. However, calorimetry studies indicated an
exotherm of approximately 120 kJ/mol upon addition of the triethylamine to the reaction
104
Manetti, F.; Este, J. A.; Clotet-Codina, I.; Armand-Ugon, M. Maga, G.; Crespan, E.; Cancio, R.; Mugnaini,
C ; Bemardini, C ; Togninelli, A.; Carmi, C ; Alongi, M.; Pretricci, E.; Massa, S.; Corelli, F.; Botta, M. J. Med.
Chem. 2005, 48, 8000-8008.
108
mixture. Because the chlorination reaction proceeds smoothly at lower temperatures
albeit with longer reaction times, it was deemed prudent to carry out the reaction
conventionally. The exotherm was used to heat the reaction initially, at which point the
reaction was heated to 100 °C in a heating mantle where it was held until TLC indicated
reaction completion: about 90 min at 100 °C giving a total reaction time of 2 h.
°r^rCH3poc,3/Et3N^Clv^rCH3
HN^N
T
SBn
MW 120°C
' .
10mln
4.11
N
N
-vf?|
SBn
Monomode: 3
^ ^
88%
Mono-mode: 14.7 mmol, 98%
Large Scale: conventional, 87.5%
4.12
Figure 4-16. Deoxychlorination of 4.11 to afford chloropyrimidine 4.12. While the
medicinal chemistry (small-scale) route was developed using microwave
irradiation with excellent results, at the large scale, the thermochemistry of the
reaction necessitated conventional heating for safety.
The final step of the multi-step sequence was an acetic acid catalyzed
nucleophilic aromatic substitution (SNAr) protocol, originally developed on the 0.1
millimole scale by Schultz and co-workers. 105 Here, aniline, chloropyrimidine 4.12, and 1
equivalent of acetic acid was heated to 150 °C in dioxane and held for 10 min. On the 1
millimole scale, a 9 1 % yield of 2-(benzylthio)-6-methyl-4-(phenylamino)pyrimidine
hydrochloride 4.13 was obtained. Scaling to 1.84 moles using the batch reactor afforded
a 75% yield of 4.13, this representing a greater than 18,000 fold increase in scale over
the original published procedure and an 1,800-fold increase in scale in house, with no
changes made to the protocol.
NH 2
C
+
V^™
3
II ^TN
N
T
SBn
4.12
Up" AcOH
i eq.
ACUH ^
dioxane
MW, 150°C
10 min
fi^V^V^V^3
r.
K ^r
jN
}
^JN „ „ .
T
Monomode: 1.84
1 mmol,
1%
AccelBeam:
mol, 975%
S B n
4.13
Figure 4-17. The acetic acid catalyzed SNAr reaction between aniline and
chloropyrimidine 4.12 utilized conditions first developed by Schultz and coworkers. A small decrease in isolated yield was noted for this transformation
when moving from the small, 1 millimole scale, to the 1.84 mole scale.
105
Wu, T. Y. H.; Schultz, P. G.; Ding, S. Org. Lett. 2003, 5, 3587.
109
Overall, the sequence employed three microwave steps and afforded 473 g of 4.13 in
38% overall yield for four steps from thiourea and ethyl acetoacetate. This is almost
identical to the 39% overall yield obtained on the developmental scale. Thus, using three
identical reaction steps (out of four) with no adjustment other that quantity of reagents
used for a sequence of four reactions developed at the one millimole scale, comparable
results were obtained, making this of great potential benefit to the process chemist.
H
CH3
A
s
ethanol
HN NH
jj^ +
MW, 125 "C
J
H 2 N^NH 2
25 min
S ^
39.2% over 4 steps, small scale
B
S
J^ +
H 2 N^NH 2
ethanol
MW,125°C
25 min
HN
^MF *"
MW,100°C
25 min
NH
I
S
u
38.0% over 4 steps, 1.82 moles
HN
N
-f*
'
^Bn
^
*" HN_,N
MW,100°C
g
**"
25 min
a 11
*•''
MW,120°C
1 0 min
100-c
2 hr
N
v^N
J
^SBn
N^.N
T
SBn
4.12
dioxane * ^ ^
MW,150°C
i0mjn
N
~f
-HCI
SBn
^
dioxane * ^
N
Y
SBn
MW ,150°C
-HCI
4.13
Figure 4-18. Side-by-side comparison of the 4-step synthesis to yield 4.13. The
"medicinal" chemistry route A provided the product in 39.2% overall yield and the
large-scale route B in the AccelBeam prototype microwave yielded a comparable
38.0% yield of 4.13.
4.15 Large-Scale Batch Solvent Heating in AccelBeam Prototype. In order to
assess the efficacy of heating efficacy of the AccelBeam prototype at the very large
scale, seven solvents - ethanol, water, acetonitrile, ethyl acetate, THF, 2-butanone
(methyl ethyl ketone, MEK), and DCM - with a range of microwave absorptivities were
heated. It was hypothesized that the typical "good" microwave absorbers such as water
or ethanol would heat efficiently, "moderate" absorbers would require a bit more input,
and the "poor" absorbers such as dichloromethane, THF, and toluene would not heat
efficiently due to the large volume that would need to be heated and the low microwave
absorbency of these solvents.
In all cases, 4.0 L of solvent was heated in the 5 liter reaction vessel until the
contents reached 150 °C. A constant 7.5 kW power delivery from the magnetrons (2.5
110
kW x 3) was utilized with the stirring set to 60 rpm and the temperature ramp from 30150 °C was recorded.
175
0:00
1:00
2:00
3:00
4:00
5:00
6:00
7:00
8:00
9:00
10:00
Time (minutes)
•2-butanone
DCM
•acetonitrile
•THF
•ethyl acetate
•ethanol
•water
Figure 4-19. Plot of temperature versus time for 4.0 liters of seven different
solvents under microwave irradiation in the AccelBeam prototype microwave.
Figure 4-19 illustrates that It took between 4-5 min to heat 4 L of MEK (4 min 13 s), DCM
(4 min 15 s), acetonitrile (4 min 21 s), THF (4 min 28 s), or ethyl acetate (4 min 58 s)
from 30-150 °C. Perhaps surprisingly, the solvents that performed the best in this study
are ones often regarded as "poor" microwave absorbers. On the other hand, those
generally touted as "good" solvents for microwave chemistry actually performed
relatively poorly: 4 L of water takes over 9 min to reach 150 °C.
An often-overlooked component to solvent selection is the heat capacity106 and,
as a result, the required amount of energy to heat it. As a rough estimate, the amount of
energy required to heat 4 L of water by 120 °C was calculated to be approximately 2.01
x 103 kJ. The same volume of ethanol requires less than half that at 924 kJ and all the
Also, though to a lesser extent, the impact of heats of vaporization cannot be overlooked. In the
AccelBeam prototype, the vessel is charged with 280 psi N2 prior to heating. In contrast, typical small-scale
microwave reactors operating in sealed-vessel mode generate autonomic pressure. As a result, when
heating with the large scale unit, energy loss due to solvent vaporization is minimized.
111
other solvents require below 900 kJ (4 L, A = 120 °C). Indeed, plots of time to reach
desired temperature versus: A) tan 5, B) dielectric loss, C) dielectric constant, and D)
calculated energy required to heat the solvent through 130 °C, showed significant
correlation only in the last case (Figure 4-20).
10:04.8
08:38.4
_
07:12.0
•|
05:45.6
|^ 04:19.2
*
R2 = 0.02257
02:52.8
01:26.4
00:00.0
0.2
0.4
0.6
0.8
10
tan 3
^
15
dielectric loss
10:04.8
10:04.8
08:38.4
08:38.4 _
07:12.0
07:12.0
in
•|
05:45.6
^
04:19.2
R2 = 0.77621
|
05:45.6
^
04:19.2
02:52.8
02:52.8
01:26.4
01:26.4
00:00.0
R2 = 0.95691
00:00.0
20
40
60
dielectric constant
80
100
500
1,000
1,500
2,000
2,500
kJ required, AT = 120 °C
Figure 4-20. Plots of time required to heat various organic solvents versus: a)
tan 5, b) dielectric loss, c) dielectric constant, and d) calculated energy required
to heat the solvents by 120 °C using a simple approximation (E=mcAT: E=energy
in kJ, m=mass of solvent, c=specific heat capacity at 25 °C, AT=120). At the
large scale, only the plot versus energy required shows significant correlation.
Thus, the data seem to indicate that only specific heat has a strong correlation
when heating using microwave irradiation. However, this is certainly counterintuitive to
any experienced user of microwaves in chemical synthesis, i.e. more absorbent systems
heat faster. However, this is only true if the entire sample cross-section can effectively
be irradiated. On the small scale this is the case, which means solvents that are highly
microwave absorbing will heat faster than those solvents that have low microwave
112
absorptivities. At larger scales, the depth to which the microwave energy can penetrate
the contents of the vessel will vary. Less absorbent solvents such as DCM or THF have
a larger cross-section that is absorbing the microwave energy as compared to more
absorbent solvents like ethanol and water. Microwave absorptivity and heating cross
section are interlinked and inversely proportional which means that dichloromethane (tan
3=0.042) can be heated to 150 °C from 30 °C (AT=120 °C) in less time than it takes to
heat ethanol (tan 3=0.941) across the same temperature range using the same applied
microwave power.
highly microwave absorbant
<
low cross-section MW penetration
*
>
y
low microwave absorbancy
high cross-section MW penetration
net result: efficient heating across a wide range of solvents
Figure 4-21. Schematic illustrating the impact of the inverse relationship between
penetration depth and microwave absorptivity. A highly absorbent solvent will
have a low penetration depth but a poorly absorbing system will have a larger
penetration depth. Therefore, with sufficiently large vessels, these two impacts
will negate one another leading to efficient heating across a wide range of
solvents and the time required to heat a reaction will only depend on the specific
heat capacity of the solvent.
The ramifications of these observations are significant. First, many researchers
intimately involved with the scale up of microwave chemistry have reluctantly viewed
microwave-mediated batch scale up as limited in scope. Many have hypothesized that
113
flow chemistry would be the only suitable approach for >1 kg scale up using microwave
heating, citing poor penetration depth as a potential issue. While there are a number of
excellent reasons to adopt flow chemistry for large-scale synthesis, they hold true
regardless of whether microwave or conventional heating is used. These results show
that, with efficient stirring and properly sized and engineered magnetrons, it should be
possible to design batch reactors capable of effectively heating reactions on significantly
larger scales than currently used.
Of course, there are limitations to this phenomenon that must be considered and
the repercussions of ignoring the limitations are significant. It can clearly be seen above
that as microwave absorptivities decrease, the depth-of-penetration increases and these
two phenomena effectively cancel one another. However, if the scientist extrapolates to
extremely non-absorbent systems, it should be apparent that a near-infinite volume will
be required to absorb all of the incident microwave irradiation.
In order to probe the limits of this phenomenon, an attempt was made to heat
toluene (tan 3=0.040,
E'=2.4, E " = 0 . 0 9 6 )
in a similar fashion. However, the overall volume
proved to be insufficient to absorb all of the incident microwave irradiation. Instead,
arcing was evidenced by a rapid increase in pressure due to decomposition of toluene
vapors into molecular hydrogen and elemental carbon black. The run was abandoned
after 15 seconds of microwave heating. Hypothetically, however, a larger solvent crosssection (e.g. 12 liters instead of 4) in conjunction with a more conservative application of
microwave power could feasibly allow for sufficient absorption of microwave energy by
the solvent and reduce the potential for arcing. At this point it should be stressed that it is
unadvisable to carry out reactions in toluene using non-absorbing reagents without
additives to increase the absorptivity of the load.
114
Figure 4-22. Result when attempting to heat toluene in the AccelBeam prototype
microwave reactor. The volume of toluene was not sufficient to absorb all of the
incident irradiation. This led to arcing and the rapid decomposition of the toluene
to its elemental constituents, H2 (g) and C (s).
4.16 Conclusions & Outlook. The application of microwave heating to larger
and larger scales is likely bumping against the upper practical limits. Certainly, in terms
of cost-to-benefit, the microwave reactor enjoys a practical and even beneficial ratio on
the small, medicinal scale, driven primarily by the time saved of the bench chemist and
the increased scientific throughput. However, as scale increases, the cost to utilize
microwave technology increases at pace not matched by the time saved of the chemist,
thus the costs increase more rapidly than the benefits.
A few areas where the application of microwave irradiation will likely see
increased utility over the next decade will include kilo-scale process development within
the pharmaceutical industry and kilo-scale process development in fine chemicals.
115
Furthermore, with continued refinement, it is possible that microwave irradiation could be
used in pharmaceutical or fine chemical arenas when a few tons of throughput is
necessary on a yearly basis. Based on the work with the AccelBeam prototype, a
microwave unit similar to this but modified as a batch stop-flow unit (which is already in
progress) could easily process twelve 30-minute reactions per 8-hour shift. At a
hypothetical two kilograms per run, 40+ kilograms could be produced each [2-shift] day,
easily surpassing one metric ton in a 4-month timeframe.
The final area that shows significant potential would be in the use of the
microwave batch reactor to perform recrystallizations for the API constituents of
pharmaceutical products. The facile access to high temperatures in conjunction with the
rapid flash cooling of the AccelBeam prototype demonstrated that new recrystallization
protocol could be developed that yield the API in the desired polymorph while potentially
reducing solvent use.
116
Chapter 5 •
Pd-Catalyzed Methodology Development:
Synthesis of Diarylmethanes
117
5.1 Diarylmethane Background & Significance. The diarylmethane structural
motif is found in a range of biologically active compounds and is incorporated into a
number of pharmaceuticals. For example, the diarylmethane substructure is present in a
number of antibacterial and HIV and AIDS treatments (Figure 5-1). Trimethoprim and
piritrexim are currently used for the treatment against opportunistic infections caused by
Pneumocystic carinii and Toxoplasma gondii in patients with full-blown AIDS as well as
for the treatment of urinary tract infections.107 Similarly, the diarylmethane substructure
can be seen in a range of anti-bacterials currently under analysis of which FN075 is one
such example.108
OCH 3
trimethoprim
0CH3
piritrexim
Figure 5-1. Structures of three representative small molecules containing the
diarylmethane moiety. Trimethoprim and piritrexim are currently approved by the
FDA for treatment of bacterial infections and see most use in patients with fullblown AIDS. FN075 is an example of one of many molecules currently under
investigation for its anti-bacterial characteristics that contain a diarylmethane
substructure.
Additionally, the diarylmethane moiety plays a prominent role in a number of
antiretroviral therapies targeting HIV1-lntegrase (IN), the enzyme mediating a crucial
step in the replication of HIV. Examples of small molecules having reached various
stages of clinical development include Shionogi/GSK's S-1360 and Merck's L-870,810.
Both compounds reached Phase lla clinical trials, but development was subsequently
halted due to various pharmokinetic issues (Figure 5-2). Though these initial leads were
scrapped, the diarylmethane moiety remains in Gilead Science's GS-9137, Elviltegravir,
which has reached Phase III clinical trials for the treatment of HIV.
107
Gangjee, A.; Devraj, R.; Queener, S. F. J. Med. Chem. 1997, 40, 470-478.
Cegelski, L; Pinkner, J. S.; Hammer, N. D.; Cusumano, C. K.; Hung, C. S.; Chorell, E.; Aberg, V.;
Walker, J. N.; Seed, P. C; Almqvist, F.; Chapman, M. R.; Hultgren, S. J. Nat. Chem. Biol. 2009, 5, 913-919.
108
118
S-1360, Shiongi/GSK
L-731,988, Merck
GS-9137 (Elviltegravir, Gilead Sciences)
Figure 5-2. Three examples of small molecules containing the diarylmethane
moiety that have reached various stages of clinical trials for the treatment of HIV.
In addition to these examples where the diarylmethane
substructure
is
incorporated into active drug entities, diarylethers and diarylamines are present in a
number of pharmaceutically active small molecules. Thus, the methylene linkage in
diarylmethanes could be seen as a "heteroatom delete group" used to probe for
structure activity relationships (SAR) of these small molecules.
dasatinib (Sprycel, BMS)
erlotinib (Tarceva, Roche/OSI/Genetech)
lapatinib (Tykerb/Tyverb, GSK)
H
H
sorafenib (Nexavar, Bayer)
gefitinib (Iressa, AZ/Teva)
nilotinib (Tasigna, Novartis)
K^f>
imatinib (Gleevec, Novartis)
Figure 5-3. A few examples of molecules containing diarylether or diarylamine
linkages. These antiretrovirals all are currently approved by the FDA for the
treatment of various cancers. Replacement of the diarylamine/ether linkage with
a methylene to afford the diarylmethane analogues could allow for pertinent SAR
studies.
119
For instance, diarylamines and ethers are especially prevalent in kinase inhibitors,
accounting for seven small molecules currently approved by the FDA for the treatment of
cancer containing a heteroatom-linked arene rings (Figure 5-3).109 The replacement of
these heteroatoms with a methylene group could provide detailed SAR data for current
and future drug entities and candidates.
Given the wide range of pertinent examples of diarylmethanes and heteroatom
analogues, it is of no surprise that there has been a recent renaissance in developing
methodologies to systematically effect the formation of the carbon-carbon bond linkage
of the diarylmethane. Furthermore, as palladium-catalyzed C-C bond formation is among
the most powerful synthetic tools available to the organic chemist,110 it follows that the
use of palladium as a catalyst in the synthesis of diarylmethanes has been widely
applied. It is also of little surprise that a number of the recent publications investigating
the synthesis of diarylmethanes have come directly from the pharmaceutical industry
(Figure 5-4).
The palladium catalyzed cross-coupling of arylboronic
acids with
benzyl
halides,111,112 or benzyl phosphates,113 coupling of aryltrifluoroborates with benzyl
halides,114 and the coupling of benzyl indium reagents with aryl iodides115,116 illustrate a
few recent approaches in the synthesis of diarylmethanes. Inverting the reactivity,
Flaherty and co-workers have successfully coupled S-benzyl-9-borabicyclo[3.3.1]nonane
1ua
Zuccotto, F.; Ardini, E.; Casale, E.; Angiolini, M. J. Med. Chem. 2009 DOI: 10.1021/jm901443h.
(a) Handbook of Organopalladium Chemistry for Organic Synthesis; Negishi, E., Ed.; Wiley-lnterscience:
New York, 2002; Vol 1. (b) Hegedus, L. S.; Transition Metals in the Synthesis of Complex Organic
Molecules, Second Edition; University Science Books: Sausalito, CA, 1999.
111
For the first reported example of a Suzuki-Miuyra coupling, in this case between a benzyl boronate and
an aryl halide, see: Maddaford, S. P.; Keay, B. A. J. Org. Chem. 1994, 59, 6501.
112
For recent examples, see: (a) Nobre, S. M.; Montiero, A. L Tetrahedron Lett. 2004, 45, 8225 (b) Burns,
M. J.; Fairlamb, I. J. S.; Kapdi, A. R.; Sehnal, P.; Taylor, R. J. K. Org Lett. 2007, 9, 5397.
113
McLaughlin, M. Org. Lett. 2005, 7, 4875.
114
Molander, G. A.; Elia, M. D. J. Org. Chem. 2006, 71, 9198.
115
Chupak, L. S.; Wolkowski, J. P.; Chantigny, Y. A. J. Org. Chem. 2009, 74, 1388.
116
Chen, Y.-H.; Sun, Mai; Knochel, P. Angew. Chem. Int. Ed. 2009, 48, 2236-2239.
110
120
with various aryl halides and triflates to yield diarylmethanes.117 Besides palladium, it
recently has been shown that iron (II) is an effective catalyst for the Negishi-type cross
coupling of diaryl zinc reagents with benzyl halides.118 A summary of these recent
methodology developments can be seen in Figure 5-4, below.
o
R
R,4r
rr**
M
CT
0Et
* , H 0 , ° B T>
Pd
^
In"
<T"
Chupak et al (Pfizer) 2009
Chen et al 2009
Pd
Pd
x
t>
Pd.
Nobre and Monteiro 2004
Burns et al 2007
Ril
X
McLaughlin (Merck) 2005
/—B
Ri"
KF,B N
\J?
Molander and Elia 2006
Fe
.Ru
+
[I
>1
Flaherty et al (AstraZeneca) 2005
x
oDG
-cr . 6
Ackermann and Novak 2009
Bedford et al (GSK) 2009
Figure 5-4. Recent advances in the synthesis of diarylmethanes. The diverse
methodologies generally employ palladium as a cross-coupling catalyst, but
recently iron (II) has been utilized as well as ruthenium C-H activation in the
presence of an ortfto-directing group (oDG). Notice the recent interest in this
transformation as evidence by publication dates in bold as well as the significant
industry interest. Those highlighted in blue were developed within the noted
pharmaceutical labs.
Although the use of transmetalation reagents (e.g. boron, tin, zinc, indium, e t c . . )
in palladium-catalyzed carbon-carbon bond formation continues to be a powerful
synthetic technique, these reagents are often synthesized from the corresponding
halides and thus reduce overall atom economy119 for the transformation. As a result, the
Flaherty, A.; Trunkfield, A.; Barton, W. Org. Lett. 2005, 7, 4975-4978.
' Bedford, R. B.; Huwe, M.; Wilkinson, M. C. Chem. Commun. 2009, 600-602.
' For a review, see: Toste, B. M. Ace. Chem. Res. 2002, 35, 695-705.
121
development of selective cross-coupling reactions that avoids a transmetalation event
would be highly desirable.
