Full text of DOI:10.1016/j.tet.2018.04.019

Accepted Manuscript
Machine assisted reaction optimization: A self-optimizing reactor system for
continuous-flow photochemical reactions
K. Poscharny, D.C. Fabry, S. Heddrich, E. Sugiono, M.A. Liauw, M. Rueping
PII:
S0040-4020(18)30398-3
10.1016/j.tet.2018.04.019
TET 29434
DOI:
Reference:
To appear in: Tetrahedron
Received Date: 9 October 2017
Revised Date: 5 April 2018
Accepted Date: 6 April 2018
Please cite this article as: Poscharny K, Fabry DC, Heddrich S, Sugiono E, Liauw MA, Rueping
M, Machine assisted reaction optimization: A self-optimizing reactor system for continuous-flow
photochemical reactions, Tetrahedron (2018), doi: 10.1016/j.tet.2018.04.019.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical Abstract
To create your abstract, type over the instructions in the template box below.
Fonts or abstract dimensions should not be changed or altered.
Machine Assisted Reaction Optimization: A
Self-Optimizing Reactor System for
Leave this area blank for abstract info.
Continuous-Flow Photochemical Reactions
K. Poscharny,a D. C. Fabry,a S. Heddrich,c E. Sugiono,a M. A. Liauw,c M. Ruepinga,b
aRWTH Aachen University, Institute of Organic Chemistry, Landoltweg 1, 52074 Aachen, Germany
bKAUST Catalysis Center (KCC), King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900
cRWTH Aachen University, Institute of Technical and Macromolecular Chemistry, Worringerweg 1, 52074 Aachen, Germany

1
ACCEPTED MANUSCRIPT
Tetrahedron
journal homepage: www.elsevier.com
Machine Assisted Reaction Optimization: A self-optimizing reactor system for
continuous-flow photochemical reactions
K. Poscharny,a D. C. Fabry,a S. Heddrichc, E. Sugiono,a M. A. Liauw,c M. Ruepinga,b
aRWTH Aachen University, Institute of Organic Chemistry, Landoltweg 1, 52074 Aachen, Germany
bKAUST Catalysis Center (KCC), King Abdullah University of Science and Technology (KAUST), Thuwal 23955-6900
cRWTH Aachen University, Institute of Technical and Macromolecular Chemistry, Worringerweg 1, 52074 Aachen, Germany.
ARTICLE INFO
ABSTRACT
Article history:
Received
Received in revised form
Accepted
A methodology for the synthesis of oxetanes from benzophenone and furan derivatives is
presented. UV-light irradiation in batch and flow systems allowed the [2+2] cycloaddition
reaction to proceed and a broad range of oxetanes could be synthesized in manual and automated
fashion. The identification of high-yielding reaction parameters was achieved through a new
self-optimizing photoreactor system.
Available online
2009 Elsevier Ltd. All rights reserved
.
Keywords:
Self-optimization
Flow chemistry
photocyclization
FTIR spectroscopy
Photoreactions and flow chemistry have been sharing an ever-
growingly entangled history. Due to the use of high-power UV-
lamps, side-reactions and decompositions of both substrate and
product were frequently observed in batch-type reactor setups
that have been overcome by sophisticated flow-reactors.1 These
allowed the precise adjustments of irradiation time as well as
reaction temperature resulting in higher conversions and space-
time-yields. One of the earliest examples of photocyclization
reactions performed in microflow systems was published by
Fukuyama and Ryu in 2004.2 Vinyl esters and enones were
successfully coupled by use of a 300 W Hg lamp and a Foturan
glass reactor giving the corresponding cyclic products in
moderate to good yields. In 2013, Oelgemöller and coworkers
reported the [2+2]-cycloaddition reaction of furanones and
alkenes.3 Thereby, a commercially available Rayonet chamber
reactor in combination with a microcapillary unit was used for
UVC-light irradiation (λ=254 nm) with excellent conversions
compared to conventional batch-type setups.
self-optimization. Nevertheless, good yields were obtained in the
50-fold bigger system that was also in accordance with the
previous investigated setup. In 2011, Poliakoff and coworkers
established continuous etherification reactions in supercritical
carbon dioxide as green solvent.6 Thereby, a super modified
simplex algorithm was used for the iterative optimization of three
(temperature, pressure, flow rate) or four (temperature, pressure
flow rate, solvent ratio) parameter sets. The self-optimization
process was not only able to optimize the yield of the main
product but could also find conditions that led to the formation of
specific side products. However, optimizations of especially
photo-flow reactions still pose an insufficiently explored research
area of organic chemistry and chemical engineering. Radical
sensitivities of high-energy photocyclization reactions towards
temperatures and residence times let these reactions suffer from
time-consuming and labor-intensive optimizations. Furthermore,
concentrations of substrates in appropriate solvents used to be a
crucial parameter that often led to drastic improvements
concerning product yields and its purity.
A rather emerging field of research has been concerned with
the establishment of self-optimizing reactor systems that made
use of optimization algorithms, monitoring technologies, and
their integration in well-known reactor setups.4 Pioneering work
was achieved by Jensen and coworkers in 2010 when they
presented the first self-optimizing Heck-reaction in both micro-
and meso-scale continuous flow systems.5 Variations of
temperature and flow rate controlled by a Nelder-Mead simplex
algorithm led to isolation of an optimal parameter space for the
micro-scale reactor system. Those conditions were then
transferred to a meso-scale setup that was further used without
Among these reactions, the [2+2] photoreaction can be seen as
model reaction that was used in batch and flow reaction setups.7
The so-called Paternò-Büchi reaction of alkenes with
photochemically activated carbonyls was first described by their
inventors E. Paternó8 and G. H. Büchi9 and later on thoroughly
investigated by Schenk10 and Griesbeck.11 The importance of this
photochemical transformation was further proven by Schreiber in
1984,12 who selectively generated a multi-substituted oxetane
building block for the synthesis of the bioactive natural product -
(±)-Asteltoxin (Figure 1)