5.2
Palladium-Enolate Coupling Background & Significance. The first
reported example of palladium-mediated coupling120 between aryl halides and activated
methylene
compounds was in 1984 when Takahashi
and co-workers
coupled
malononitrile with various aryl bromides.121 They reported that using 10 mol % Pd(PPh3)4
with sodium hydride as the base in refluxing THF afforded good to excellent yields (7295%) of the arylated malononitrile (Figure 5-5 A).
The first example of a coupling between a carbonyl-activated methylene was the
intramolecular variant carried out by Ciufolini and Browne in 1987 of a 1,3-diketone
(Figure 5-5 B).122 Similarly, Reissig and co-workers carried out an intramolecular
coupling of a single ketone-activated methylene compound (Figure 5-5 C).123
The first intermolecular examples of Pd-enolate coupling between ketones and
aryl bromides were reported nearly simultaneously by three groups in 1997-98. In 1997,
Hartwig and coworkers124 utilized BINAP as a ligand for palladium and in the presence of
sodium f-butoxide, were able to couple the a-carbon of aryl and alkyl ketones with aryl
bromides (Figure 5-5 F). Similarly, Buchwald and coworkers125 utilized potassium
hexamethyldisilazide or f-butoxide in THF with
1,1'-i)/s(diphenylphosphino)ferrocene
(dppf) and bulkier diarylphosphino derivatives of dppf to carry out similar transformations
between the ketones and aryl bromides (Figure 5-5 E). Shortly thereafter, the group of
Miura was able to use a simple PdCI2 salt with the triphenylphosphine ligand added
For the first example of Ni-mediated coupling, see: Semmelhack, M. F.; Stauffer, R. D.; Rogerson, T. D.;
Tetrahedron Lett. 1973, 14, 4519-4522.
121
Uno, M.; Seto, K.; Takahashi, S. J. Chem. Soc. Chem. Commun. 1984, 932-933.
122
Ciufolini, M. A.; Browne, M. E. Tetrahedron Lett. 1987, 171-174.
123
Khan, F. A.; Czerwonka, R.;Reissig, H-U Synlett 1996, 533-535.
124
Hamann, B. C; Hartwig, J. F. J. Am. Chem. Soc. 1997, 119, 12382.
125
Palucki, M.; Buchwald, S. L. J. Am. Chem. Soc. 1997, 119, 11108.
122
(Figure 5-5 D).126 Unlike the research groups of Buchwald and Hartwig, this laboratory
utilized weaker carbonate bases in DMF at elevated temperatures (130 °C) to carry out
the coupling in 1998, though their only reported example was between bromobenzene
and deoxybenzoin.
xr.
THF, NaH
10% Pd(PPh3)4
N C ^ ^ C N reflux, 4 hr.
CN
72-95%
Uno, M.; Seto, K.; Takahashi, S. J. Chem. Soc. Chem.
Commun. 1984,932-933
20% Pd(PPh3)4 Me02C.
t-BuOK, THF
»42%
Ph
Br F. A.; Czerwonka, R.;Reissig, H-U Synlett 1996, 533-535
Khan,
7.5-10% Pd(dba)2
R ' ^ A Ar
R' - H alkvl
y
'
DMF, NaH
10%Pd(PPh3)4
130°C, 6hr.
CN
DPPF and DPPF derivatives
KN(SiMe3)2 or NaOt-Bu
, ,
Refluxing THF, 0.75-5 h
FT
47-94%
Palucki, M.; Buchwald, S. L. J. Am. Chem. Soc. 1997, 119,11108
»•
68%
Ciufolini, M. A.; Browne, M. E. Tetrahedron Lett. 1987, 171-174
O
PdCI2,5%,
Ph v
Br
0
PPh3,10-20%
6
+
p h
^ A
p h
CssCOsOrKsCOs"
DMF, 130 °C, 2hr
f
L
96% 71,
(LC2239-2246
yield)
Bull. Chem. Soc. Jpn. 1998,
Br T.; Inoh, J.-l.;0 Kawamura, Y.; Miura, M.; Nomura, M.
Satoh,
U
1-5%Pd2(dba)3 R'.
D,
D„
+ H N / V H
3.6%BINAP
R' = H, alkyl
NaOt-Bu, THF,
R" = alkyl, aryl 70=0, 4-12 h
63-93%
&>
Hamann, B. C ; Hartwig, J. F. J. Am. Chem. Soc. 1997, 119,
12382
Figure 5-5. First examples of palladium catalyzed arylation of activated
methylene compounds.
Over the past decade, this transformation has been examined in further detail
and numerous improvements have been made in this palladium catalyzed coupling
reaction.127 However, there have been no examples that avoid the use of phosphine or
A/-heterocyclic carbene (NHC) ligands to affect this transformation, and none that
examine the feasibility of this transformation in aqueous media. Thus, the a-arylation of
ketones using simple ligandless palladium catalysts was examined. An aqueous media
utilizing TBAB as a phase transfer catalyst and as a colloidal palladium stabilizing agent
was selected for the initial optimization conditions. The Leadbeater research group has
Satoh, T.; Inoh, J.-l.; Kawamura, Y.; Miura, M.; Nomura, M. Bull. Chem. Soc. Jpn. 1998, 71, 2239-2246.
For excellent reviews, see: (a) Johansson, C. C. C; Colacot, T. J. Angew. Chem. Int. Ed. 2010, 49, 676707. (b) Culkin, D. A.; Hartwig, J. F. Ace. Chem. Res. 2003, 36, 234. For representative examples, see: (c)
Viciu, M. S.; Germaneau, R. F.; Nolan, S. P. Org. Lett. 2002, 4, 4053, (d) Grasa, G. A.; Colacot, T. J. Org.
Lett. 2007, 9, 5489, (e) Hama, T; Hartwig, J. F. Org. Lett. 2008, 10, 1545, (f) Ruan, J.; Saidi, O.; Iggo, J. A.;
Xiao, J. J. Am. Chem. Soc. 2008, 130, 10510, (g) Ehrentraut, A.; Zapf, A.; Beller, M. Adv. Synth. Catal.
2002, 344, 209-217.
127
123
had considerable success utilizing similar conditions to carry out Suzuki-Miyaura and
Heck coupling reactions in water, thus these conditions seemed a logical starting point.
5.3 Colloidal Palladium Background. Ligands have two main roles in palladium
catalysis. First, these bulky, electron-rich ligands will facilitate in the oxidative insertion
step, which is usually rate-limiting. The second role is the ligands' ability to stabilize the
palladium and keep it homogeneous. Often, without the use of bulky organic ligands,
palladium will "black out" of solution. This finely divided palladium (0) metal has limited
catalytic activity when compared to homogeneous catalysis due to the diminished
surface area of the catalyst. Similar to bulky organic ligands, TBAB has a stabilizing
effect, generating a highly active "colloidal palladium" which is in equilibrium with the less
reactive palladium black species.128
simple palladium salt
precatalysts
Br
low TOF
r
o
TBAB
dissoulution
nucleation,
precipitation
Pd"- R insoluble palladium black
low catalytic activity
high TOF
colloidal palladium
high catalytic activity
favored by HIGH Pd loading
Br
p
,„
f|^f'
favored by LOW Pd loading
Figure 5-6. The highly active colloidal palladium-TBAB species is in equilibrium
with insoluble palladium metal precipitate. It has been shown that the turnover
frequency (TOF) is much higher when employing low levels of palladium catalyst
in these systems. Thus, the reaction is often more effective when utilizing lower
palladium loadings.
The amount of palladium black can be minimized, however, and it seems to be
the case (as with all transitions from soluble substrates to insoluble) that there needs to
be a nucleation point to begin the process of palladium insolubility. As such, in cases
Reetz, M. T.; de Vries, J. G. Chem. Commun. 2004,1559-1563.
124
where soluble colloidal palladium is the active catalyst, decreasing palladium loading can
actually increase the catalytic activity.129 In conjunction with TBAB, low levels of
palladium remain suspended in solution in a wide variety of solvents, including water,
NMP, DMF, ethyl acetate, etc. Of course, this inverse loading relationship holds true only
to a point, but it appears that the optimum loading of TBAB is approximately 0.25-1.0 M
which can stabilize up to 0.01 molar equivalents (1 mol %) of the palladium salt.130,131
This phenomenon is illustrated in Figure 5-7, below. The two palladium-catalyzed
reactions were identical in all aspects except the reaction on the right utilized twice that
of the reaction on the left (2 mol % vs. 1 mol %). Additionally, the reaction using less
catalyst was more efficient (82% vs. 50% completion).
Figure 5-7. Side-by-side comparison of two reactions in NMP utilizing TBAB as a
phase-transfer reagent and Pd(OAc) 2 to effect the reaction. Both reactions were
0.5 M in substrate, 0.5 M in TBAB and were allowed to react for 20 minutes at
120 °C. The left reaction utilized a 1 mol % loading of the palladium catalyst while
the reaction on the right utilized 2 mol %. The reaction on the left was - 8 2 %
complete by crude H NMR, while the one on the right showed ~ 50%
conversion.
129
(a) de Vries, A. H. M.; Mulders, J. M. C. A.; Mommers, J. H. M.; Henderickx, H. J. W.; de Vries, J. G. Org.
Lett. 2003, 5, 3285-3288. (b) Slagt, V. F; de Vries, A. H. M.; de Vries, J. G.; Kellogg, R. M. Org. Proc. Res.
Dev. 2010, 14, 30-47.
130
Qualitative observations by the author. No comprehensive screening has ever been undertaken in order
to discover the 'optimal' palladium-TBAB ratio under aqueous reaction conditions. For a discussion of this
effect in NMP, see: Reetz, M. T.; de Vries, J. G. Chem. Commun. 2004,1559-1563.
131
For the first reported examples utilizing tetra alkyl ammonium salts with palladium in an aqueous
environment, see: (a) Jeffery, T. Chem. Commun. 1984, 1287-1289. (b) Jeffery, T. in Advances in MetalOrganic Chemistry, ed. L. S. Liebeskind, Jai Press Inc., London, 1996, Vol. 5, pp. 153-260.
125
5.4 Diarylmethane Synthesis: Initial Discovery & Screening. Initial studies
aimed to investigate the feasibility of coupling acetophenone with two equivalents of 4bromoanisole to form a,a-bis-(4-methoxyphenyl)acetophenone 5.2. The reactions were
performed under phase-transfer conditions in an aqueous sodium hydroxide-TBAB
emulsion. A commercially available palladium ICP standard solution as a convenient
ligandless palladium source (0.1 mol %) was employed to ensure precise palladium
loading on the small scale. Initially, these reactions were performed in the microwave at
high temperatures. The first reaction attempted gave a promising result, yielding the
desired diarylated product in 73% isolated yield after column chromatography. As can be
seen in Table 5-1, an overall trend developed where reducing the reaction temperature
while extending the time led increase yields. Thus, an optimized set of reaction
conditions utilizing of 0.025 mol % Pd, 100 °C for 20 hours was developed which
provided a 95% isolated yield of the diarylated product (Table 5-1, Entry 7). Although the
reaction progresses at even lower catalyst loadings (Table 5-1, Entry 8) it was deemed
unreasonably sluggish.
0
O
C^ **&
r
2 equivalents
3AJU
5.2
Entry
1
2
3
4
5
6
7
8
Time (hours)
.33
.5
1
.5
7
20
20
20
Temp. (°C)
150
150
150
130
110
100
100
100
.OCH3
r^V
OCH 3
Pd (mol %)
0.1
0.1
0.1
0.05
0.05
0.05
0.025
0.005
Yield (%)
73
63
64
81
90
95
95
58/28/14
Table 5-1. Optimization of conditions for the diarylation of acetophenone with 4bromoanisole. Conditions: 1 mmol acetophenone, 2.1 mmol 4-bromoanisole, 0.5
mmol TBAB, 2.0 ml 2.0 M aqueous NaOH. Entries 1-5 performed using
microwave heating, 6-8 utilizing an oil bath. Palladium loadings are based upon
4-bromoanisole. All yields are isolated yields after column chromatography on
silica gel with the exception of entry 8 (crude conversion deduced via 1H NMR).
126
After developing conditions for the diarylation of acetophenone, a short screen of
conditions was probed for conditions that might be amenable for the monoarvlation of
acetophenone to yield a-(4-methoxyphenyl)acetophenone 5.1, though this is not a trivial
notion (Table 5-2). The major obstacle is that the enolate coupling is closely tied to the
acidity of the a-proton on acetophenone. Because the monoarylated intermediate 5.1
contains a proton that is significantly more acidic than the starting acetophenone (pKa
acetophenone = 24.7, 132 pKa deoxybenzoin = 17.7,133), trapping an intermediate that is
more susceptible to enolization—and thus Pd-enolate coupling—proves difficult (Figure
5-8). Indeed, a few reactions were all that were needed to show that there would be little
hope of optimizing for the mono-arylation product beyond the expected statistical
distribution.
The best result was a 1/0.38 ratio (45% regiomeric excess) of
monoarylated/diarylated product when utilizing a 3-fold excess of acetophenone (Table
5-2, Entries 3-5).
Table 5-2. Attempts to afford the mono-arylated intermediate 5.1 by reacting 4bromoanisole with deoxybenzoin were met with a statistical distribution of 4bromoanisole starting material and products 5.1 and 5.2. Excess 4-bromoanisole
afforded the diarylated product in 63% isolated yield after chromatography (Entry
8).
Ace. Chem. Res. 1988, 21, 456, 463.
Can. J. Chem. 1990, 68, 1714.
127
relatively slow f
Br
PdL
T
relatively fast
II base
p
dL
base
base
0"
/^/t^.
pKa=17.7
^•^
pKa=24.7
Strategies to intercept:
-bulky ligands to palladium
•bulky bases
•multiple equivalents of starting ketone
Figure 5-8. Schematic illustrating the difficulties faced in the palladium-catalyzed
synthesis of the mono-arylated variant. The overall reactivity hinges on the
acidity of the a-proton and after the first arylation thus making the intermediate
far more reactive than the starting material was. Three common strategies are
employed to counteract the inherent reactivity differences: the use of bulky
bases, the use of bulky ligands, and utilizing multiple equivalents of the starting
carbonyl compound.
While others have painstakingly developed acceptable conditions to halt the
reaction at the mono-arylated stage, a common factor in developing mono-arylation
conditions for methyl ketones is in the utilization of multiple equivalents of the
appropriate coupling partner, swamping the system with starting material and favoring
the desired mono-arylated product.134 Even then, yields of the desired product in these
cases rarely exceed 80-90%, with the remainder of the aryl halide consumed to afford
the diarylated product. Furthermore, while the bulk of previous research has focused oh
utilizing
bulky
phosphine
ligands or bulky N-heterocyclic
carbene
ligands135 in
conjunction with bulky bases to promote this reaction, the use of colloidal palladium and
aqueous hydroxide base eliminates the requisite steric aspects that may help suppress
the second arylation.
134
Desai, L. V.; Ren, D. T.; Rosner, T. Org. Lett. 2010 DOI:10.1021/ol1000318.
For a review, see: (a) Johansson, C. C. C; Colacot, T. J. Angew. Chem. Int. Ed. 2010, 49, 676-707. (b)
Culkin, D. A.; Hartwig, J. F. Ace. Chem. Res. 2003, 36, 234.
135
128
For this reason, attention was re-focused on the diarylation and a small screen of
three additional substrates to probe electronic effects was undertaken. However, while
the previously optimized conditions worked well for the electron-rich 4-bromoanisole, the
conversion to the desired diarylated products dropped when electron neutral or poor
systems were
used. Of
interest,
however, was the observation of
symmetric
diarylmethanes as by-products, and in the case of 1-bromo-4-fluorobenzene,
a
significant amount of 4,4'-difluorodiphenylmethane was observed (Figure 5-9).
Figure 5-9. A small screen with electronically neutral or slightly electron-poor
substrates quickly revealed the limited scope of this methodology. Only electronrich substrates would afford high conversions.
It stands to reason that the inductively electron-poor system with the a,a-bis-(4fluorophenyl)acetophenone is better able to stabilize the carbanionic diarylmethane
leaving group. Gratifyingly, this observation helped to explain why yields suffered initially
when the reaction between acetophenone and 4-bromoanisole was performed at higher
129
temperatures (Table 5-1, entries 1-4). While there are limited examples of isolating
diarylmethanes from the corresponding ketone intermediates for characterization
purposes,136 no systematic approach has been exploited that employs this methodology
in order to synthesize a range of substituted diarylmethanes. This mechanistically
interesting route to diarylmethanes warranted further exploration.
The replacement of acetophenone with deoxybenzoin as the ketone-coupling
partner allowed non-symmetric arylations, where the deoxybenzoin can effectively be
thought of as a benzylating reagent. Because the coupling for the electron-rich 4bromoanisole should be the most demanding, it was again employed in order to develop
optimal conditions for generating diarylmethanes. Suitable conditions were quickly
developed, a 1:1 ratio of deoxybenzoin: aryl bromide at the 1.0 mmol scale, 2.0 ml of a
3.0 M aqueous solution of NaOH, 0.5 mmol TBAB generated the desired 4methoxydiphenyl 5.3 in 9 1 % isolated yield after chromatography. Palladium was loaded
at 0.1 mol %, utilizing a commercially available palladium standard solution in 5%
aqueous HCI.137 The higher concentration of hydroxide was utilized in order to drive the
reaction to completion, as an equivalent of benzoic acid is also generated over the
course of the reaction. The optimal conditions involved a two-stage process: heating at
130 °C for 30 minutes, followed by heating at 160 °C for 30 minutes.
Further increases in temperature led to diminished yields. Similarly, simply
increasing the reaction time was of no benefit, possibly due to the limited lifetime of the
catalyst under these harsh reaction conditions. This was further evidenced by the fact
that more of the deoxybenzoin and 4-bromoanisole starting materials were evident in
crude 1H NMR at 160 °C than when the reaction was run at 150 °C. Two-stage reaction
1Jb
(a) Lagrave, R. Ann. Chimie 1927, 8, 363, (b) Koelsch, C. F. J. Am. Chem. Soc. 1931, 53, 1147
-1000 ppm (1 mg/ml), available from Aldrich (207349). This is a convenient method to deliver low
palladium loadings. -106 uL is delivered via pipette for a 1 mmol scale reaction at 0.1 mol % catalyst
loadings (compare the 0.17 mg of PdCb that would be required if the catalyst was delivered in solid form).
137
130
conditions allowed for the bulk of the palladium-mediated coupling to occur at 130 °C, at
which
point the temperature
was
increased to expedite the formation of the
diarylmethane.
Application of these optimized conditions to a range of aryl bromides afforded the
corresponding diarylmethanes in good to excellent yields for 3- and 4-substitued aryl
bromides
(Figure
5-10).
Furthermore, the
use
of
(3-bromostyrene
showed
the
applicability toward using this methodology to couple vinyl halide, though 2.4 equivalents
of the styrene had to be utilized to consume all of the deoxybenzoin starting material,
likely due to the decomposition of p-bromostyrene to phenylacetylene.
5.12 78%
5.13 84%
5.14 99%
5.15 89%
Figure 5-10. Screen of various aryl bromides coupled with deoxybenzoin to
afford substituted diarylmethanes. Yields were generally high, but orthosubstituted aryl bromides proved generally unreactive with the exception of 1bromonaphthalene. Additionally, no product was observed when utilizing the very
electron-rich 4-bromoaniline under these conditions. Conditions: 1 mmol
deoxybenzoin, 1 mmol aryl halide (2.4 for 3-bromostyrene), 0.5 mmol TBAB, 2.0
ml 3.0 M aqueous NaOH, 0.1 mol % palladium, heat to 130 °C for 30 minutes,
then to 160 °C for 30 minutes.
Notable exceptions (Figure 5-11) to this coupling included the very electron-rich
4-bromoaniline, where no coupling was observed and only starting materials were
recovered. A second limitation to this methodology is the very limited tolerance of orthosubstituted aryl bromides, and only when the C-Br bond is very activated as in 1-
131
bromonaphthalene, could successful coupling be realized (Figure 5-10, 5.11). In the
coupling with 2-bromotoluene, only a trace of the arylated intermediate and no
diarylmethane was observed after 60 minutes. The attempted reaction with 2bromoanisole returned only starting materials.
Br
0.1 mol % PdCI2
^
^
^
fTV^r^R
20/NaOH/TBAB
\^)
\^>
MW, 60-120 min.
130-160 °C
not observed:
H
Figure 5-11. Under the colloidal palladium reaction conditions, both sterics and
electronics can inhibit the initial arylation of deoxybenzoin. 4-bromoaniline, 2bromotoluene, and 2-bromoanisole were all essentially unreactive under the
developed conditions.
Next, a series of compounds was prepared to yield substitution on both rings of
the diarylmethanes. Immediately, the attenuated reactivity of a slightly electron rich
deoxybenzoin versus a slightly electron-poor one can be seen in the two orthogonal
syntheses
of
5.16
(Table
5-3,
entries
1
and
2).
Reacting
a-(4'-
chlorophenyl)acetophenone with 4-bromotoluene (Table 5-3, Entry 1) afforded 5.16 in a
95% isolated yield, but with reversal of the substitution pattern (Table 5-3, Entry 2)
reduces the reactivity and a modest 50% yield of 5.16 was realized. An even more
pronounced
effect
is
seen
in
the
two
attempted
syntheses
of
5.17.
After
chromatography, a moderate yield of 5.17 (82%) was isolated when reacting a-(4'chlorophenyl)acetophenone with 4-bromoanisole. However, the crude 1H NMR of the
reaction between a-(4'-methoxyphenyl)acetophenone and 4-chlorobromobenze showed
poor conversion (25%, Table 5-3, fourth entry) of the starting materials even after 2
hours. This non-symmetric property of equality illustrates that consideration must be paid
to the electronics of the system in order to optimize the yield of the desired product.
Additionally, it sheds some light on the mechanism of this reaction (Figure 5-12).
132
0
0^" • * 0
o
Product
5.16
5.16
5.17
5.17
5.18
5.19
5.20
5.21
5.22
<R1
Ri
CH 3
CI
CI
OCH-,
CH 3
CI
CI
CH,
OCHs
0.1 mol % PdCI2
H20/NaOH/TBAB
MW, 120 min.
130-160 °C
R2
4-CI
4-CH,
4-OCH 3
4-CI
4-OCH3
4-F
3-CI '
4-Ph
3,5-(OCH3)?
^
Rl
5.16-5.22
Yield (%)
50
95
82
(15/10/75)
83
80
62
87
(-/10/90)
Table 5-3. The palladium-catalyzed coupling of various aryl bromides with 4'substitued deoxybenzoins was undertaken in order to develop a qualitative
appreciation of electronic effects on the reactivity. The differences are best
appreciated in the formation of 5.16 and 5.17 where striking differences in
diarylmethane yield were seen. Indeed, when the electron-rich a-(4'-methoxyphenyl)acetophenone was used (5.17 and 5.22), conversion was so low that the
products were not isolated.
5.5 Putative Reaction Mechanism (Qualitative Observations). From the initial
screening studies, a qualitative appreciation was developed regarding some mechanistic
aspects of this transformation (Figure 5-12) and a few assertions can be made. First,
from the orthogonal syntheses of 5.16 and 5.17, it can be seen that the acidity of the ahydrogen likely plays a role in the overall rate of the reaction (Figure 5-12, /(1): when the
chlorine atom was incorporated into the deoxybenzoin, the reaction was significantly
faster. Similarly, the electron-withdrawing ability of the chlorine could be implicated in the
stabilization of the transition state (Figure 5-12, 9) during the dissociation of the
intermediate to form the diarylmethane and benzoate anion (Figure 5-12,
k8).
Interestingly, an electron-withdrawing group also likely would stabilize the nonproductive intermediate (Figure 5-12, 7a, k6), and perhaps inhibit the formation of
diarylmethane. Electron-withdrawing groups are known to facilitate the oxidative
insertion of palladium into the carbon-halogen bond of aryl halides. Thus, an electronwithdrawing group on the aryl bromide might increase the overall rate of the reaction
(Figure 5-12, k2) if the oxidative insertion is the rate-limiting step. Generally, the oxidative
133
insertion is the rate-limiting step, though not always. 1 3 8 Furthermore, this elaborate
transformation may have two transition states with very close activation energies.
R'
plausibleTS 9
Figure 5-12. A plausible reaction mechanism can be developed from the
qualitative observations from the screening reactions. The mechanism can be
split into two main components: one that requires palladium and one where
palladium is not involved in the mechanism. Depending on the nature of
substitution, significant reactivity differences were noticed when screening
substrates. When R or R' = EWG, it can be hypothesized that the individual rate
constants, k^ k2, and k8, would be activated. Conversely, an EWG may attenuate
the reaction somewhat due to the stabilization of the non-mechanisticallyproductive intermediate 7a.
Mathew, J. S.; Klussmann, M.; Iwamura, H.; Valera, F.; Futran, A.; Emanuelsson, E. A. C; Blackmond,
D. G. J. Org. Chem. 2006, 71, 4711-4722.
134
5.6
Initial
Diarylmethane
Kinetic
Studies.
In order to gain
additional
mechanistic insight and to explore this methodology further, a series of kinetic studies
were devised. The development and implementation of a robust approach was devised
after a period of difficulty. The first hurdle that had to be cleared was choosing the best
method to evaluate the reaction's progress and a number of options were considered. A
commercially available aftermarket module that interfaces with any HPLC system called
the iChem Explorer was considered.139 The automated sampling and the ability to
simultaneously run multiple reactions sounded promising, as one of the most important
aspects to successful kinetic studies is multiple trials and data points. However, after
working with the manufacturers and this kit for 2-3 weeks, it was determined that this
would not be a viable approach. The biphasic nature of the reaction requires vigorous
and reproducible stirring from sample to sample and the integrated stir bar was unable to
provide this essential aspect. Furthermore, when sampling 1-5 uL of the emulsion to
characterize the reaction, there was little guarantee that the sample would draw from a
"greasy" portion where the reactants, intermediates, and products lay, or simply draw 5
uL of a sodium hydroxide solution. As such, the relative intensities from sample to
sample varied by 1-2 orders of magnitude and few kinetic conclusions could be drawn
with any certainty.
Next, the use of GC was considered, as it would be straightforward to set up
calibration curves for known concentrations of starting materials, intermediate, and
product. The biggest drawback to this approach, however, was the considerable time
element that would be required to employ GC. The most efficient set of conditions
developed required a 35-minute run time to elute all components, not including cooldown time. Thus, each reaction with its 10 aliquots would require approximately 6
person-hours in front of a gas chromatograph. Furthermore, each aliquot would require
139
Kedia, S. B.; Mitchell, M. B. Org. Proc. Res. Dev. 2009, 13, 420-428.
135
an extensive work up to remove all palladium, all TBAB, and all water, as each of these
components severely shorten the lifetime of a GC column.
Thus, 1H NMR was selected and proved an ideal match to monitor this reaction.
The aliquots required only a short work-up and an experienced NMR spectroscopist can
take 10 proton NMRs in under 30 minutes. One half an hour of spectroscopy time per
reaction is much more reasonable than the 6 hours required of GC. Furthermore, the
starting material, intermediate, and product each contained a methylene or methine
proton signal that was spin isolated with a unique chemical shift, i.e. all were clean
singlets. After verification that NMR could be utilized to quantify the mole fraction of each
component in the reaction, it was decided that this would be the way forward to
quantitatively monitor reaction progress.
,
F=™
1
,
pn
8.0
1
7.8
,
1
7.6
,—-,
7.4
r
1
72
1
r
7.0
,
,
6.8
1
1
6.6
,
1
6.4
r—,
62
i
P—|
6.0
,
1
5.8
1
|
5.6
i
|
5.4
.
|
52
,
1
5.0
,
,
4.8
1
1
4.6
,
1
4.4
I,
1
i
42
I|
4.0
i
|
r
3.8
Figure 5-13. Typical 1H NMR spectrum generated when monitoring the formation
of diarylmethanes. Each compound exhibits an isolated singlet and the
comparative integration values are used to plot the reaction's progress (Figure 514, below).
There were a number of other difficulties encountered when developing the
protocol to accurately monitor the progress of this reaction. First, there was an induction
136
period for the formation of the active palladium colloid. Furthermore, the palladium in this
form and at these relatively low temperatures has a limited stability. Once palladium
black begins to precipitate, the reaction's progress soon ceases. Finally, this reaction is
a complex, multistep process and poses significant potential pitfalls for the kineticist.
That said, enough information was generated to glean some kinetic insight from the
data.
The formation of diarylmethane from deoxybenzoin and bromobenzene initially
was monitored at 96 °C (Figure 5-14). Right away, the use of microwave heating under
sealed vessel conditions with high temperatures can be appreciated. During the initial
methodology development, the yields of diarylmethanes were high (Figure 5-10), the
reaction was complete in 60 minutes, and only required 0.1 mol % of the PdCI2 catalyst.
During these kinetic studies, however, the reaction conditions were limited to openvessel (1 atm) conditions, and subsequently lower reaction temperatures. At 0.1 mol %
palladium loading, the reaction is sluggish, taking 6 hours to reach 20.8% conversion
and only reaching 56.7% conversion after 22.5 hours at 96 °C. Furthermore, the
palladium had been deactivated and precipitated out of solution which was evidenced by
the palladium black. No further conversion occurred.
As such, the palladium loading was increased to 1 mol %. This catalyst loading
sufficiently increased the rate and the transformation could be monitored over a
reasonable timeframe, as can be seen in Figure 5-14. The reaction was monitored at a
range of temperatures 140 and a plot of In [deoxybenzoin] versus time yielded nearly
straight lines, which indicates that it is a first order reaction (or very nearly first order),
where the slope of the line is equal to the negative rate constant (m=-k).
Ideally, it is best to monitor a reaction across as wide a range of temperatures as possible. However, due
to the dynamics of the reaction, the temperature range was limited to only 10 degrees: the boiling point of
water limited the upper range, and below 88 °C the TBAB (mp 102 °C) was slow to melt/dissolve into the 3.0
M NaOH solution, which impeded the reaction more than expected due to temperature alone.
137
3.0 M NaOH (aq)
*Br
0.5 eq. TBAB
1 mol % PdCI2
96 °C
5.4
Catalyst pre-activation
Obeys 1 st order kinetics
J.
I
Catalyst deactivation
100
200
300
400
500
600
time (min)
—•—diarylmethane --•— deoxybenzoin —*— intermediate
Figure 5-14. Typical plot of mole fraction of substrate versus time for the
formation of diphenylmethane 5.4. At the outset of the reaction (shown in purple),
there is an introductory catalyst activation period to generate the colloidal
palladium. Shown in green, the reaction is nearly first-order as evidenced from
Figure 5-15, below.
3.0 M NaOH (aq)
0.5 eq. TBAB
1 mol % PdCI2
5.4
89-99 °C
-0.2
-0.4
g -0.6 o -0.8
O)
•o
"^ - 1 . H
-1.2 -|
-1.4
50
100
150
200
250
300
time (min)
Figure 5-15. Plot of the natural log of deoxybenzoin concentration versus time in
the linear region for selected temperatures. The slope of the lines is equal to the
negative rate constant (m=-k).
138
Next, a plot of the calculated values of the rate constants versus 1/T was plotted
to yield a straight line (Figure 5-16). Using the slope = -Ea/R, the activation energy was
calculated to be 120 kJ/mol for the palladium catalyzed formation of diphenylmethane
from bromobenzene and deoxybenzoin.
-4
-4.4 Ea
-4.8
+ 34.226
y = -14440x
R2 = 0.992
= 120 kJ/mol, 28 kcal/mol
*v*
JC
c
-5.2 -
#\
-5.6 -b
0.00266
r
'
0.0027
0.00274
'
0.00278
1/T
Figure 5-16. Plot of the In k versus 1/T yields a straight line, y=mx + b, where m
- -Ea/R, thus Ea = 120 kJ/mol (28 kcal/mol) for the reaction.
The next step for this project is to screen four or five aryl bromides in order to
determine krel values as the substrates change electronically. The substrates need to
cover a range of a values and should be selected accordingly. Table 5-4 below lists a
range of aryl bromides currently available in the Leadbeater research group and their
respective sigma values. The determination of relative values for the rate constants (kx)
should be straightforward. Two trials for each substrate at 93 °C can be compared to the
rate constant determined for bromobenzene at 93 °C (k0). Next, a plot of In {kxlk0) versus
the sigma values for each substrate will yield a trend, called the reaction
constant,
sensitivity constant, or Rho (p). A positive p indicates that negative charge is building on
the transition state of the rate-determining step where a negative p value indicates that
positive charge is building.
139
Table 5-4. A range of a values should be screened to determine relative rate
constants and electronic impacts for the formation of diarylmethanes.
Next, to gain a qualitative appreciation for the dynamics of the second step of the
reaction, preliminary studies have been undertaken to determine the relative rates for
the hydroxide-catalyzed decomposition of the diarylated intermediate to form the
diarylmethane and benzoate anion (Figure 5-17). To accomplish this, the arylated
intermediates were synthesized and isolated.141
These substrates were subjected to the initial reaction conditions, and the
conversion of the intermediate starting material to form substituted diarylmethanes was
monitored. To accomplish this, a solution of 3.0 M NaOH (aq) and 0.5 molar equivalents
TBAB was brought to the desired temperature (95.5 °C). At this point, the solid
diarylated intermediate was introduced to the reaction flask in one portion and the
conversion to the substituted diarylmethane was monitored by 1H NMR spectroscopy.
From Figure 5-17, trends can be appreciated, and the reaction rate increases following
the trend: 4-OCH 3 5.23 < 4-CH 3 5.24 < 3-CH 3 5.25. Unfortunately, the unsubstituted a,adiphenylacetophenone 5.22 is not well-behaved in this system and other aspects of this
substrate have a greater impact on the reactivity than do electronic effects. The melting
141
For reaction conditions, see Appendix 5.
140
point of 5.22 is 135 °C, but all other substrates have a melting point lower than the
reaction temperature. Because the solution is an emulsion, the reactivity of the a,adiphenylacetophenone 5.22 is greatly hindered initially, as it is slow to melt. However,
after the slow start, the formation of some diphenylmethane renders the system a bit
"greasier," the a,a-diphenylacetophenone 5.22 melts completely, at which point the rate
increases. The plot of 1/[a,a-diarylacetophenone] versus time (Figure 5-17, right) yields
straight lines in all cases except for the decomposition of a,a-diphenylacetophenone
5.22, indicating second-order kinetics.
0
100
time (min)
200
0
100
time (min)
200
—•—-H —•—4-methyl —»—3-methyl —*—4-methoxy —•—-H —•—4-methyl —»—3-methyl —*—4-methoxy
Figure 5-17. Left: Plot of conversion to diarylmethane versus time for the
diarylated intermediates 5.22-5.25. Right: The non-ideal behavior of 5.22 is
evident in the plot of 1/[s.m.] versus time (bold blue line). All other substrates
yield straight lines, indicating second-order reaction kinetics. The high melting
point of 5.22 initially hinders the reaction.
In the future, one of these substrates should be selected, e.g. 5.24, in order to
determine an activation energy for the second step of the reaction. Qualitatively, it
appears that the activation energy for the second step is lower than the calculated 120
kJ/mol for the complete reaction, indicating that the formation of the diarylated
intermediate is the rate-limiting step for the overall reaction. However, it would more
thorough to put a quantitative value here. At this point, a determination of kre, for various
141
substitution on the acetophenone ring for three or four substrates to determine electronic
effects on this ring in the overall rate of the reaction would complete the kinetic studies
project (Figure 5-18).
Figure 5-18. Future work to determine the impact of acetophenone ring
substitution on the overall rate of reaction should be undertaken. The hypothesis
would be that electron-withdrawing groups would facilitate the reaction, as it may
lead to the increased acidity of the a-proton ad a higher effective concentration of
the enolate.
5.7 Conclusions & Future Outlook. Besides the continued studies into the
kinetics of the diarylmethane formation, the application of this coupling protocol to
heteroaryl bromides would yield interesting small molecules and warrants further
investigation. Additionally, Appendix 5 discusses subjecting the diarylated intermediate
to
conditions
that
include
an
atmosphere
of
oxygen
to
yield
unsymmetric
benzophenones. These future potential applications of the palladium-catalyzed arylation
of deoxybenzoin are summarized in Figure 5-19.
Figure 5-19. Future applications of the Pd-catalyzed arylations of deoxybenzoin.
142
Appendix 1 •
Detailed Raman Signal Strength-Concentration Correlations
For the organic chemist, a good first order approximation is that Raman signal
intensity is proportional to substrate concentration. Really, however, it is a bit more
complex. Below is an outline of each component to Raman signal strength, a definition
of those aspects, and an indication which of the components can be ignored by the
organic chemist and why as well as a simplified equation that should be suitable for the
organic chemist in most situations.
Raman Signal Strength (S)
S = P0 P D K CI T Q ts, where;
P 0 = laser intensity (photons»sec"1)
Laser intensity should be fairly straightforward and can be appreciated by the
wattage indicated on the laser, e.g. 500 W, which can be related to photons per second.
For the organic chemist, this is important: higher-powered laser leads to stronger
signal strength or the ability to use shorter acquisition times.
p = differential Raman cross section (cm2,molecule»sr"1)
This particular aspect is mired in physical chemistry but concerns a molecule's
changes in Raman activity as it stretches, bends, and vibrates. The Raman signal (precalibration) is Gaussian in nature and signal strength is actually area under the curve,
not absolute peak intensity. However, nearly all commercially available
Raman
manipulation software automatically calibrates this utilizing known Raman cross-sections
that are widely agreed upon by the analytical chemistry community. Indeed, one of the
first steps while setting up the Raman spectrometer is to place a vial of acetonitrile
(cyclohexane is common, too) and calibrate the Raman signal to the known standard.
Secondly, p accounts for the fact that a molecule will be more or less likely to be
sympathetic with the incident irradiation and begin vibrating depending on its orientation
to the incident laser light. This can be exploited to determine crystal structure or polymer
configuration, i.e. a crystal will give a different Raman signal when looking down the
144
(100) face compared to the (010) or the (001) faces. However, in solution chemistry, this
can be completely ignored as at any given moment in time the distribution orientation of
the billions and billions of molecules will remain constant.
For the organic chemist, this has few ramifications in the day-to-day utilization of
spectroscopy to monitor reactions is solution.
D = number density (molecules#cm"3)
This should be straightforward: number density = concentration. Obviously, this is
extremely important for the organic chemist. Indeed, the Raman spectrometer can be
used to measure change in concentration assuming all other parameters are held
constant.
K = 3rd Dimension correction factor (cm)
In absorption spectroscopy, this would be the 'path length.' Raman signal
strength is proportional to path length, but in many cases, the path length is actually the
laser's "depth of field," as the flask has exceeded the maximum penetration depth for the
laser.
The practical ramifications of this for the organic chemist is that one needs to
employ the same size reaction flask (ideally, the same exact flask) when using the
Raman spectrometer for kinetic studies or to determine concentration, etc.
Q = solid angle of collection (sr)
Raman spectrometers generally will acquire data either 90° or 180° with respect
to the incident irradiation. This is accounted for within the spectrometer and can for all
practical purposes be ignored by the organic chemist. The spectrometer in the
Leadbeater research group has a 180° solid angle of collection.
T = transmission factor (unitless)
This accounts for signal loss due to filters within the spectrometer and can be
ignored.
145
Q = quantum efficiency of the detector (electrons*photon~1)
This is simply how efficiently the CCD translates detected photons into electrons
(binary data) and sends the data to the PC. Analogous to the number of photons that
must strike the retina before the brain processes 'sight.'
Unimportant to the organic chemist.
ts = measurement time (sec)
This is the 'acquisition' time of the Raman spectrometer.
For the organic chemist, longer acquisition times lead to stronger Raman signal
and better signal to noise ratios.
Signal intensity S is measured in units of electrons (data bits received by the PC) as
illustrated in the units equation, below:
S=
[photonsV
\
cm
U molecules], v J electrons
1— (cm)(sr)
(sec) = electrons
sec /^molecule »sr/\ cm
/
^ photon /
However, there is one more aspect that must be considered, and that is
temperature (this is discussed in Chapter 3 as well). Most Raman spectrometers only
record the Stokes' shift in the Raman spectrum, ignoring the anti-Stokes. This is
because the anti-Stokes' shift is redundant in terms of spectroscopic data, i.e. the
relative shifts are the same for both Stokes' and anti-Stokes'. Thus by ignoring the antiStokes' side of the spectrum the computer can cut in half its calculations (or do twice as
much in the same time).
Because only the Stokes' shift is monitored, the Raman signal strength is
inversely proportional to temperature. As molecules get hotter, there are fewer in the
ground vibrational state and the Stokes' shift signal intensity suffers accordingly.
For the organic chemist, this means that calibration curves must obtained at each
discrete temperature that the reaction is being monitored at if it is desirable to convert
those units of Raman signal intensity to units of concentration.
146
So, for the organic chemist, a more practical formula for signal intensity can be
developed:
S = P0 [substrate] K T 1 fs4*
Where
Po = laser intensity, [substrate] = substrate concentration, K = path length (vessel size),
T = temperature, ts = acquisition time, and M* = a constant.
147
Appendix 2 •
Interface of Raman Spectrograph & Microwave; Pertinent Equipment Data
General Information.
Excitation Source: NIR, frequency stabilized, narrow line width diode laser at 785 nm.
Laser power, 500 mW. Laser power at sample: -200 mW. Line width < 2 cm"1. Fibercoupled laser output (100 urn, 0.22 NA), NA=numerical aperture of the fiber optic cable.
Fiber-optic Probe: Permanently-aligned two single fiber combination 100 urn excitation
fiber, 200 urn collection fiber (0.22 NA). Working distance: 8 mm (standard). Rayleigh
rejection: O.D. > 7 at laser wavelength.
CCD Detector: High sensitivity linear CCD array. Temperature regulated (at 13 °C)
operation for long integration time and stable dark reference subtraction. Pixel Size: 14
urn x 200 urn (2048 Pixels); 16-bit digitization.
Spectrograph: Symmetrical crossed Czerny-Turner design. Resolution: - 1 0 cm"1 at 785
nm. Excitation spectral coverage: 200-2400 cm"1. Acquisition spectral coverage: 2502250 cm"1. Built-in software calibration.
System Software: EZ Raman 3.5.4MAS. Data files exported into Microsoft Excel.
Interfacing the Raman Spectrometer with the Discover Microwave. This
should already be set up. However, in the event that the microwave unit needs to be
replaced, here is a step-by-step procedure detailing how to interface the two units and a
step-by-step procedure outlining the proper sequence to obtain accurate Raman data.