2
Tetrahedron
ACCEPTED MANUSCRIPT
Figure 1: Synthesis of Asteltoxin with
photocycloaddition key step.
a
[2+2]
Figure 2: Capillary-type continuous photo-flow reactor.
Herein, we report the advantage of using a capillary-type13
microfluidic photochemical reactor for Paternó-Büchi reactions
and a computer-assisted self-optimization of flow rates.
To maximize the productivity, the reaction time was evaluated
by changing either the reactor volume or the flow rate (Table 2).
After adjusting the reaction time manually from 50 to 120
minutes by looping the reaction mixture back into the reactor, it
was shown that the yield reached already excellent values after
100 minutes (Table 2, entry 4). Furthermore, an online-analysis
approach was developed to obtain the optimal residence time by
connecting an in-line ReactIR spectrometer and a computer based
communication interface to the aforementioned flow reactor
(Figure 3). Fortunately, applying the newly developed self-
optimizing reactor setup we observed a comparable yield of 97%
after 83 minutes (Table 2, entry 3).
The first investigations were carried out in order to determine
the most suited light source. A Rayonet chamber reactor was
therefore charged with lamps with emitting wavelength of 300,
350, and 400 nm, respectively. After 8 and 15 h reaction times
the conversion of the batch-mode Paternó-Büchi reaction of
benzophenone 2a and furan 1 was determined by GC-analysis
(Table 1).
Table 1: Evaluation of the light source.
Table 2: Evaluation of the flow rate and reaction time.
Entry
light (nm)
time (h)
2[b]
3a[b]
Conv. (%) Yield. (%)
Entry
Reactor
volume
(mL)
flow rate
(µL/min)
τ(total)
(min)
Yield
(%)
1
2
3
4
5
6
300
350
400
300
350
400
8
8
8
15
15
15
83
77
82
69
62
70
15
20
15
26
32
26
1
2
3
4
5
10
5
5
10
10
600
500
360
1200
500
17
2
14
8
69
80
97
96
98
[a] Reaction conditions: 2a (2.73 mmol), furan (20 mL). [b]
GC conversions/yields (%).
20
It was observed that formation of the product 3a at 350 nm
was slightly higher than at the other wavelengths (Table 1, entry
2, 5). This fact was confirmed by the known absorption spectrum
of benzophenone 2a in non-polar solvents, where the excitation
of the important nπ* shift is located at that wavelength.14
However, the usage of the photoreactor equipped with a powerful
medium-pressure Hg lamp and a Pyrex-glass filter, gave a
superior yield of 80% after 1 h reaction time (Figure 2).
[a] Reaction conditions: 2a (1.37 mmol), furan (10 mL). [b]
Yields after column chromatography (%).

3
been determined from the crude reaction mixture and turned out
ACCEPTED MANUSCRIPT
to be mostly excellent, the yield of the purified product was party
reduced. This observation was tracked back to the difficult
separation of the remaining substrate from the synthesized
product. In some cases, multiple purification procedures had to
be performed to obtain a good separation.
Table 3: Scope of the Paternó-Büchi reaction.a
Yield[b]
(%)
GC[c]
(%)
Entry
R1
R2 R3
3
1
2
3
4
5
6
7
8
H
H
H
H
H
H
H
H
H
H
H
H
Me
H
H
H
H