1) Remove the protective spill cup from the Discover.
2) Carefully remove the top of the Discover S-Class to access the side port.
149
3) Remove the side port's Teflon insert by unscrewing the nut holding it in place.
4) Replace the cover to the Discover and insert the modified spill cup with hole,
ensuring that it lines up with the access port on the Discover.
5) On the Raman system, connect the 6" quartz light pipe to the laser by gently
tightening the provided locking nut.
6) Gently slide the light pipe in the Discover S-Class through the access port, ensuring
that the spill cup is still aligned.
7) Once the microwave cavity has been breached, replace the attenuator on the
microwave and insert a 10-ml glass tube.
8) Continue sliding the light pipe into the microwave until it touches the tube, then pull it
back until there is approximately 1-2 mm space between the light pipe and the
microwave tube.
9) Connect the Raman system to the desktop computer using a USB 2.0 cable. Turn
Raman unit on by turning the key to the "ON" position. The must be done before
opening the Enwave software or the software will not recognize the Raman laser and
display an error message.
150
Data Acquisition Procedure
10) Open the Enwave Software (V3.5.4 MAS).
jj&iasl
0b ££&«* igspfee1 &3*gu*a B^e-iSaft- "SW© 3per^ iJefp
!airf glig&fcti.&g J J | JiiJikJ'd
i»&i$
--
**S»
r
jft* |
rrXJUisScate
:
>, aw»?
s? M
—
•1
-1
YJUdsScife
^u*<
Intensity
"Laser
;
H>
2358
-
388 408 SIB ISO 708 S B
990 1809 1108 1280 1 3 » 1408 1500 1090 1708 1890 1900 2WJ 2MB HOB 23TO
Raman SNA (cm-1)
QpatHe;
pSPrcgjsm Flasi£2*a8S5ss¥3t4 liftSSS*
MS*M
a-iaoxp
11) On the main screen, set Integration to the desired acquisition time (in seconds),
Average to 1, and Boxcar to 3.
12) From the Time Chart pull-down tab, select Time Trend.
13) In the "Cycle Time" field enter 4.
14) Click on the Dark Scan icon.
15) At the prompt "Please use same integration, average, and boxcar parameters for
dark subtraction," click OK.
16) At the prompt, "Turn on laser and save a reference spectrum for subtraction," click
Yes. This should be done while the reaction is still in the "pre-stirring" stage in the
Discover microwave.
151
17) Click OK at the. "Dark Spectrum Saved" prompt.
18) Click the Continuous Scan.
19) Click OK at the "Start scans without saving files?" prompt.
20) Click OK at the "Turn on the laser?" prompt. Do this when the Discover has 2-3
seconds remaining in the "pre-stirring" stage. The first scan with subtracted starting
materials will be the t=0 spectrum and is essentially baseline noise.
CAUTION: The laser is very powerful (500mW) and all necessary precautions
should be taken to avoid putting oneself in its path.
21) Now the Raman software is monitoring the reaction's progress.
22) Click the Stop Scan icon.
23) At the prompt "Stop continuous scans?" click Yes.
24) At the prompt, "Save all data as a .txt file in your application directory?" click NO.
25) Click the Save Data icon. At the "Do you want to export all data to Excel?" click
YES. This automatically transfers all of the raw data to Excel in .xls format which will
allow for facile data manipulation in the form of charts and graphs, etc. Click NO at
the prompt, "Do you want to save all RAW data?"
152
Appendix 3 •
Insertion of d10 Metals into TPP
Results and Discussion. Metalloporphyrins find use in a wide array of
applications including light-harvesting systems, chemotheraputics, light-emitting diodes,
and chemical sensors.142 Many preparative routes to metalloporphyrins involve heating
the free porphyrin together with a metal salt in a high-boiling solvent for a prolonged
period of time. Although thermodynamically favorable, metal insertion is kinetically
inhibited due to the relative inflexibility of the porphyrin ligand conformation as well as
the nature of the metal salt used. The scope and limitations of using microwave heating
to facilitate the insertion of d 10 metals into various porphyrins was assessed in an
attempt to reduce overall reaction times.
Reactions were first performed on the unsubstituted tetraphenylporphyrin (TPP)
in order to optimize conditions suitable for rapid insertion of nickel, palladium, and
platinum. Due to the robust nature of TPP, relatively harsh conditions could be employed
with good results. Indeed, utilizing 3 equivalents of Ni(OAc)2*4H20 in pyridine, complete
conversion to the Ni*TPP was realized in only 15 minutes at 180 °C (Table A4-1, Entry
b). These conditions proved ideal in the synthesis of the Pd*TPP analogue, though Pd
(acac)2 was used as the preferred metal salt precursor. When moving to platinum, more
forcing conditions had to be utilized. The solvent was changed from pyridine to
benzonitrile. Quantitative insertion of the metal to afford Pt*TPP was realized in 15
minutes, though the reaction temperature was increased to 250 °C (Table A4-1, Entry i).
Kadish, K.M.; Smith, K.M.; Guilard, R., eds. The Porphyrin Handbook, Academic Press, San Diego 2000.
154
Table A3-1. d10 metal insertion into tetraphenylporphyrin and derivatives. *1 eq.
Ni(OAc)2 was utilized. **Ptl2 and K2PtCI4 were also employed with no conversion.
***Numerous other products (decomposition) were evident by TLC. The attempts
to insert Pd and Pt into the diol resulted only in decomposition, affording none of
the desired product.
Conclusions. The facile access to high temperatures afforded by the scientific
microwave apparatus led to the rapid insertion of d 10 metals into tetraphenylporphyrin as
well as the tetraphenylporpholactone with good yields. However, in the case of the
chlorin-diol, the high temperatures led mainly to decomposition of the starting materials,
and only the nickel insertion provided any of the desired metal-inserted porphyrin.
155
Appendix 4 •
Aza-Michael Synthesis of A/-Aryl p-Amino Acid Derivatives
The preparation of p-amino acids and their derivatives is seeing increasing
research interest due to their applications in medicinal chemistry due to their biological
activity.143,144 They undergo little or no degradation by peptidases and nonpeptidic pamino acids are found in (3-lactam-antibiotics, HIV-protease inhibitors, and enzyme
inhibitors.
The addition of primary and secondary cyclic or acyclic amines to a,punsaturated esters to generate the p-amino ester is facile. However, the use of
substituted anilines to generate A/-aryl substituted p-amino esters is problematic owing to
the decreased nucleophilicity of the nitrogen atom when adjacent to an aromatic system.
Thus, the rapid synthesis of A/-aryl substituted p-amino acids was investigated using
microwave irradiation as a tool to access high temperatures in order to overcome the
high activation energy barrier of this transformation, in order to seek to improve yields,
eliminate the use of expensive145 or toxic146 catalysts as well as to reduce reaction
times. 147 ' 148149
Using 10 mol % of acetic acid as a catalyst, it was found that moderate yields of
the corresponding Michael adduct could be obtained simply by heating aniline with an
equivalent of methyl acrylate in the absence of solvent. A range of times and
temperatures were screened. It was decided the most efficient method was to heat the
reaction to 200 °C and hold for 20 minutes. These conditions repeatedly yielded the
desired mono-adduct in 78-82% yield. Longer reaction times at lower temperatures
could be utilized (Table A5-1, Entry b) with similar results. However, higher temperatures
143
Juaristi, E.; Soloshonok, V. A. Eds Enantioselective Synthesis of P-Amino Acids, Wiley-lnterscience, New
York, 2005.
144
(a) Fulop, F. Chem. Rev. 2001, 101, 2181. (b) Hamuro, Y.; Schneider, J. P.; DeGrado, W. F. J. Am.
Chem. Soc. 1999, 121, 12200. (c) Gademann, K.; Hintermann, T.; Schreiber, J. V. Curr. Med. Chem. 1999,
6, 905-925.
145
(a) Li, K.; Horton, P. N.; Hursthouse, M. B.; Hii, K. K. J. Organomei. Chem. 2003, 665, 250. (b) (a) Li, K.;
Phua, P. H.; Hii, K. K. Tetrahedron 2005, 61, 6237.
146
Barluenga, J.; Villamafia, J.; Yus, M. Synlett. 1981, 375.
147
Southwick, P. L.; Seivard, L. L. J. Am. Chem. Soc. 1949, 71, 2532.
148
Werbel, L. M.; Kesten, S. J.; Turner, W. R. Eur. J. Med. Chem. 1993, 28, 837.
149
Siddiqui, A. A.; Arora, A.; Siddiqui, N.; Misra, A. Ind. J. Chem. B. 2005, 44, 838.
157
(Table A5-1, Entries e & f) led to significant formation of the bis-adduct, and lowered
overall yield of the desired mono-addition product.
N
a "i
Entry
a
b
c
d
e
f
g
h
O
Jk.
10mol%AcOH
Temp (°C)
140
170
170
200
215
230
200
200
^ ^ .
u
Time (min)
30
30
10
10
10
10
20
20
H
O
Conv.
48
82
57
75
75
73
81
23
Table A4-1. Initial optimization of the aza-Michael addition of aniline to methyl
acrylate utilizing 10 mol % acetic acid to facilitate the reaction. The optimum
conditions heated the reaction to 200 °C for 20 minutes. Higher temperatures
and/or longer times led to significant formation of the bis-adduct. For entry h, the
acetic acid catalyst was omitted. Conversion determined by 1H NMR of the crude
reaction mixture.
Next, a screen of various anilines was examined to investigate the impact of both
sterics and electronics on the overall reactivity. Electron-withdrawing groups (Table A52, Entries e, f & i) attenuated somewhat the overall nucleophilicity of the aniline at the
nitrogen atom. Sterics played an even larger role hindering the overall reactivity as
evidenced by the low conversion when utilizing 2,6-dimethylaniline (Table A5-2, Entry d)
or /V-ethylaniline (Table A5-2, Entry c). When sterics and electronics counteracted one
another such as when utilizing o-anisidine (Table A5-2, Entry g), conversion recovered,
though when sterics and electronics reinforced each other, conversion plummeted, as
was the case with 2-nitroaniline (Table A5-2, Entry e).
158
^ N H
l
\^
Entry
J
fi
2
II
10mol%AcOH
..
200 °C, 20 min
Aniline
B ^ ^ O M e
1"J
0
Conv.
Product
H
a
aniline
81
KJ
O
H
b
r ^=Y'
p-toluidine
H
c
A/-ethylaniline
d
2,6-dimethylaniline
3
N
\-^- Y ' O M e
C ^
79
°
43
?H3H
29
^ C H
° '
3
NO2H
e
o-nitroaniline
f
p-nitroaniline
4
56
0
2
N ^
°
H
g
o-anisidine
70
U-0CH3 °
h
m-anisidine
i
3'-aminoacetophenone
J
2-bromoaniline
O
l^J
U
79
H
decomposition/explosion
34
0
Table A4-2. Short screen of various anilines with methyl acrylate. It can be seen
that steric hindrance plays a significant role and yields suffer accordingly (Entries
c & d). Similarly, electron-withdrawing groups further decrease the nucleophilicity
of the aniline nitrogen (Entries e, f, and j). In the case of 2-bromoaniline,
decomposition (explosion) was encountered when the reaction was attempted.
Conversion determined by 1H NMR of the crude reaction mixture.
Finally, a short screen of Michael acceptors was investigated. While the addition
of aniline to n-butyl acrylate proceeded smoothly, substitution of f-butyl acrylate led to no
conversion. This is not unexpected, as at these temperatures, the f-butyl acrylate will
decompose to afford /'sobutylene and acrylic acid via a 3,3-sigmatropic rearrangement.
Other limitations to this methodology include the need for the Michael acceptor to be
159
unsubstituted at the p-carbon. In the attempted coupling of aniline with methyl cinnamate
(Table A5-2, Entry f), only starting materials were recovered. Finally, when utilizing
acrylamide as the potential Michael acceptor (Table A5-2, Entry g), the polymerization of
the acrylamide at these temperatures was facile, leading to the inability to determine the
reaction conversion.
iTY
K^
NH 2
EWQ
II
Michael
acceptor
a
methyl
acrylate
b
n-butyl
acrylate
c
f-butyl
acrylate
d
acrylonitrile
e
phenyl vinyl
sulfone
g
j^
/ ^ C U / R
*
Entry
f
10mol%AcOH
methyl
cinnamate
acrylamide
200 °C, 20min
N
Conv.
Product
H
81
H
83
decomposition
0
^ c
60
Q
r
N
0-^30,™
96
no reaction
0
polymerization
n.d.
Table A4-3. Short screen of various Michael acceptors with aniline. Most
proceeded efficiently, however decomposition was observed when employing the
f-butyl ester (Entry c). Again, steric factors dominate, as no reaction was
observed when utilizing methyl cinnamate (Entry f). Conversion determined by 1H
NMR of the crude reaction mixture.
Conclusions. A simple methodology was developed for the aza-Michael addition
of anilines to various activated olefins. Solvent-free conditions, high temperatures, and a
catalytic amount of acetic acid facilitated the formation of the desired adduct in
moderate-good yields. The utility of this methodology is limited to cases where the
activated olefin is unsubstituted, especially at the p-position. Furthermore, stericallydemanding anilines also show attenuated reactivity.
160
This methodology was utilized on a number of occasions to test the efficiency of
different scientific microwave apparatus, especially in the scale-up arena (Chapter 4).
Because the aniline-methyl acrylate reaction mixture was not particularly absorbent, a
qualitative analysis could be made of the magnetron's overall efficiency when heating
large quantities of this reaction mixture.
161
Appendix 5 •
Synthesis of Benzophenones, Initial Studies
During the initial kinetic investigations into the synthesis of diarylmethanes,
reactions were run open to the atmosphere. Granted, this is generally not good practice
especially with respect to transition metal-catalyzed reactions, as oxygen can poison the
catalyst as well as cause undesired side-products. However, because the reaction was
developed without measures to exclude oxygen and yields never suffered, it was
reasoned that in this case there was little interference by atmospheric 0 2 .
In an attempt to isolate the discrete steps of the reaction, the diarylated
intermediate,
a-(4-methylphenyl)-a-phenylacetophenone
A5.1, was
isolated.
The
kinetics of the final step to afford 4-methyldiphenylmethane A5.2 could be examined
before attempting to de-convolute the kinetic data from the full reaction. In this case,
however, formation of 4-methylbenzophenone A5.3 was noted as a side product.
Figure A5-1. Initial kinetic studies to extract data from the second step of the
reaction yielded the unexpected substituted benzophenone as a significant side
product.
From the conversion
data plotted
below
in
Figure A5-2, two
rather
unusual
characteristics became apparent. First, the rate of formation of benzophenone increased
as temperature decreased. Secondly, the rate of formation of benzophenone appears to
be zero-order. However, the apparently zero-order reaction can be explained in an
alternative manner.
163
Conversion of A5.1
I °-8 $S>
500
4.
. . . 1000
time (mm)
-•-110 C — 1 0 0 C
—*-90C
1500
-o-80 C
Formation of 4-methyldiphenylmethane (A5.2)
a>
c
c 0.8
re
|
0.6
&0A
*°:tt
>
500
1000
1500
time (min)
- • - 1 1 0 C - " - 1 0 0 C —*— 90 C - < ^ 8 0 C
Formation of 4-methylbenzophenone (A5.3)
0.5
0>
c 04
o
c
.c 0.3
Q.
o
N
C
0)
DO
>P
0.2
0.1
)
500
1000
time (min)
— 1 1 0 C - " - 1 0 0 C -*— 90 C
-*»-80C
1500
Figure A5-2. Plots of conversion versus time for a) the consumption of starting
material, b) the formation of the desired 4-methyldiphenylmethane, and c) the
formation of 4-methylbenzophenone.
164
To evaluate whether the reaction is indeed zero order, a working hypothesis was
constructed. First, the oxidation of A5.1 to 4-methylbenzophenone A5.3 likely was due to
the presence of molecular oxygen, as the reaction was run open to the air. Research
into the solubility properties of gaseous oxygen revealed a couple of salient features. 150
First, the solubility of 0 2 decreases as temperature increases. Thus, as the reaction
temperature was lowered, it is hypothesized that there was more 0 2 dissolved and
hence the formation of the benzophenone side product increased as temperature
dropped. Secondly, 0 2 is significantly more soluble in non-polar solvents than it is in
water. This second phenomenon was very important when attempting to interpret the
data above. It may be tempting to look at the data for the formation of the benzophenone
and assign a zero-order kinetic dependence on the reaction. However, it is hypothesized
that this was not actually the case (and is likely 2 nd order). Instead, it is posited that the
formation of benzophenone begins slowly when the reaction (a phase-transfer emulsion
in water) is very polar. As some of the 4-methyldiphenylmethane A5.2 was formed,
however, the reaction becomes more hydrophobic/lipophilic. Thus, the reaction will be
better able to dissolve oxygen as it progresses. As such, the decrease in concentration
of the starting material will counteract the increase in 0 2 concentration and will result in
kinetics that are apparently zero order, i.e. the apparent rate will remain steady. That
said, a systematic evaluation needs to be undertaken to determine the kinetics of the
reaction—including overall order and order with respect to each reagent—before any
concrete assertions are made!
A few cursory investigations were carried out on this system before returning to
the kinetic studies of the diarylmethane formation. Two main adjustments were made.
First, in order to test the importance of the "greasiness" of the system in the formation of
150
(a) Golovanov, I. B.; Zhenodarova, S. M. Russian Journal of General Chemistry 2005, 75, 1795-1797. (b)
Battino, R.; Rettich, T. R.; Tominaga, T. J. Phys. Chem. Ref. Data 1983, 2, 163-178.
165
the benzophenone, a 50:50 toluene: water biphasic reaction was employed. Secondly,
the reaction was carried out under 100 psi of oxygen in order to help favor the formation
of the benzophenone and avoid formation of the diarylmethane. Eureka! On the first
attempt, - 9 7 % conversion to the desired benzophenone was achieved in only 2 hours
even though the reaction was run at the same temperature of 100 °C as above when the
"side product" 4-methylbenzophenone A5.3 was first discovered.
X, H ?
r^Y ^ -" 'J
O
fi^Y
' ^ V ^
5
^
KJ Ml
kJ
^ ^ A5.1
O
1 mL 3.0 M NaOH (aq.)
^"
1 ml toluene,
TBAB , 0.5 eq
100 |DSi 0 2
^r^Y^i
fl
^
^
+ rT^
*J IA* ^
^ ^ C H
3
A5.2
Entry
Time
Temp
Adjustments
a
2hr.
100 "C
none
b
2hr.
100 : C
no TBAB
2hr.
100 °C
4 mmol NaOH
1 eq. BHT
d
2hr.
100 °C
1 eq. DABCO
complete conversion
to benzophenone
e
2hr.
100 °C
2 eq. DMSO
complete conversion
to benzophenone
f
1 hr.
100 °C
none
n
.'50 nun
flO C
nurifj
c
Conversion
97% benzophenone
2% diarylmethane
1%s. m.
25% benzophenone
75% s. m.
complete conversion
to benzophenone
T l f ^ l
<i^
^
\-^\~,,
^ ^
CH3
A5.3
Notes
BHT radical
scavenger
DABCO a known
singlet oxygen
scavenger
DMSO a known
singlet oxygen
scavenger
96% benzophenone
4% diarylmethane
o0"^ bcn/optic-nniii'
40% s.m.
Table A5-1. Oxidation a,a-diarylated acetophenone under 100 psi 0 2 using a
biphasic water-toluene solvent system and TBAB as a phase-transfer reagent
yields the non-symmetric benzophenone. The reaction proceeds smoothly even
in the presence of the radical trap 1,6-di-f-butyl-4-methylphenol (BHT), as well as
DABCO and DMSO, known singlet 0 2 quenchers.151
The use of strong bases under an oxygen-rich atmosphere to decompose aryl ketones
to the corresponding benzoic acids plus a ketone/aldehyde has been demonstrated,152
151
(a) Schaap, A. P.; Recher, S. G.; Faler, G. R.; Villasenor, S. R. J. Am. Chem. Soc. 1983, 105, 16911693. (b) Shaap, A. P.; Faler, G. R. J. Am. Chem. Soc. 1973, 95, 3381-3382. (c) Ovannes, C; Wilson, T. J.
Am. Chem. Soc. 1968, 90, 6528.
152
(a) Doering, W. von E.; Haines, R. M. J. Am. Chem. Soc. 1953, 76, 482-486. (b) Gersmann, H. R.; Bickel,
A. F. J. Chem. Soc. B 1971, 2230-2237.
166
though there has been no systematic approach to utilizing this methodology for the
purpose
of
generating
non-symmetric
benzophenones.
Furthermore,
the
exact
mechanism is still uncertain,153 though a recent theoretical investigation into the
mechanism of the oxidation of linalool by 0 2 may shed some light on potential
mechanisms.154 Ideally, a two-step, one pot procedure (Figure A5-3) could be employed
where an equivalent of deoxybenzoin would be coupled utilizing Pd-enolate chemistry.
After the first step, a solution of sodium hydroxide and TBAB in water could be added
and the reaction pressurized with 0 2 and allowed to react at 100 °C for approximately 60
minutes. This hypothetical protocol appears to be feasible based upon initial studies that
have been undertaken. The primary obstacle would be to find a set of conditions to
affect the initial coupling, but that do not decompose the intermediate to form the
diarylmethane.
not isolated
Figure A5-3. Hypothetical 2-step, 1-pot synthesis of non-symmetric
benzophenones. The second step of the reaction is facile and should prove trivial
across a wide range of substrates.
The
isolation
of
the
intermediate
is
straightforward
for
unsubstituted
bromobenzene, or for the electron-rich 3-bromotoluene, 4-bromotoluene, and 4bromoanisole. For these substrates, aqueous K 2 C0 3 was substituted for the stronger
hydroxide solution. Utilizing 0.5 equivalents of TBAB as a phase transfer reagent, 0.2
mol % PdCI2, the intermediate was isolated in better than 95% yield after 6-10 hours,
depending on the substrate. In all cases, less than 1 % of the diarylmethane was
detected in the crude 1H NMR.
153
(a) Gersmann, H. R.; Bickel, A. F. J. Chem. Soc. B 1971, 2230-2237. (b) Backtrop, C; Wass, J. R. T. J.;
Panas, I; Skold, M.; Borle, A.; Nyman, G. J. Phys. Chem. A 2006, 110, 12204-12212.
154
Backtorp, C; Wass, J. R. T. J.; Panas, T.; Skold, M.; Borje, A.; Nyman, G. J. Phys. Chem. A 2006, 110,
12204-12212.
167
o
0.2 mol % PdCI2
+
^ \
Br H20/K2C03/TBAB H
95°C, 6-10hr.
R_£"^f
R = 4-OCH3, 4-CH3, 3-CH3
90-95% isolated yields
o(p-OCH3)= -0.12, o(p-CH3)= -0.14,
a(m-CH3)= -0.06
R
= 4-CF3, directly to diarylmethane
R
= 3 " C| . -50:50 mix of
intermediate:diarylmethane
a(m-CI)=0.37, a(p-CF3)=0.53
Figure A5-4. Intermediate trapping is straightforward when using unsubstituted
bromobenzene or slightly electron rich systems. Changing the aqueous base
from hydroxide to carbonate and running the reaction at 95°C for 6-10 hours
affords isolated yields of 90-95% of the arylated intermediate. When moving to
more electron-withdrawing systems, however, the reaction progressed to yield
the (now unwanted) diarylmethane as a significant by-product.
However, the isolation of the intermediate is not feasible for all substrates. It is most
problematic when the substitution of the aryl bromide has an electron-withdrawing group,
as
it
leads
to
stabilization
of
the
carbanionic
leaving
group.
Utilizing
3-
chlorobromobenzene as the coupling partner, a range of conditions were examined in an
attempt to suppress the formation of the diarylmethane. The most success was met
utilizing NMP or DMF as the solvent with potassium carbonate as the base. There was
some impact when employing phosphine ligands to stabilize the palladium in terms of
catalyst longevity, though in terms of initial, reactivity, palladium (II) chloride bisacetonitrile proved most reactive. The slight variation in activity can be appreciated in the
five examples that were monitored and the progress of the reaction was plotted. The
comprehensive data is in tabular (Table A5-2 and Table A5-3) and graphical (Figure A55) forms, below.
168
o
conditions
DMF
Entry
Time
(hr.)
Temp
(°C)
0.5
140
0.5
140
18
140
18
18
130
0.5
150
0.5
150
18
120
18
120
Pd source,
loading (mol
.%)
PdCI.(MuCN»„
1 mol %
PdCI2(MeCN)2,
1 mol %
PdCI2(MeCN)?,
1 mol %
P'iOl,(PPh,) 2 ,
1 mol %
PdCl.idppc).
1 mol ••;,
PdCI2(dppe),
1 mol %
PdCI2(MeCN)?,
1 mol %
PdCI2(MeCN)2,
1 mol %
PdCI 2 .
1 mol %
base
%A /%B/%C
K.CO
85.5/14.5/0
Cs 2 C0 3
51 / 3 8 / 1 1
K 2 C0 3
ground
K.CO-,
ground
K.CO
ti,2=30 min
ijniLind
k2co.,
325 mesh
K ? C0 3
325 mesh
K.CO,
325 mesh
K 2 C0 3
325 mesh
t 1/2 =129min
t- -'62 mm
no reaction
27/72/~2
ti/2=465 min
ti, z =265 min
Lab book c.f.
and notes
JRS4-128a
1 eq. TBAB
JRS4-128b
1 eq. TBAB
JRS4-128C
Graph A5-1
JRS4-129
Graph A5-2
JRS4-130
Graph A5-3
JRS4-131a
In dioxane
JRS4-131b
JRS4-132
Graph A5-4
JRS4-133
Graph A5-5
Table A5-2. Optimization of conditions to trap the arylated intermediate.
169
JRS4-128C
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
200
400
- starting material
600
•
800
intermediate
1000
1200
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3 0.2 0.1
0
0
—*•-•- diarylmethane
JRS4-129
200
400
—
200
400
starting material
600
800
•— intermediate
800
1000 1200
-starting material
1400
- diarylmethane
JRS4-131
0
600
JRS4-132
1000
1200
200
1400
—*— diarylmethane
400
600
800
1000
1200
1400
-starting material —•—intermediate —*—diarylmethane
JRS4-133
200
400
- starting material
600
800
- intermediate
1000
1200
1400
diarylmethane
Figure A5-5. Plots of conversion versus time for the trapping of the arylated
intermediate of deoxybenzoin (Aryl l= 3-chlorophenyl) for selected entries in
Table A5-2, above. Excel file: intermediate_trapping.xlsx
170
o
Br
conditions
NMP
CI
CI
Entry
Time
(min)
Temp
a
30
140
b
30
160
c
30
150
d
30
150
e
30
150
f
60
150
g
60
150
h
240
140
i
30
160
J
30
160
k
60
150
I
60
150
m
16 hr
130
16 hr
130
16 hr
130
16 hr
130
17 hr
130
17 hr
130
17 hr
130
17 hr
130
17hr
130
Pd source,
loading
PdCI2(MeCN)2>
1 mol %
PdCI 2 (MeCN) 2 ,
1 mol %
Pd(OAc) 2
1 mol %
PdCI 2 (PPh 3 ) 2 ,
1 mol %
PdCI 2 (MeCN) 2 ,
1 mol %
PdCUPPh,),,
1 mol %
PdCI 2 (PPh 3 ) 2 ,
2 mol %
PdCI 2 (PPh 3 ) 2 ,
2 mol %
PdCI 2 (PPh 3 ) 2 ,
2 mol %
PdCI.,(PPh.).,.
1 mol %
Pd(PPh 3 )„,
1 mol %
Hermann's
1mol%
Hermann's
1 mol %
Hermann's
2 mol %
PdCI ? (MeCN) ? .
1 mol %
PdCI2(PPh.,)2.
1 mol %
PdCI 2 (MeCN) 2 ,
0.5 mol %
PdCI 2 (MeCN) 2 ,
0.1 mol %
PdCI 2 (PPh 3 ) 2 ,
0.5 mol %
PdCI 2 (PPh 3 ) 2 ,
0.1 mol %
PdCI 2 ,
0.5 mol %
CI
% A/ % B / % C
Lab book c.f. and notes
36/64/none
detected
JRS4-137a
18/80/2
JRS4-137b
44/56/<1 ,
JRS4-137C
17/82/1
JRS4-137d
1 7 / 8 1 12
JRS4-138a
11/86/3
JRS4-138b
8/90/2
JRS4-138C
41/54/5
JRS4-139a
23/63/12
JRS4-139b
9/86/5
JRS4-139C
42/55/3
JR-S4-140J
31/66/3
JRS4-140b
44/54/2
JRS4-141a
59/38/3
JRS4-141b
27 / 72 / 1
JRS4-141C
20 / 78 / 2
JRS4-141d
30/66/4
JRS4-142a
43/47/10
JRS4-142b
28/69/3
JRS4-142C
40/50/10
JRS4-142d
32/63/5
JRS4-142e
Table A5-3. Intermediate trapping using NMP as the solvent.
171
Conclusions & Future Development. This project was orphaned after a few
weeks to return to the kinetic studies of diarylmethanes, but holds great potential for
continued development. If conditions could be developed in toluene, a 2 step, one pot
procedure would give facile access to a wide range of non-symmetric benzophenones. A
recent paper in this area may provide some additional input to the further development
of this project, especially with regard to solvent selection.155 Other
appropriate
references may also be beneficial.156 Furthermore, as much as it pains the author to say
this, it might be beneficial to screen a range of activating phosphine ligands. This might
allow for slightly lower reaction temperatures as well as alternative solvents. Indeed,
toluene may prove to be the ideal solvent, as it should inhibit the formation of the
diarylmethane due to the solvent's inability to stabilize that high-energy carbanionic
leaving group. Toluene was abandoned in favor of DMF and later NMP, as it appeared
that the palladium coupling was suppressed when using the simple palladium salts in
toluene. Additionally, while the 325 mesh K 2 C0 3 worked well in NMP, other bases such
as K3PO4 may be more effective when moving to other solvents.
155
Desai, L. V.; Ren, D. T. Rosner, T. Org. Lett. 2010, 12, ASAP Article DOI: 10.1021/ol1000318.
(a) Satoh, T.; Kametani, Y.; Terao, Y.; Miura, M.; Nomura, M. Tetrahedron Lett. 1999, 40, 5345-5348.
(b)Satoh, T.; Inoh, J.-I.; Kawamura, Y.; Kawamura, Y.; Miura, M.; Nomura, M. Bull. Chem. Soc. Jpn. 1998,
71, 2239-2246. (c) Churruca, F.; SanMartin, R.; Tellitu, I.; Dominguez, E. Org. Lett. 2002, 4, 1591-1594. (d)
Churruca, F.; SanMartin, R.; Carril, M.; Tellitu, I.; Dominguez, E. Tetrahedron 2004, 60, 2393-2408.
156
172
Appendix 6 •
Investigations into the Synthesis of Tamoxifen
Background & Introduction. At a group meeting exercise, the structure of
Tamoxifen A6.1 was presented to the group as a retrosynthetic analysis problem. As the
Leadbeater group deals extensively with palladium-catalyzed reactions, it was no
surprise that the two organic chemists of the group quickly and independently developed
similar retrosynthetic analyses. As a follow-up, investigations into current synthetic
protocol via patent literature as well as a number of recent attempts to synthesize
Tamoxifen in a more efficient manner were evaluated.
\
N—\
/
^ O
Tamoxifen A6.1
Tamoxifen is an orally available drug used in the treatment of breast cancer.157
Developed by ICI Pharmaceuticals (now Astra-Zeneca), this off-patent drug is currently
the world's top-selling drug for this purpose. In addition to treating breast cancer, recent
studies suggest that Tamoxifen and other tetra-substituted olefins may have may have
cancer preventative properties158 and effectiveness in the preservation of bone mineral
density related to osteoporosis.159 Thus, these tetrasubstituted olefins have potential for
broader medicinal applications.160
A6.2 Recent Advancements in the Synthesis of Tamoxifen. Perhaps it is of
no surprise that a number of research groups have recently been attempting to
synthesize Tamoxifen in a more concise manner with the aid of transition metal
157
Furr, B. J. A.; Jordan, V. C. Pharmacol. Ther. 1984, 25, 127.
Jordan, V. C. J. Steroid Biochem. Mol. Biol. 2000, 74, 269.
159
Grey, A. B.; Stapleton, M. C ; Evans, M. C; Tatnell, R. W.; Ames, B.; Reid, I. R. Am. J. Med. 1995, 99,
636.
160
Grese, T. A.; Dodge, J. A. Estrogen Recptor Modulators: Effects In Non-Tradition Target Tissues. In
Annual Reports in Medicinal Chemistry; Bristol, J. A. Ed., Academic Press: San Diego, CA, 1996; Vol. 31, pp
181-190.
158
174
catalysis. Figures A6-1 through A6-5 illustrate a number of recent protocol developed for
the synthesis of Tamoxifen.
1%PdCI2(PhCN)2
3 eq. K 2 C0 3
2:1 DMF:H 2 0
45°C, 24 h
removed by column
chromatography
68% yield, > 95% pure
Zhou, C ; Larock, R. C. J. Org. Chem. 2005, 70, 3765.
13% overall yield, 3 steps
Figure A6-1. In 2005, Larock and co-workers reported on a three-component
one-step procedure that yielded Tamoxifen in 68% yield. No discussion was
presented on the synthesis of the unique phenylboronic acid. Drawbacks to this
methodology include the need to use a large excess of both the boronic acid and
the iodobenzene and the necessary column chromatography to remove the
significant quantities (2-fold excess!) of biphenyl produced.
NBu4HS04
aq. NaOH
PhOH + CI(CH2)2CI
OCH2CH2CI
/=(
Ph
OH
JFAA, r.t., 70 h
84-94%
85%
0CH2CH2CI
OCH2CH2CI
1)KH, THF
»2) Tf2NPh, LiBr
63%
OCH2CH2CI4 e q
P h Z n C | 5o/o
0-S0 2 CF 3
p d (pp h3 ) 4
110°C, 1h
»»
99% (0.5 mmol scale)
Potter, G. A.; McCague, R. J. Org. Chem. 1990, 55, 6184.
5 steps, < 50% overall yield
Figure A6-2. In 1990, Potter and co-workers described a 5 step synthetic
procedure that yielded Tamoxifen in approximately 50% overall yield. Drawbacks
include the need for chromatography after the first three steps and the large
excess of aryl zinc reagent required.
175
1)1.4eq. Et2AICI
1.4 eq. Cp2TiCI2
•SiMe,
^—?
2) NBS, -78°C
85%
SiMe3
— /
PhZnCI, Pd(PPh3)4 (5 mol.%)
THF, reflux
»~
SiMe3
fl\
95%
iBr
OMe
Br2, DCM, NaOMe/MeOH
-78°C to r.t.
p-MeOC6H4ZnCI
Pd(PPh3)4 (5 mol.%), THF
k
to-
84%
85%
1)NaSEt,DMF, reflux
2) CI(CH2)2NMe2 HCI
NaOEt, EtOH
»3) dry HCI, Et 2 0
4) 0.5 M NaOH
60%
Miller, R. B.; Al-Hassan, M. I. J. Org. Chem. 1985, 50, 2121.
34.6% overall, 7 steps
Figure A6-3. In 1985, Miller and co-workers described a 7-step synthesis built
around metal-catalyzed coupling reactions, including Al, Ti, Zn, and Pd.
Drawbacks to this approach include the relatively long 7-step linear synthesis,
rather high palladium loading, and two cryogenic steps.
0(CH 2 ) 2 NMe 2
0(CH 2 ) 2 NMe 2
10%PdCI 2 (PPh 3 ) 2
10%Cul
NEt3, THF, 18h
2) 5% Pd(PPh3)4
3.6 eq. Phi
=—+-
OH
0(CH 2 ) 2 NMe 2
1)3.2eq.PhMgCI
tol. reflux, 16 h
83%
72%
"OH
1 ) 3 e q . DMP, DCM, r.t., 12 h
2) KOf-Bu, Ph3PCH3Br
THF reflux, 16 h
0(CH 2 ) 2 NMe 2
*3) H 2 , Pd/C, r.t., 2 h
66%
7 steps, 35.5% overall
Tessier, P. E.; Penwell, A. J.; Souza, F. E. S.; Fallis, A. G. Org. Lett. 2003, 5, 2989.
Figure A6-4. In 2003, Tessier and co-workers reported a 7-step route to
Tamoxifen in a reported 35.5% overall yield. Drawbacks to this approach include
high palladium loading, Dess-Martin oxidation, and the poor atom economy Wittig
reaction.
176
0(CH 2 ) 2 NMe 2
OH
0(CH 2 ) 2 NMe 2
m
2) CI(CH 2 ) 2 NMe 2
tol. reflux, 12 h
1) C 0 2 (1 atm), 20% Ni(cod)2
10 eq. DBU, 3 eq. Ph2Zn
THF, 40°C, 20 h
1%PdCI 2 (PPh 3 ) 2
2% Cul
1)KOH, EtOH
Ph
NEt3, 6 h
80%, 2 steps
2) CH 2 N 2
63%, 22% undesired isomer
Ph
C0 2 Me
1 ) 3 e q . DIBAL-H,-78°C
Dess Martin ox.
2)
3) KOf-Bu, Ph3PCH3Br
4) H 2 , Pd/C
Me 2 N(H 2 C) 2 d
71%
Me 2 N(H 2 C) 2 0
36% overall, 8 steps
Shimizu, K.; Takimoto, M.; Mori, M.; Sato, Y. Synlett, 2006, 18, 3182.
Figure A6-5. In 2003, Shimizu and co-workers reported a 8-step route to
Tamoxifen very similar to the route reported by Tessier (Figure A6-4) in a
reported 36% overall yield. Drawbacks to this approach include high nickel
loading in the carbonylation step, again Dess-Martin oxidation, and poor atom
economy for the Wittig reaction.
A6.3. Retrosynthetic Approach. After examining reported protocol, it was
decided that the retrosynthetic approach that had been developed warranted further
investigation. On paper, it appeared that this approach would have a number of
advantages over previously reported work, including: 1) it was a four-step synthesis from
iodo- or bromophenol, 2) it utilized relatively inexpensive and readily available materials,
and 3) it should require no chromatography, a big plus. The proposed building blocks
with approximate prices can be seen in Figure A6-6 and the proposed retrosynthetic
approach can be seen in Figure A6-7.
$37/mol
,N V
$12/mol
V
$112/mol
\
//
HO
X=Br: $62/mol
X=l:$510/mol
CI
Bromine: $23/mol
Figure A6-6. Proposed Building blocks with approximate cost per mole for the
synthesis of Tamoxifen.
177
Suzuki
(rans-specific
dibromination
Sonogashira
alkylalion
Figure A6-7. Proposed retrosynthesis for the synthesis of Tamoxifen utilizing two
palladium catalyzed steps including a dual 1,2-diphenylation of a dibromoalkene
utilizing a Suzuki coupling protocol.
A6.4. Discussion and Results. The alkylation of iodo- and bromophenol as well
as the Sonogashira reaction with 1-butyne were reported elsewhere with good results.161
The
trans-specific
dibromination
of
alkynes
has
been
reported162
but
further
investigations were undertaken in order to optimize bromination conditions on a system
that would better typify the Tamoxifen alkyne intermediate. Conditions were quickly
developed first using 1-phenylpropyne as a model alkyne. The conditions utilized 10 mol
% of lithium bromide and one equivalent of Br2 which was added drop wise to an ethanol
solution of the alkyne over approximately 10 minutes. Running the reaction at room
temperature yielded an 8:1 ratio of the desired frans-dibrominated alkene to the
undesired c/s-dibromoalkene. However, upon cooling the reaction to 0 °C, only the
desired c/s-dibromoalkene was formed.
The final step in the synthesis of Tamoxifen was to be a double Suzuki coupling
on the frans-1,2-dibromoalkene. However, initial analogue studies on frans-1,2-dibromo1-phenylpropene quickly revealed that the final step of the route would be destined for
failure, as the Pd-catalyzed elimination to form 1-phenylpropyne was facile (Figure A68).
Kormos, C. M., Ph.D. Thesis, University of Connecticut, 2010.
(a) Pincock, J.A.; Yates, K, J. Am. Chem. Soc. 1968, 90, 5643. (b) Chiappe, C; Caprato, C; Conte, V.;
Pieraccini, D. Org. Lett. 2001, 3, 1061.
162
178
only trace detected
byGCMS
frans-1,2-dibromo1-phenylpropene
elimination
regenerates alkyne
Figure A6-8. Attempts to effect the dual Suzuki reactions were problematic
owing to the fact that the (intramolecular) palladium-catalyzed elimination of
frans-1,2-dibromo-1-phenylpropene to afford 1-phenylpropyne was more facile
than the palladium catalyzed Suzuki reaction. A trace of a product with an m/z 270 was detected by GCMS, indicating limited success forming the desired 1,1,2triphenylpropene.
Though some diarylated product was detectable via GCMS, the elimination to afford the
alkyne was more favorable under the conditions screened, and it was hypothesized that
the intermolecular
Suzuki reaction likely would never
be competitive with the
intramolecular elimination that affords the alkyne. Though the first three steps of the four
step synthesis of Tamoxifen were high yielding and at no point was chromatography
needed, the inability to effect the double Suzuki reaction on the final step unfortunately
spelled the demise of this approach as an effective methodology. As such, the project
was orphaned at this point.
Appendix 7 •
Detailed Experimental Protocol
General. All solvents are purchased from Aldrich, Baker, or Fischer and used as
received without re-distillation. Similarly, all reagents are used as received from various .
vendors without purification. Phenylboronic acid is purchased from Optima Chemical
Group, Douglas, GA and used as received. For automated flash chromatography, prepacked columns available from Biotage or Luknova were employed in 25, 40, or 50 g
sizes, as appropriate. All pre-packed columns utilized 60 mesh silica gel. All reactions
described are prepared in air. Unless specifically stated otherwise open-vessel reactions
are run open to the air.
Chapter 2 Raman Spectroscopy Monitoring of Microwave Assisted Reactions
General. All reactions were carried out in a CEM Discover S-Class microwave.
Typical procedure for monitoring the formation of 3-acetylcoumarin (2.1).
Into a 50.00 ml volumetric flask was placed salicylaldehyde (6.106 g, 50.00 mmol) and
ethyl acetoacetate (6.507 g, 50.00 mmol). The reagents were diluted to 50.00 ml with
ethyl acetate. This solution was transferred to a 50 ml long-necked round bottom flask
equipped with a Teflon-coated stir bar. The flask was placed into the microwave cavity,
making sure that any company glassware markings were orthogonal to the Raman laser
path. The microwave attenuator was then locked in place. A 2" adapter was connected
to the round bottom flask to allow a Claisen adapter with septum inlet to be placed atop
the reaction flask (Figure 18). A reflux condenser was placed on the Claisen adapter.
The septum inlet was capped with a rubber stopper with a 22-gauge syringe needle
inserted through. The Raman probe was inserted into the microwave cavity until the
quartz light pipe just made contact with the side of the reaction flask. The reaction
mixture was brought to reflux (83-84 °C) using a microwave power of 50 watts. Once at
reflux, the microwave power was dropped to 7 watts and the temperature was set
181
artificially high at 90°C to ensure the microwave was always applying continuous power
and reflux was maintained. At this time, a background scan of the reaction was
recorded. This background is automatically subtracted from all subsequent scans,
removing any signal due to starting materials and solvent. The Raman spectrometer was
set to acquire a spectrum every 10 s (actual observed acquisition rate was 12.2 s), with
10 s integration times, "boxcar" set to 3 and "average" set to 1. Continuous scans were
then begun. After the first scan (t=0), the piperidine catalyst (436 mg, 4.0 mmol, 8 mol
%) was rapidly injected into the reaction mixture. As the reaction is dynamic and the
change in concentration is initially linear, the first acquisition after the catalyst is injected
was set to half of the observed acquisition rate (e.g. 6.1 s). Each subsequent acquisition
is set to 6.1 + 12.2n seconds. After running the reaction for the requisite period of time
the Raman data acquisition and the microwave heating were halted. The reaction
mixture was allowed to cool to room temperature and the product isolated was stopped.
Upon cooling, the product was collected by vacuum filtration and recrystallized from
ethanol. Reactions carried out below reflux (35-75°C) were done in a procedure
analogous to the Knoevenagel condensation procedure below.
Monitoring formation of
ethyl-(E/Z)-2-acetyl-3-(2-methoxyphenyl)-acrylate
(2.4) Into a 50.00 ml volumetric flask were placed 2-methoxybenzaldehyde (6.808 g,
50.00 mmol) and ethyl acetoacetate (6.507 g, 50.00 mmol). The reagents were diluted to
50.00 ml with ethyl acetate. This solution was transferred to the 50 ml long-necked round
bottom flask and placed into the microwave cavity. The reaction mixture was brought to
the desired temperature (35-75 °C) at which point a background scan of the reaction that
would be subtracted from all subsequent scans was taken. The Raman spectrometer
was set to take continuous scans using the same parameters as with the 3-acetyl
coumarin synthesis. Continuous scans were begun.
After the first scan (t=0), the
stopper was momentarily removed and the piperidine catalyst in toluene (4.0 M, 1.0 ml,
182
4.0 mmol, 8 mol %) was rapidly injected into the reaction mixture.
Upon cooling, the
reaction mixtures were combined. To drive the reaction to completion and thus allow for
isolation and characterization, a portion of the reaction mixture was transferred to a 35
ml thick-walled reaction vessel, more ethyl acetoacetate added and then the vessel
sealed and heated to 120°C until completion was reached as judged by Raman
monitoring (-12 min). The crude reaction mixture was poured over aqueous HCI (2.0M)
and extracted with ethyl acetate. The organic layer was washed with brine, dried over
MgS0 4 , and the solvent was removed under vacuum. The E/Z isomers co-distill (184°C
@8-10 mmHg) so 500 mg of the distillate was placed on a 20-cm silica gel column (80 g
silica gel) and eluted using 6:3:0.5 pet ether:DCM:ethyl acetate solvent system to
separate the E a n d Z isomers. Rf=0.342 (Z-isomer), 0.210 (E-isomer).
General
procedure
for
the
synthesis
and
monitoring
of
dihydropyrimidinones (2.5-2.14) A 50.0-ml solution was prepared by dissolving urea
(50.0 mmol, 3.00 g), ethyl acetoacetate (50.0 ml, 6.514 g), and benzaldehyde (50.0
mmol, 5.306 g) in absolute ethanol. This solution was placed in a 100-ml long-necked
round bottom flask with a Teflon-coated stir bar and equipped with a modified Claisen
head adapter with reflux condenser. The flask was placed in the microwave cavity and
brought to reflux at which point a dark scan would be collected by the Raman
spectrometer. At this point, continuous scans were started. Immediately after the first
scan was collected, one milliliter a 6.0 M solution of H 2 S0 4 in ethanol was quickly
introduced to the reaction (t=0). Raman scan would be collected every 10 seconds or
until the dihydropyrimidinone product crystallized in the flask. Upon completion, the
reaction was allowed to cool and the product was collected by vacuum filtration.
General procedure for the synthesis and reaction monitoring of chalcones
(2.15-2.28) Five hundred milliliters of a stock solution of acetophenone (250 mmol,
30.037 g, 0.5M) and benzaldehyde (250 mmol, 26.530 g, 0.5 M) in 95% ethanol is
183
prepared. This solution is allowed to stand at room temperature for 2 hours. Twenty-five
milliliters of this solution is placed in a 50-ml long-necked round bottom flask with a
Teflon-coated stir-bar, stoppered with a glass stopper, placed in the microwave cavity,
and brought to the desired temperature. When the reaction reaches the desired
temperature, a background scan is taken that is subtracted from all subsequent scans.
At this point, continuous Raman scans begin. With 5-second integration times, a scan is
collected approximately every 7.2 seconds. After the first scan is completed the glass
stopper is quickly removed and 500 pi of a 2.00 M aqueous solution of NaOH is injected
(8.0 mol %). The reaction is monitored for the first 15-20 scans, the scans are halted,
and the microwave is stopped. This process is repeated 20 times, 4 times at each of five
temperatures. The reactions are combined, ice is added, and the combined reactions
are allowed to stir at 5 °C overnight. The solid product is collected, re-crystallized from
ethanol, and allowed to dry under vacuum overnight. This analytically pure product is
then used to set up the calibration curve, see above. In some instances, deviations from
this general procedure are required:
1) chalcone, no deviations
2) 4'-chlorochalcone, 0.25 M in reagents, 315 pi 2.00 M NaOH (10 mol %) catalyst
3) 2'-chlorochalcone, 0.25 M in reagents, 315 pi 2.00 M NaOH (10 mol %) catalyst,
product is a viscous liquid and is purified via column chromatography (9:1,
hexanes:ethyl acetate)
4) 4-chlorochalcone, 0.25 M in reagents, 315 pi 2.00 M NaOH (10 mol %) catalyst
5) 2-chlorochalcone, 0.40 M in reagents, 315 pi 2.00 M NaOH (10 mol %) catalyst
6) 4'-phenylchalcone, 0.15 M in reagents, 190 pi 2.00 M NaOH (10 mol %) catalyst, 15second integration time for reaction monitoring as well as calibration curve, reaction
monitored at 80, 70, 65, 60, and 50 °C
7) 4'-methoxychalcone, no deviations
184
8) 4,4'-dimethoxychalcone, no deviations
9) 4'-fluorochalcone, 0.25 M in reagents, 315 pi 2 00 M NaOH (10 mol %) catalyst
10)4'-methylchalcone, no deviations
11)3',4'-dimethoxychalcone, 0.25 M in reagents, 315 |_il 2 0 0 M NaOH (10 mol %)
catalyst, product is a viscous liquid and is purified via column chromatography (9:1,
hexanes:ethyl acetate), reaction monitored at 80, 75, 70, 65, and 60 °C
12)3'-bromo-4'-methoxychalcone, 0.193 M in reagents, 315 pi 2.00 M NaOH (13 mol %)
13)3'-chloro-5'-methoxy-2',4'-dimethylchalcone, 0.2926 M in reagents, 315 pi 2.00 M
NaOH (8.5 mol %), 10-second integration time for reaction monitoring as well as
calibration curve, reaction monitored at 80, 75, 70, 65, and 60 °C
14)3-phenyl-1-pyridin-2-yl-prop-2-ene-1-one, 0.25 M in reagents, 315 pi 2.00 M NaOH
(10 mol %) catalyst, for isolation a procedure developed by J.B.F.N. Engberts and
co-workers163 was utilized, as the desired chalcone, after initial formation, readily
reacts with a second equivalent of the acetylpyridine in a Michael reaction:
Otto, S.; Bertoncin, F.; Engbert, J.B.F.N. J. Am. Chem. Soc. 1996, 118, 7702-7707.
185
Chapter 4 Scale up of Microwave-Mediated Transformations, Biotaqe Advancer
Suzuki coupling reaction to afford 4'-methyl-2-biphenylcarbonitrile (4.1) In a
350 ml Teflon vessel were combined 2-bromobenzonitrile (9.10 g, 50 mmol), ptolylboronic acid (7.14 g, 52.5 mmol), sodium carbonate (15.90 g, 150 mmol), 1000 ppm
Pd stock solution (500 ml, 0.01 mol %'), ethanol (100 ml) and water (100 ml). The vessel
was placed into the Biotage Advancer, the top plate closed and the reaction mixture
heated to 140 °C using an initial microwave power of 1200 W. The contents of the vessel
were then held at this temperature for 10 min. The reaction mixture was agitated
throughout the run using the built-in mechanical stirrer. Upon completion of the heating
stage, the reaction mixture was cooled adiabatically by rapid expansion into the stainless
steel collection vessel. The product was purified and isolated in an identical manner to
that use for in the coupling of 4-bromoanisole and phenylboronic acid to give 4'-methyl2-biphenylcarbonitrile 4.1 as a white crystalline solid (9.484 g, 98 % yield).
Preparation of methyl A/-phenyl-3-aminopropanoate (4.8) In a 350 ml Teflon
vessel were combined: aniline (137 ml, 1.50 mol), methyl acrylate (135 ml, 1.50 mol) and
acetic acid (8.6 ml, 0.15 mol, 10 mol %). The vessel was placed into the Biotage
Advancer, the top plate closed and the reaction mixture heated to 200 °C using an initial
microwave power of 1200 W. The contents of the vessel were then held at this
temperature for 20 min. The reaction mixture was agitated throughout the run using the
built-in mechanical stirrer. Upon completion of the heating stage, the reaction mixture
was cooled adiabatically by rapid expansion into the stainless steel collection vessel and
the product obtained in 76 % conversion as determined by NMR spectroscopy.
Preparation of 3-acetylcoumarin
(4.3) In a 350 ml Teflon vessel were
combined salicylaldehyde (32.0 ml, 0.300 mol), ethyl acetoacetate (38.5 ml, 0.305 mol),
and piperidine (2.36 ml, 8 mol %). Ethyl acetate was added to make a total volume of
300 ml. The vessel was placed into the Biotage Advancer, the top plate closed and the
186
reaction mixture heated to 130 °C using an initial microwave power of 1200 W. The
contents of the vessel were then held at this temperature for 8 min. The reaction mixture
was agitated throughout the run using the built-in mechanical stirrer. Upon completion of
the heating stage, the reaction mixture was cooled adiabatically by rapid expansion into
the stainless steel collection vessel and the product collected by vacuum filtration and
recrystallized from ethanol to give 3-acetylcoumarin 4.3 in 7 1 % yield (40.10 g).
Synthesis of allyl phenyl ether (4.5) In a 350 ml Teflon vessel were combined
allyl bromide (24.20 g, 200.0 mmol), phenol (24.40 g, 260.0 mmol, 1.3 eq), and K 2 C0 3
(30.4 g, 220 mmol, 1.1 eq). Reagent grade acetone (99%) was added to make a total
volume of 250 ml. The vessel was placed into the Biotage Advancer, the top plate closed
and the reaction mixture heated to 120 °C using an initial microwave power of 1200 W.
The contents of the vessel were then held at this temperature for 20 min. After this time,
the mixture was allowed to cool passively to 60 °C, at which point the reaction chamber
was opened. The contents were poured into a separatory funnel containing cold water.
The aqueous phase was extracted three times with diethyl ether (100 ml). The organic
extracts were combined, washed sequentially with 2.0 M NaOH (25 ml), H 2 0 (50 ml),
and saturated NaCI (50 ml). The organics were dried over MgS04 and the diethyl ether
removed under reduced pressure until a constant weight was observed to afford allyl
phenyl ether 4.5 (23.86 g, 89%) as an amber-colored oil.
Synthesis of 3-(2-hydroxyphenyl)-1-propene (4.9) In the Advancer's 350 ml
Teflon vessel were combined allyl phenyl ether 4.5 (60 ml, 0.437
mol) and
tetrabutylammonium bromide (16.12 g, 50 mmol, 11.4 mol %). The vessel was placed
into the Biotage Advancer, the top plate closed and the reaction mixture heated to 245
°C using an initial microwave power of 1200 W. The contents of the vessel were then
held at this temperature for 30 min. The reaction mixture was agitated throughout the run
using the built-in mechanical stirrer. Upon completion of the heating stage, the reaction
187
mixture was cooled adiabatically by rapid expansion into the stainless steel collection
vessel. NMR spectroscopy showed complete conversion of the ether—83% conversion
to 3-(2-hydroxyphenyl)-1-propene 4.9 and 17% undesired conversion to phenol.
Chapter 4 Scale up of Microwave-Mediated Transformations, AccelBeam Prototype
Preparation of 3-acetylcoumarin (4.3) To the 9 L reaction flask was added 1.26
liters (12.0 moles, 1,470 g) salicylaldehyde and 1.56 L (12.0 moles, 1560 g) ethyl
acetoacetate. This mixture was diluted to 8.00 L with ethanol (1.50 M). The lid was
placed on the reaction and the fiber optic temperature probe was inserted into the
reaction mixture. The stirring paddle was fitted to the motor and the ejection tubing was
inserted into the reaction vessel. At this point, piperidine (120 mmol, 10.2 g) was added
through a small access port. The reaction chamber was securely closed and prepressurized to 280 psi with nitrogen. The reaction was heated to 130 °C using 7500 W
(2500 W x 3), the ramp time taking approximately 7 min (4 min for 2 liter scale, 13 min
for 12 L scale). At this point the magnetron power was modulated to remain at the
desired 130 °C (300-600 W; 100-200 W x 3) for 20 min. After this time, microwave
heating was stopped and the contents were ejected without cooling into a 5-gallon
receiving vessel containing 1500 ml ethanol. The solid was collected via vacuum
filtration, rinsed with 1000 ml ethanol and allowed to dry at ambient temperature for 3
days, yielding 1.664 kg (73.7%) 3-acetylcoumarin 4.3
Preparation of dihydropyrimidinone (4.4) To the 5 L reaction flask was added
ethyl acetoacetate (4.4 moles, 557 ml), urea (4.0 moles, 240 g), and benzaldehyde (4.0
moles, 404 ml) and diluted to 4 liters with ethanol (1.0 M). Without waiting for the urea to
dissolve, the lid was placed on the reaction and the fiber optic temperature probe was
inserted into the reaction mixture. The stirring paddle was fitted to the motor and the
ejection tubing was inserted into the reaction vessel. At this point, 67 ml 12.0 M HCI (aq.,
188
20 mol %) was added via an inlet port on the reaction lid. The reaction chamber was
securely closed and pre-pressurized to 280 psi with nitrogen. The reaction was heated to
120 °C using 7500 W (2500 W x 3), the ramp time taking approximately 6.5 min. At this
point the magnetron power was modulated to remain at the desired 130 °C (300-600 W;
100-200 W x 3) for 20 min. After this time, microwave heating was stopped and the
contents were ejected without cooling into a 5-gallon receiving vessel containing 1500 ml
ethanol. The product was allowed to cool and the solid was collected via vacuum
filtration and rinsed with ethanol. The solid was then allowed to dry overnight in an oven
(100 °C) to yield 576 g (55.4%) of 4.4.
Preparation of 4'-methoxybiphenyl (4.2) A solution was prepared by adding 4bromoanisole (502 ml, 4.00 mol) to enough ethanol to make 4.0 L of solution. A second
solution containing phenylboronic acid (536 g, 4.4 mol), sodium hydroxide (320 g, 8.0
mol), and enough water to make 3.8 L was prepared. A third solution was prepared by
diluting 1600 uL of a commercially available palladium stock solution (Aldrich 207349,
1.001 mg/ml in 5% aq. HCI, 1600 ug palladium, 15.0 umol, 0.0004 mol %) to 200 ml with
deionized water. The first two solutions were combined in the 9-liter reaction flask. The
lid was placed on the reaction and the fiber optic temperature probe was inserted into
the reaction mixture. The stirring paddle was fitted to the motor and the ejection tubing
was inserted into the reaction vessel. The reaction chamber was securely closed and
pre-pressurized to 280 psi with nitrogen. The reaction was heated to 150 °C using 7500
W (2500 W x 3), the ramp time taking approximately 13.3 min. At this point the
magnetron power was modulated to remain at the desired 150 °C (300W; 100 W x 3) for
an additional 5 min. At the end of the reaction, the reaction contents were ejected into a
receiving flask containing 2.0 kg ice. The white solid was collected via vacuum filtration,
rinsed with 2.0 L water and was dried for three days at 50 °C to yield 669 g of 4methoxybiphenyl, 4.2 (91.8%).
189
Preparation
of 2-(benzylthio)-4-chloro-6-methylpyrimidine
(4.12) A
1-L,
three-necked round bottom flask was charged with 232 g (1.00 mol) of S-benzyl-6methylthiouracil 4.11 and placed in a heating mantle. To the flask was added 333 ml
POCI3 (3.6 mol) and fitted with an addition funnel, an overhead stirring shaft, and a
thermometer. To the stirred solution, Et3N (102.0 g, 1.00 mol) was added drop-wise at a
rate to maintain approximately 80 °C, taking approximately 30 min. The heating mantle
was turned on and the reaction was heated to 100 °C and held at this temperature until
the reaction reached completion as indicated by TLC, approximately an additional 90
min. At this point, heating was discontinued and the reaction was allowed to cool to 50
°C. The reaction contents were slowly added portion wise to 1500 ml of a saturated
aqueous solution of NaHC0 3 which was simultaneously being stirred vigorously with an
overhead stirrer, adding additional NaHC0 3 as needed (upon cessation of C 0 2
generation). Approximately 2 kg of bicarbonate were needed to completely quench the
reaction contents and bring the pH=7. This aqueous layer was decanted from the solid
precipitate into a 4.0L separatory funnel. The precipitate was rinsed with -300 ml ethyl
acetate. The aqueous layer was extracted with 900 ml ethyl acetate in three 300-ml
portions. The organic extracts were combined, washed sequentially with 300 ml water
and 100 ml saturated sodium chloride, dried over MgS0 4 , and the solvent was removed
under reduced pressure to yield 219.5 g (87.5%) of 4.12 as a reddish brown oil of a
purity greater than 95% (1H NMR) that was used without further purification in the next
step.
Synthesis of 2-(benzylthio)-6-methyl-4-(phenylamino)pyrimidine*HCI
(4.13)
The 5 L reaction vessel was charged with 2-(benzylthio)-4-chloro-6-methylpyrimidine
4.12 (461 g, 1.838 mol), aniline (172 g, 1.84 mol), and acetic acid (110 g, 1.84 mol) in
dioxane (3.5 L solution, 0.53 M). The lid was placed on the reaction and the fiber optic
temperature probe was inserted into the reaction mixture. The stirring paddle was fitted
190
to the motor and the ejection tubing was inserted into the reaction vessel. The reaction
chamber was securely closed and pre-pressurized to 280 psi with nitrogen. The reaction
was heated to 150 °C using 7500 W (2500 W x 3), the ramp time taking approximately
11 min. At this point the magnetron power was modulated to remain at the desired 150
°C (300-600 W; 100-200 W x 3) for 10 min. After this time, microwave heating was
stopped and the solution was ejected into a receiving flask containing 2 L water, leaving
behind a spongy solid in the reaction flask. The solid was allowed to cool then filtered
under vacuum and dried overnight at 100 °C to yield the pale yellow solid (433 g,
68.4%). The aqueous solution was extracted using ethyl acetate (4 x 300 ml). The
organic extracts were combined washed with brine (100 ml), dried over MgS04 and the
solvent and residual 1,4-dioxane was evaporated under reduced pressure to yield an
additional 40.0 g pale yellow solid bringing the total yield of 2-(benzylthio)-chloro-6methyl-4-(phenylamino)pyrimidine to 473 g (74.7%). A small sample of this solid was
combined with 2.0 M NaOH (aq.) and extracted using ethyl acetate. The organic layer
was washed with brine, dried over MgS0 4 and the solvent was removed under vacuum
to afford the freebase of 4.13.
191
Chapter 5 Pd-Catalyzed Methodology Development: Synthesis of Diarylmethanes
a,a-bis(4-methoxyphenyl)acetophenone (5.2) To a 10-ml borosilicate glass
vial was placed a Teflon-coated stir-bar, tetrabutylammonium bromide (TBAB, 160 mg
0.50 mmol), acetophenone (122 mg, 1.0 mmol), 4-bromoanisole (392 mg, 2.1 mmol),
followed by addition of aqueous sodium hydroxide solution (2.0 M, 2.0 ml, 4 mmol). 50
uL (0.05 mol %) of the palladium stock solution was added, the vial was capped and
placed into a temperature-controlled oil bath set to 100 °C where it was allowed to stir for
20 h. At this point the reaction was stopped and allowed to cool to room temperature.
The reaction mixture was extracted with ethyl acetate ( 5 x 3 ml). Approximately 1 gram
of silica gel was added to the organic extract and the excess solvent was removed under
vacuum. Column chromatography of the crude product (95:5 pet ether:ethyl acetate)
yielded 314 mg (95%) of the product as a pale yellow oil.
Typical Procedure for the Synthesis of Diarylmethanes. All reactions with
deoxybenzoin to afford diarylmethanes (5.3-5.15) substituted on only one ring follow an
identical procedure for the synthesis of 4-methoxydiphenylmethane 5.3. All reactions
with deoxybenzoin to afford diarylmethanes substituted on both rings follow an identical
procedure for the synthesis of 4-methoxydiphenylmethane 5.3, but the reaction time is
extended to 60 minutes in the second leg of the reaction at 160°C.
(4-methoxylphenyl)phenylmethane (5.3). To a 10-ml borosilicate glass vial
was placed a Teflon-coated stir-bar and 160 mg (0.50 mmol) tetrabutylammonium
bromide (TBAB), deoxybenzoin (216 mg, 1.1 mmol), 4-bromoanisole (185 mg, 0.99
mmol), followed by addition of aqueous sodium hydroxide solution (3.0 M, 2.5 ml, 7
mmol). 106 uL (0.1 mol %) of the palladium stock solution was added, the vial was
capped and placed into the microwave cavity. The sample was heated to 130 °C and
held for 30 minutes, then ramped to 160 °C and held for an additional 30 minutes. At this
point the reaction was stopped and allowed to cool to room temperature. The reaction
192
mixture was extracted with ethyl acetate ( 5 x 3 ml). Approximately 1 gram of silica gel
was added to the organic extract and the excess solvent was removed under vacuum.
Column chromatography of the crude product yielded (99:1 pet ether: ethyl acetate) 175
mg (91%) product as a pale yellow oil.
Procedure for the formation of a-(4-methylphenyl)-a-phenylacetophenone (5.24).
Deoxybenzoin (50.0 mmol, 9.81 g), 4-bromotoluene (50.0 mmol, 8.55 g), TBAB (25
mmol, 8.0 g), and 100 ml 3.0 M K 2 C0 3 (aq) was added to a 250 ml round bottom flask.
The flask was placed in an oil bath set to 105 °C with vigorous magnetic stirring. After
approximately 10 minutes, PdCI2 (0.05 mmol, 8.8 mg, 0.1 mol %) was added to the
emulsion. A reflux condenser was placed on the reaction vessel, open to the
atmosphere. TLC showed complete conversion after approximately 9 hours. A crude 1 H
NMR of the reaction mixture showed complete conversion of starting materials to
product as evidenced by the singlet at -6.1 ppm for the product and no singlet apparent
at 4.25 ppm for the deoxybenzoin. The contents of the flask were transferred to a
separatory funnel and extracted with diethyl ether (3 * 80 ml). The organic extracts were
combined, washed sequentially with water (2 * 50 ml) then brine (50 ml), and dried over
MgS0 4 . The residual solvent was removed under vacuum to yield 14.1 g (98.6%) of a
grayish solid. The 1H NMR of this unpurified product was >95% pure. A recrystallization
from ethanol using activated charcoal removed all signs of palladium to yield 10.9 g
(76%) of a white crystalline material. A similar procedure was used to prepare 5.22,
5.23, and 5.25.
193
Appendix 8 •
Spectral Data for Thesis Compounds
General. Unless otherwise specified, all NMR spectra are recorded in CDCI 3 and
chemical shifts are calibrated to the residual CHCI 3 solvent peak, 5 = 7.26 ppm.
o
OX
3-acetylcoumarin (2.1) 300 MHz 1H NMR: 5 8.51 (s, 1H), 7.67 (m, 2H),
7.40 (m, 2H), 2.73 (s, 3H).
13
C NMR: 6 195.4, 159.2, 155.3, 147.4, 134.4, 130.2, 125.0,
124.5, 118.2, 116.7, 30.5. IR (dry film): 1740,1674, 1612, 1554, 1208,973,756 cm"1.
o
3-benzoylcoumarin (2.2). 300 MHz 1H NMR: 5 8.07 (s, 1H), 7.85 (d,
2H, J=7.7 Hz), 7.60 (m, 3H), 7.46 (t, 2H, J=7.9 Hz), 7.37 (d, 1H, J=7.7 Hz), 7.35 (d, 1H
J=7.7 Hz).
13
C NMR: 5 191.6, 158.4, 154.7, 145.4, 136.2, 133.8, 133.7, 129.6, 129.3,
128.6, 126.8, 125.0, 118.1, 116.8.
o
ox
OEt
coumarin-3-carboxylic acid ethyl ester (2.3). 400 MHz 1 H NMR: 5
8.44 (s, 1H), 7.57 (dt, 2H Ja= 7.6 Hz Jb=M
Hz), 7.27 (dd, 1H, Ja= 7.8 Hz Jb=1-0 Hz),
7.23 (dd, 1H, Ja=7.7 Hz J„=1.0 Hz), 4.32 (q, 2H, J=8A Hz), 1.32 (t, 3H, J=8.1 Hz),
o
o
— 3 ethyl-(Z)-2-acetyl-3-(2-methoxyphenyl)-acrylate
(Z-2.4) 1H NMR (300MHz,
CDCI3): 5 7.97 (s, 1H), 7.38 (t, 1H), 7.28 (d, 1H), 6.93 (m, 2H), 4.32 (q, 2H), 3.88 (s,
3H), 2.32 (s, 3H), 1.35 (t, 3H). 75 MHz
13
C NMR: 5 202.9, 164.8, 157.7, 136.8, 133.9,
131.9, 130.2, 122.3, 120.7, 110.9, 61.4, 55.3, 31.0, 14.2. EIMS: m/z (% base peak): 248
(10), 233 (12), 217 (100), 203 (15), 189 (20), 171 (18), 161 (18), 131 (20), 43 (20).
195
o
o
-OEt
^^OCH3
ethyl-(£)-2-acetyl-3-(2-methoxyphenyl)-acrylate
(E-2.4). 300 MHz
1
H
NMR (CDCI3): 5 7.96 (s, 1H), 7.40 (m, 2H), 6.94 (m, 2H), 4.30 (q, 2H), 3.89 (s, 3H), 2.45
(s, 3H), 1.24 (t, 3H). 75 MHz 13C NMR: 5 195.2, 167.8, 158.0, 137.3, 134.6, 132.2,
129.2, 122.3, 120.6, 110.8, 61.4, 55.6, 26.6, 13.9.
EtO^O
HN
^^OCH2CH3
NH
T
o
5-ethoxycarbonyl-6-methyl-4-(3'-methoxy-4'-ethoxy)phenyl-
3,4-dihydropyrimidin-2(1H)-one (2.5). 400 MHz 1H NMR (d6-DMSO): 5 9.12 (s, 1H),
7.64 (s, 1H), 6.83 (m, 2H), 6.69 (dd, 1H,\7a=8.3 Hz Jb=2.2 Hz), 5.08 (d, J=3.3 Hz), 3.97
(m, 4H), 3.70 (s, 3H), 2.23 (s, 3H), 1.28 (t, 3H, J=6.9 Hz), 1.11 (t, 3H, J=7A Hz). 100
MHz
13
C NMR: 6 166.9, 153.6, 150.0, 149.5, 148.6, 138.7, 119.1, 114.1, 111.8, 100.5,
64.6, 60.0, 56.2, 54.3, 18.3, 15.3, 14.7.
EtO^-O
uon3
^L/ OCH 3
hCyXyJ " ^ X ) C H
3
T0
5-ethoxycarbonyl-6-methyl-4-(2',3',4'-trimethoxy)phenyl-3,4-
dihydropyrimidin-2(1H)-one (2.6). 400 MHz 1H NMR (d6-DMSO): 5 9.15 (s, 1H), 7.67
(S..1H), 6.51 (s, 2H), 5.10 (s, 1H), 4.01 (q, 2H, J=7.0 Hz), 3.71 (s, 6H), 3.61 (s, 3H), 2.23
(s, 3H), 1.11 (t, 3H, J=7.0 Hz). 75 MHz 13C NMR: 5 165.9, 153.2, 148.9, 140.9, 137.3,
103.9,99.5,60.4,59.7,56.2,54.3,18.2,14.6.
EtO-^O
\
^^,OCH
h 1
3
WyKJ^
HN
NH
T
o
5-ethoxycarbonyl-6-methyl-4-(4'-methoxy)phenyl-3,4dihydropyrimidin-2(1H)-one (2.7). 300 MHz 1 H NMR (d6-DMSO): 9.16 (s, 1H), 7.68 (s,
196
1H), 7.15 (d, 2H, J= Hz), 6.88 (d, 2H, J=8.7 Hz), 5.10 (s, 1H), 3.98 (q, 2H, J=8.7 Hz),
3.72 (s, 3H), 2.25 (s, 3H), 1.10 (t, 3H, J=7.0 Hz). 75 MHz
13
C NMR: 5 165.8, 158.9,
152.6, 148.4, 137.5, 127.8, 114.2, 100.0, 59.6, 55.5, 53.8, 18.2, 14.6.
HN
NH
T
o
5-ethoxycarbonyl-6-methyl-4-(3',4'-methylenedioxy)phenyl-3,4-
dihydropyrimidin-2(1H)-one (2.8). 400 MHz 1H NMR (d6-DMSO): 5 9.18 (s, 1H), 7.69
(s, 1H), 6.84 (d, 1H, J=7.9 Hz), 6.75 (d, 1H, J=1.6 Hz), 6.68 (dd, 1H, Ja=7.9 Hz Jb=1.6
Hz), 5.98 (s, 2H), 5.08 (d, 1H, J=3.3 Hz), 3.99 (q, 2H, J=7.0 Hz), 2.25 (s, 3H), 1.12 (t,
3H, J=7.0 Hz). 100 MHz
13
C NMR: 5 165.8, 152.5, 148.7, 147.7, 146.8, 139.3, 119.8,
108.5, 107.1, 101.4, 99.8, 59.6, 54.1, 18.2, 14.5.
HN
T
NH
o
5-ethoxycarbonyl-6-methyl-4-(4'methyl)phenyl-3,4-
dihydropyrimidin-2(1H)-one (2.9). 400 MHz 1H NMR (d6-DMSO): 6 9.16 (s, 1H), 7.69
(s, 1H), 7.13 (apparent singlet, 4H), 5.12 (s, 1H), 3.98 (q, 2H, J=7A Hz), 2.26 (s, 3H),
2.25 (s, 3H), 1.11 (t, 3H, J=7.0 Hz). 100 MHz
13
C NMR: 5 166.0, 152.9, 148.8, 142.6,
137.0, 129.5, 126.8, 100.2, 59.8, 54.3, 21.3, 18.4, 14.7.
HN
NH
T
o
5-ethoxycarbonyl-6-methyl-4-phenyl-3,4-dihydropyrimidin-2(1H)-
one (2.10). 400 MHz 1H NMR (d6-DMSO): 5 9.19 (s, 1H), 7.74 (s, 1H), 7.35-7.34 (m,
5H), 5.17 (s, 1H), 3.99 (q, 2H, J=7.0 Hz), 2.26 (s, 3H), 1.09 (t, 3H, J=7.0 Hz). 100 MHz
13
C NMR: 5 165.8, 152.6, 148.8, 145.3, 128.8, 127.7, 126.7, 99:7, 59.6, 54.4, 18.2, 14.5.
197
HN.
Y
.NH
CH 3
O
5-ethoxycarbonyl-6-methyl-4-(2'-methyl)phenyl-3,4-
dihydropyrimidin-2(1H)-one (2.11). 300 MHz 1H NMR (d 6 -DMSO): 5 9.16 (s, 1H), 7.64
(s, 1H), 7.10-7.18 (m, 4H), 5.41 (s, 1H), 3.89 (q, 2H, J=7.0 Hz), 2.42, (s, 3H), 2.30 (s,
3H), 1.02 (t, 3H, J=7.0 Hz). 75 MHz
13
C NMR: 5 165.7, 152.0, 148.9, 143.7, 135.1,
130.5, 127.6, 127.0, 99.6, 59.5, 50.9, 19.1, 18.1, 14.4.
»3CyjyX^
HN
NH
T
°
5-ethoxycarbonyl-6-methyl-4-(4'-nitro)phenyl-3,4-
dihydropyrimidin-2(1H)-one (2.12). 400 MHz 1 H NMR (d 6 -DMSO): 6 9.35 (s, 1H), 8.22
(d, 2H, J=8.7 Hz), 7.89 (d, 1H, J=3.2 Hz), 7.50 (d, 2H, J=8.7 Hz), 5.28 (d, 1H, J=3.2 Hz),
4.00 (q, 2H, J=7A Hz), 3.32 (s, 3H), 2.27 (s, 3H), 1.10 (t, 3H, J=7.1 Hz). 100 MHz
13
C
NMR: 5 165.5, 152.5, 152.2, 149.8, 147.2, 128.1, 124.3, 98.7, 59.8, 54.2, 18.3, 14.5.
HN
T
NH
o
5-ethoxycarbonyl-6-methyl-4-(4'-chloro)phenyl-3,4-
dihydropyrimidin-2(1H)-one (2.13). 300 MHz 1 H NMR (d 6 -DMSO): 6 9.25 (s, 1H), 7.78
(s, 1H), 7.38 (d, 2H, J=8.5 Hz), 7.25 (d, 2H, J=8.5 Hz), 5.15 (s, 1H), 3.98 (q, 2H, J=7A
Hz), 2.25 (s, 3H), 1.11 (t, 3H, J=7.1 Hz). 75 MHz
13
C NMR: 5 165.7, 152.4, 149.2, 144.2,
132.2, 128.8, 128.6, 99.3, 59.7, 53.9, 18.3, 14.5.
HN
NH
CI
T
o
5-ethoxycarbonyl-6-methyl-4-(2'-chloro)phenyl-3,4dihydropyrimidin-2(1H)-one (2.14). 400 MHz 1 H NMR (d 6 -DMSO): 5 9.27 (s, 1H), 7.69
198
(s, 1H), 7.40 (d, 1H, J=8.3 Hz), 7.26-7.33 (m, 3H), 5,64 (s, 1H), 3.89 (q, 2H, J=7.1 Hz),
2.31 (s, 3H), 1.00 (t, 3H, J=7.1 Hz). 100 MHz
13
C NMR: 5 165.6, 152.0, 149.9, 142.4,
132.4, 130.0, 129.7, 129.4, 128.4, 98.6, 59.7, 52.2, 18.3, 14.6.
1,3-diphenylpropne-1-one,
chalcone (2.15) 300 MHz
8.05 (d, 2H, J=7.0 Hz), 7.84 (d, 1H, J= 15.7 Hz), 7.6-7.4 (m, 9H).
1
H NMR: 5
13
C NMR: 5 190.5,
144.8, 138.2, 134.9, 132.8, 130.5, 129.0, 128.6, 128.5, 128.46, 122.2.
o
^
c r
^
4'-chlorochalcone (2.16) 300 MHz 1H NMR: 5 7.94 (d, 2H, J=6.4
Hz), 7.78 (d, 1H, J=11.8 Hz), 7.61 (m, 2H), 7.50-7.40 (m, 6H).
13
C NMR: 5 189.1, 145.3,
139.2, 136.5, 134.7, 130.7, 130.0, 129.0, 128.95, 128.5, 121.5.
O
^ ^ c i
^ 2 ' - c h l o r o c h a l c o n e (2.17) 300 MHz 1H NMR: 5 7.6-7.3 (m, 10H), 7.15
(d, 1H, J=16.1 Hz). 13C NMR: 5 193.8, 146.3, 139.1, 134.4, 131.4, 131.3, 130.9, 130.3,
129.4, 129.0, 128.6, 126.9, 126.3.
o
ci4-chlorochalcone (2.18) 300 MHz 1H NMR: 5 8.03 (d, 2H, J=7.6),
7.75 (d, 1H, J=15.7 Hz), 7.6-7.45 (m, 6H), 7.39 (d, 2H, J=8.1 Hz).
13
C NMR: 5 190.1,
143.2, 138.0, 136.4, 133.4, 132.9, 129.6, 129.2, 128.7, 128.5, 122.5.
o
^
c i - ^ ^ - c h l o r o c h a l c o n e (2.19) 300 MHz 1H NMR: 5 8.20 (d, 1H, J=15.8 Hz),
8.03 (d, 2H, J=7.1 Hz), 7.70 (m, 1H), 7.6-7.4 (m, 5H), 7.30 (m, 2H).
13
C NMR: 5 190.3,
140.5, 137.9, 135.5, 133.2, 133.0, 131.2, 130.3, 128.7, 128.6, 127.8, 127.1, 124.8.
199
4'-phenylchalcone (2.20) 300 MHz 1 H NMR: 5 8.14 (d, 2H,
J=8.3 Hz), 7.88 (d, 1H, J=15.7 Hz), 7.8-7.4 (m, 13H).
13
C NMR: 5 189.9, 145.5, 144.7,
139.9, 136.9, 135.0, 130.6, 129.1, 129.0, 128.5, 128.2, 127.3, 122.1.
o
H3ccr ^
^
4'-methoxychalcone (2.21) 300 MHz 1H NMR: 5 8.01 (d, 2H,
J=6.6 Hz), 7.77 (d, 1H, J=11.7 Hz), 7.60 (m, 2H), 7.56 (d, 1H, J=11.7), 7.36 (m, 3H),
6.93 (d, 2H, J=6.5 Hz), 3.81 (s, 3H).
13
C NMR: 5 188.6, 163.5, 143.9, 135.1, 130.8,
130.3,128.9,121.9,113.9,55.5.
o
Haccr^
^ s S ^ O C H 3 4,4'-dimethoxychalcone (2.22) 300 MHz 1H NMR: 6 8.04
(dd, 2H, Ja=8.9 Hz, Jb=2.1 Hz), 7.79 (dd, 1H, Ja=15.6 Hz, 4=1.6 Hz), 7.61 (dd, 2H, Ja=
8.6 Hz Jb=3.4 Hz), 7.43 (dd, 1H, dd, 1H, Ja=15.6 Hz, 4=2.1 Hz), 7.0-6.9 (m, 4H), 3.87
(apparent dd, 6H, Ja=9.9 Hz, 4=4.9 Hz). 13C NMR: 5 188.7, 163.3, 161.5, 131.4, 130.7,
130.1, 127.8, 119.6, 114.4, 113.8, 55.5, 55.4.
o
F
^ V f l u o r o c h a l c o n e (2.23) 300 MHz 1H NMR: 5 8.07 (dd, 2H, 4=8.8
- k ^
Hz, Jb=5.5 Hz), 7.83 (d, 1H, J=15.7 Hz), 7.65 (m, 2H), 7.52 (d, 2H, J=15.7 Hz), 7.43 (m,
3H), 7.18 (t, 2H, 4=8.6 Hz). 13C NMR: 5 188.8, 165.6 (d, J=255 Hz), 145.0, 134.8, 134.5,
131.1 (d, J=9 Hz), 130.6, 129.0, 128.5, 115.9, 115.6.
o
^4'-methylchalcone (2.24) 400 MHz 1H NMR: 5 7.93 (d, 2H, J=8A
Hz), 7.80 (d, 1H,.J=15.8 Hz), 7.62 (m, 2H), 7.52 (d, 1H, J=15.8 Hz), 7.4 (m, 3H), 7.28 (d,
200
2H, J=8.1 Hz), 2.42 (s, 3H). 13C NMR: 5 190.0, 144.4, 143.6, 135.7, 135.0, 130.4, 129.4,
129.0, 128.7, 128.4, 122.1,21.7.
o
II
H3CO
6cH
3',4'-dimethoxychalcone (2.25) 300 MHz 1H NMR: 5 7.82 (d,
3
1H, J=15.6 Hz), 7.7 (m, 4H), 7.61 (d, 1H, J=15.6 Hz). 7.43 (m, 3H), 6.94 (d, 2H, J=8.4
Hz), 3.982/3.977 (overlapping singlets, 6H).
13
C NMR: 5 186.6, 153.3, 149.3, 144.0,
135.1, 131.3, 130.3, 128.9, 128.4, 123.0, 121.7, 110.9, 110.0, 56.1, 56.0.
o
ft
H3C0
S'-bromo^'-methoxychalcone (2.26) 300 MHz 1 H NMR: 5 8.27
Br
(d, 1H, J=2.1 Hz), 8.02 (dd, 1H, J a =8.6 Hz, Jb=2A Hz), 7.82 (d, 1H, J=15.6 Hz), 7.65 (m,
2H), 7.50 (d, 1H, J=15.6 Hz), 7.44 (m, 3H), 6.97 (d, 1H, J=8.6 Hz). 13 C NMR: 5 187.5,
159.5, 144.7, 134.8, 134.0, 132.1, 130.6, 129.7, 129.0, 128.5, 121.2, 112.0, 111.2, 56.5.
EIMS: m/z (% base peak): 316/318 (100), 237 (40), 213/215 (38), 165 (16), 131 (42),
103 (44), 77 (43), 63 (25), 32 (45).
CH 3 O
1
H3c
1
^
°
11
OCH3
~
3'-chloro-5'-methoxy-2',4'-dimethylchalcone
(2.27) 300 MHz
H NMR: 5 7.53 (m, 2H), 7.40 (m, 3H), 7.26 (d, 1H, J=16.2 Hz), 6.95 (d, 1H, J=16.2 Hz),
6.75 (s, 1H), 3.76 (s, 3H), 2.45 (s, 3H), 2.27 (s, 3H).
13
C NMR: 5 197.1, 154.5, 146.1,
138.0, 134.5, 134.1, 130.7, 128.9, 128.5, 128.4, 111.3, 55.9, 21.5, 17.4.
o'
3-phenyl-1-pyridin-(2-yl)-prop-2-ene-1-one
(2.28)
300
MHz
1
H
NMR: 5 8.74 (dd, 1H, Ja =4.7 Hz Jb =0.8 Hz), 8.32 (d, 1H, J=16.0 Hz), 8.19 (dd, 1H, Ja
=6.8 Hz Jb =0.8 Hz), 7.95 (d, 1H, J=16.0 Hz), 7.83 (dt, 1H, Ja =6.8 Hz Jb =0.8 Hz), 7.73
201
(m, 2H), 7.45 (m, 4H).
1M3 ,
C NMR: 5 (189.4, 154.2, 148.8, 144.7, 137.0, 135.2, 130.5,
128.9, 128.8, 126.9, 122.9, 120.9.
4'methyl-2-biphenylcarbonitrile (4.1) 300 MHz 1H NMR: 5 7.74-
CN
7.72 (m, 1H), 7.63-7.58 (m, 1H), 7.49-7.37 (m, 4H), 7.29 (d, 2H, J=8.0 Hz), 2.41 (s, 3H).
13
C NMR: 5 145.6 138.7, 135.3, 133.8, 132.8, 130.0, 129.5, 128.7, 127.3, 119.0, 111.2,
21.3. EIMS: m/z (% base peak): 193 (100), 177 (5), 165 (30), all remaining peaks have
relative peak intensities < 5%: 152. 140, 126, 113, 101, 95, 89, 82, 75, 63, 5 1 , 39.
H 3 CO'
4-methoxybiphenyl (4.2). 400 MHz 1H NMR: 5 7.58 (m, 4H), 7.45 (t,
^
2H), 7.37 (t, 1H), 7.03 (d, 2H), 3.90 (s, 3H). 100 MHz
13
C NMR: 5 159.4, 141.1, 134.0,
129.0, 128.4, 127.0, 126.9, 114.4, 55.6. EIMS: m/z (% base peak): 184 (100), 169 (48),
141 (44), 115(30).
o
-^Ny^o
3-acetylcoumarin (4.3) 300 MHz 1H NMR: 5 8.51 (s, 1H), 7.67 (m, 2H),
7.40 (m, 2H), 2.73 (s, 3H).
13
C NMR: 6 195.4, 159.2, 155.3, 147.4, 134.4, 130.2, 125.0,
124.5, 118.2, 116.7,30.5.
H3C
HNV.NH
Y
o
5-ethoxycarbonyl-6-methyl-4-phenyl-3,4-dihydropyrimidin-2(1 H)-
one (4.4). 400 MHz 1 H NMR (d 6 -DMSO): 5 9.19 (s, 1H), 7.74 (s, 1H), 7.35-7.34 (m, 5H),
5.17 (s, 1H), 3.99 (q, 2H, J=7.0 Hz), 2.26 (s, 3H), 1.09 (t, 3H, J=7.0 Hz). 100 MHz
13
C
NMR: 5 165.8, 152.6, 148.8, 145.3, 128.8, 127.7, 126.7, 99.7, 59.6, 54.4, 18.2, 14.5.
202
Allyl phenyl ether (4.5) 300 MHz 1H NMR: 5 7.32 (m, 2H, J=7.5 Hz), 6.95
(m, 3H, J=7.0 Hz), 6.12 (ddd, 1H), 5.45 (d, 1H, v7=15.8 Hz), 5.31 (d, 1H, J=10.5 Hz), 4.59
(2H, d, 5.6 Hz). 7 5 M H z
13
C NMR: 5 154.1, 136.2, 136.18, 127.8, 125.2, 120.8, 115.7,
116.3,35.3.
cr
6H
Benzyl-(4-chloro-3-methylphenyl)ether (4.6) 400 MHz 1 H NMR: 5
3
7.47-7.31 (m, 5H), 7.28 (d, 1H, J=8.7 Hz), 6.92 (d, 1H, J=2.7 Hz), 6.79 (dd, 1H, Ja=8.8
Hz, Jb=2.7 Hz), 5.07 (s, 2H), 2.40 (s, 3H). 100 MHz
13
C NMR: 5 157.6, 137.3, 137.0,
129.9, 128.9, 128.3, 127.7, 126.3, 117.8, 113.7, 70.4, 20.6.
jy°
Benzyl-(4-chlorophenyl)ether (4.7) 400 MHz 1H NMR: 5 7.50-7.34
C-^N^
(m, 5H), 7.30 (d, 2H, J=8.8 Hz), 6.96 (d, 2H, J=8.8 Hz), 5.09 (s, 2H). 100 MHz
13
C NMR:
5 157.6, 136.9, 129.6, 128.9, 128.4, 127.7, 126.1, 116.5, 70.6.
H
Qf^Y
0
Methyl A/-phenyl-3-aminopropanoate (4.8) 300 MHz 1 H NMR: 6
7.18 (t, 2H, J=7.5 Hz), 6.70 (t, 1H, J=7.3 Hz), 6.62 (d, 2H, J=7.7 Hz), 4.00 (bs, 1H), 3.70
(s, 3H), 3.46 (t, 2H, J = 6.4 Hz), 2.63 (t, 2H, J = 6.4 Hz). 75 MHz
13
C NMR: 6 172.8,
147.6, 129.3, 117.7, 113.0, 51.8, 39.4, 33.7.
^^^-"^>3-(2-hydroxyphenyl)-1-propene
(4.9) 300 MHz 1 H NMR: 5 1H NMR 7.09-
7.14 (m, 2H), 6.79-6.89 (m, 2H), 5.96-6.06 (m, 1H), 5.12-5.16 (m, 2H), 5.00 (s, 1H), 3.40
(d, 2H, J=6.3 Hz). 100 MHz
13
C NMR 5 35.0, 115.8, 116.4, 120.9, 125.3, 127.8, 130.4,
136.4,154.0.
203
HN
NH
6-methylthiouracil (4.10). 400 MHz 1H NMR: 5 12.2 (bs, 2H), 5.67 (s, 1H),
s
2.06 (s, 3H). 100 MHz
W
1J
C NMR: 5 176.3, 161.4, 153.5, 104.1, 18.5.
CH,
HN^N
S-benzylthiouracil (4.11). 400 M H z 1 H NMR: 5 7.20-7.43 (m, 5H), 5.99 (s,
SBn
0.86H), 5.39 (s, 0.14 H), 4.38 (s, 2H), 2.33 (s, 0.4 H), 2.21 (s, 2.6 H). 100 M H z 13 C N M R :
5 164.5, 164.4, 137.9, 129.5, 128.9, 128.9, 127.7, 107.2, 3 4 . 1 , 23.6.
CI
WCH3
Chloropyrimidine (4.12). 400 M H z 1 H N M R : 5 7.45 (d, 2 H , J=8.7 Hz),
SBn
7.28-7.33 (m, 3H), 6.87 (s, 1H), 4.42 (s, 2H), 2.46 (s, 3H). 100 M H z 13 C N M R : 6 172.2,
169.0,160.8,137.2,129.2,128.4,127.3,115.9,35.5,23.7.
H
SBn
2-(benzylthio)-6-methyi-4-(phenylamino)pyrimidine
(4.13)
400
MHz 1 H N M R : 5 7.17-7.45 (m, 10H), 6.73 (s, 1H), 6.27 (s, 1H), 4.43 (s, 2H), 2.33 (s, 3H).
100 M H z
13
C N M R : 5 170.7, 166.6, 160.8, 138.17, 138.12," 129.4, 129.0, 128.4, 127.0,
124.8, 122.5, 98.7, 3 5 . 1 , 2 4 . 1 . T O F H R M S m/z+1 calc. 308.1216, found 308.1227. IR:
3270, 3038, 1 6 2 1 , 1 5 7 1 , 1495, 1278, 756, 691 cm" 1 .
XCH3
a,a-bis(4-methoxyphenyl)acetophenone
(5.2) 400 M H z
1
H
NMR: 5 7.98 (d, 2 H , J=7.5 Hz), 7.49 (t, 1H, J=7.4 Hz), 7.39 (t,
J)
OCH
2 H , J=7.5 Hz), 7.17 (d, 4 H , J=8.6 Hz), 6.84 (d, 4 H , J=8.6 Hz),
3
5.95 (s, 1H), 3.76 (s, 6H). 100 MHz
13
C N M R : 5 198.7, 158.6,
136.9, 132.9, 131.6, 130.1, 128.9, 128.6, 114.2, 57.8, 55.2.
204
H 3 co' ^
^
Phenyl(4-methoxylphenyl)methane
(5.3) 1H NMR: 5 7.34 (m,
2H), 7.25 (m, 3H), 7.18 (d, 2H), 6.91 (d, 2H), 4.00 (s, 2H), 3.84 (s, 3H);
13
C NMR: 5
158.1, 141.6, 133.3, 129.9, 128.9, 128.5, 126.0, 114.0, 55.3, 41.1.
Diphenylmethane (5.4) 1 H NMR: 5 7.34 (m, 10H), 4.07 (s, 2H); 13C NMR:
141.2,129.0, 128.5,126.1,42.0.
F ^
^
(m, 2H);
13
(4-fluorophenyl)phenylmethane (5.5). 1H NMR: 6 7.30 (m, 7H), 7.07
C NMR: 5 163.1, 159.9, 136.9, 136.8, 130.4, 130.3, 128.9, 128.6, 126.3,
115.4, 115.1,41.1.
4-benzylbiphenyl (5.6) 1H NMR: 5 7.64-7.56 (m, 4H), 7.45 (t, 2H),
7.36-7.27 (m, 8H), 4.08 (s, 2H);
13
C NMR: 5 141.0, 140.3, 139.1, 129.4, 129.0, 128.7,
128.6,127.2,127.1,127.0,126.2,41.6.
\ ^
\ ^
1,3-diphenyl-1-propene, mixture of cis- and trans- (5.7) 1 H NMR: 5
7.35 (m, 10H), 6.7 (d, J=11.4, 0.07 H, cis-), 6.50 (m, 1.78H, trans), 5.97(dd, 0.07 H, cis),
3.79 (d, 0.14 H, J=7.5, cis-), 3.65 (d, 1.79 H, J=6.3, trans-);
13
C NMR: 5 140.3, 137.6,
131.2, 130.8, 130.1, 129.3, 128.8, 128.7, 128.6, 128.4, 128.3, 127.2, 126.9, 126.3,
126.2, 39.5(trans-), 34.8(c/s-).
H 3 CO v
0CH
3
(3,5-dimethoxyphenyl)phenylmethane (5.8) 1H NMR: 5 7.30 (m,
5H), 6.40 (m, 3H), 3.98 (s, 2H), 3.81 (s, 6H);
13
C NMR: 5 160.9, 143.5, 140.8, 128.9,
128.7, 126.2, 107.2, 98.1, 55.3, 42.2.
205
H3CO
6cH
3
(3,4-dimethoxyphenyl)phenylmethane
(5.9) 1 H NMR: 5 7.26, (m,
5H), 6.83 (d, 1H), 6.77 (m, 2H), 3.98 (s, 2H), 3.89 (s, 3H), 3.87 (s, 3H);
13
C NMR: 5
149.0, 147.5, 141.4, 133.7, 128.8, 128.5, 126.1, 120.9, 112.3, 111.3, 55.93, 55.83, 41.5.
J
1-(phenylmethyl)naphthalene (5.11) 1H NMR: 5 8.12 (m, 1H), 7.97 (m,
1H), 7.86 (d, 1H), 7.54 (m, 3H), 7.36 (m, 6H), 4.56 (s, 2H);
13
C NMR: 5 140.7, 136.7,
134.1, 132.3, 128.9, 128.8, 128.6, 127.4, 127.3, 126.2, 126.1, 126.0, 125.7, 124.4, 39.1.
c r
v^
4.01 (s, 2H);
F3C ^
^
13
(4-chlorophenyl)phenylmethane
(5.12) 1H NMR: 5 7.27 (m, 9H),
C NMR: 5 140.6, 139.6, 132.0, 130.3, 128.9, 128.6, 126.4, 41.3.
^
(4-trifluoromethylphenyl)phenylmethane
2H), 7.30 (m, 7H), 4.08 (s, 2H);
13
(5.13) 1H NMR: 6 7.59 (d,
C NMR: 5 145.3, 145.2, 140.0, 129.2, 129.0, 128.7,
126.5, 126.2, 125.5, 125.4, 125.39, 125.34, 122.5, 41.7.
H3C ^
^
(4-methylphenyl)phenyImethane (5.14 & A5.2) 1H NMR: 5 7.42 (m,
2H), 7.33 (m, 2H), 7.24 (s, 4H), 4.09 (s, 2H), 2.47 (s, 3H); 13C NMR: 141.6, 138.2, 135.6,
129.3, 129.0, 128.96, 128.6, 126.1, 41.7, 21.1. GCMS
•^ 1,4-dibenzylbenzene (5.15) 1H NMR: 5 7.30 (m, 4H), 7.23 (m, 6H),
7.16 (s, 4H), 3.99 (s, 4H). 13C NMR: 5 141.2, 138.9, 129.0, 128.9, 128.5, 126.0, 41.6.
H 3 C- >v^
^
- c , (4. met hy|phenyl)-4-chlorophenylmethane
(d, 2H), 7.16 (m, 6H), 3.97 (s, 2H), 2.40 (s, 3H)
(5.16) 1 H NMR. 5 7.30
13
C NMR: 5 140.0, 137.6, 135.9, 131.9,
130.2,129.3, 128.8, 128.6,40.9,21.1.
206
0CH
CI
(5.17) 1 H NMR: 5
3 (4-chlorophenyl)-4-methoxyphenylmethane
7.30 (d, 2H), 7.16 (m, 4H), 6.90 (d, 2H), 3.95 (s, 2H), 3.84 (s, 3H)
13
C NMR: 5 158.2,
140.2, 132.7, 131.8, 130.2, 129.9, 128.6, 114.0, 55.3, 40.4.
^^
H 3 CO"
^*"
^CH 3
(4-methylphenyl)-4-methoxyphenylmethane
5 7.18 (m, 6H), 6.90 (d, 2H), 3.98 (s, 2H), 3.85 (s, 3H), 2.41 (s, 3H)
13
(5.18) 1H NMR:
C NMR: 5 158.0,
138.6, 135.5, 133.6, 129.9, 129.2, 128.8, 113.9,55.3,40.7,21.1.
r
>*-
v ^ - C | (4. c hlorophenyl)-4-fluorophenylmethane
(5.19) 1H NMR: 5 7.28
13
(d, 2H), 7.13 (m, 4H), 7.01 (M, 2H) 3.95 (S, 2H)
C NMR: 5 162.8, 160.3, 139.4, 136.3,
132.1, 130.2, 128.7, 115.3,40.4.
cu
"^r
>r " ^
ci3,4'-dichlorodiphenylmethane (5.20) 1 H NMR: 5 7.2 (m, 8H), 3.94
(s, 2H)
1J
C NMR: 5 142.6, 138.7, 134.4, 132.3, 130.3, 129.8, 129.0, 128.8, 127.0, 126.6,
40.9.
4-(p-methylbenzyl)-biphenyl
(5.21)
1
H NMR: 5 7.65 (d, 2H),
7.63 (d, 2H), 7.58 (t, 2H), 7.46 (t, 1H), 7.39 (d, 2H), 7.19 (apparent singlet, 4H)
13
C NMR:
5 141.1, 140.6, 139.0, 138.0, 135.7, 129.3, 129.27, 128.9, 128.8, 127.3, 127.1, 127.07,
41.2,21.1.
o
^
a.a-diphenylacetophenone
(5.22) 400 MHz 1H NMR: 5 8.01 (d, 2H,
J=8.0 Hz), 7.45 (t, 1H, J=8.0 Hz), 7.38 (t, 2H, J=8.0 Hz), 7.3-7.1 (m,
10H), 6.01 (s, 1H). 100 MHz
13
C NMR: 5 198.6, 139.5, 137.2, 133.5,
129.6, 129.4, 129.2, 129.0, 127.6, 59.9. EIMS: m/z (% base peak): 272 (<2), 167 (34),
152(14), 105(100), 77(20).
207
OCH3
a-(4-methoxyphenyl)-a-phenylacetophenone
1
(5.23) 500 MHz
H NMR: 5 8.05 (dd, 2H, Ja=8.05 Hz, Jb=0.85 Hz), 7.53 (t, 1H,
J=8.0 Hz), 7.44 (t, 2H, J=7.95 Hz), 7.36 (t„ 2H, J=8.0 Hz), 7.30
(m, 5H), 6.90 (dd, 2H, Ja=6.7 Hz Jb=2.0 Hz), 6.04 (s, 1H), 3.80 (s, 3H). 125 MHz
13
C
NMR: 6 198.5, 158.7, 139.5, 136.9, 133.0, 129.1, 128.98, 128.72, 128.68, 128.67, 127.1,
114.2,58.6,55.2,30.9.
CH 3
a-(4-methylphenyl)-a-phenylacetophenone (5.24 & A5.1) 400
MHz 1H NMR: 6 8.10 (d, 2H, J=8A Hz), 7.6-7.2 (m, 12H), 6.11 (s,
1H), 2.39 (s, 3H). 100 MHz
13
C NMR: 5 198.4, 139.4, 136.9,
136.8, 136.1, 133.0, 129.5, 129.1, 129.02, 128.99, 128.7, 128.6, 127.1, 59.1, 21.1.
a-(3-methylphenyl)-a-phenylacetophenone
(5.25) 500 MHz
1
H
NMR: 5 8.05 (d, 2H, J=8.0 Hz), 7.53 (t, 1H, J=8.0 Hz), 7.44 (t, 2H,
J=8.4 Hz), 7.2-7.4 (m, 6H), 7.00-7.14 (m, 3H), 6.05 (s, 1H), 2.35 (s,
3H). 125 MHz
13
C NMR: 5 198.3, 139.2, 138.9, 138.4, 137.0, 133.0,
129.2, 129.0, 128.7,128.6, 128.0, 127.1, 126.2,59.4,21.5.
Figure A8-1. The "Ar-X" shelf in the Leadbeater Research Group laboratory.
About the Author •
209
Jason Schmink was born in 1978 in Pontiac, Illinois, the first-born son of P. David
and Melissa Y. Schmink. Jason's father was a public school administrator, having
earned a Bachelor's degree in Psychology, a Master's in Special Education, and finally
earning his Ed.D. in Education Administration. His mother, too, earned her fair share of
degrees, earning both a Bachelor's and a Master's in Sociology before switching fields
and earning a Bachelor's and Master's in Accounting. Jason reckons that between his
two parents, they had well over 50 credit hours of psychology courses, which have likely
contributed to his "interesting" mental and psychological attributes, i.e. he is odd.
Jason and his younger brother, Alex, grew up mainly in the small central Illinois
town of Monticello. Jason was a jack-of-all-trades and was active in a number of extracurricular activities in high school, including: being captain of the football team as an allconference offensive and defensive lineman, drum major of the state champion
Marching Sages marching band, and a member of the elite Madrigals singing group. Of
course, Jason loved to participate in the more 'nerdy' offerings his school had to offer
such as the scholastic bowl team, math team, and the Junior Engineering, Technology,
and Sciences (JETS) team.
After graduating from Monticello High School in 1996, Jason began his collegiate
career at the University of Illinois, Urbana-Champaign, majoring in instrumental music
education (trombone). After two years, however, he began to realize that music would
not satisfy his yearning to be creative: Jason wanted to make things! After considering
both culinary school and an education leaning toward architecture, he instead decided to
"get his hands dirty" and work full-time as an apprentice carpenter for a small, familyowned custom home construction company. During this time, he became quite good with
his hands and can tackle just about any project around the house with confidence!
However, after four hard years working in the sweltering August sun and the
bitter January wind, Jason decided that he should consider a career path that better took
210
advantage of his intellect as well as kept him comfortably inside a climate-controlled
building! Thinking back to the things he had enjoyed most in his education, he fondly
returned to the chemistry courses he had taken in high school where some of his fondest
memories involved sodium metal in water, boiling water + dry ice (it is vigorous, be
assured), burning magnesium, and numerous other laboratory 'exploits.'
Jason returned to the University of Illinois, and in December 2005 earned his
Bachelor's degree in Liberal Arts and Sciences. In January 2006 he began his doctoral
studies in chemistry at the University of Connecticut, working for Prof. Nicholas E.
Leadbeater in the area of microwave-assisted organic chemistry. Jason had a
successful tenure at UCONN. The accomplishments he is most proud of include: earning
the Waring scholarship for highest GPA for a chemistry or materials science graduate
student, being awarded the Outstanding Graduate Teaching Award by the Institute for
Teaching and Learning in 2008, being named the GSK Division of Organic Chemistry
Fellow in 2009-2010, and last but not least, becoming a rather good home brewer in his
free time!
Jason has accepted a post-doctoral fellowship that is a joint position between
Merck & Co. and the University of Pennsylvania where he will be working with Gary
Molander, Marissa Kozlowski, and Patrick Walsh, co-Pis. This appointment will begin
May 2010.
211
Документ
Категория
Без категории
Просмотров
0
Размер файла
10 569 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа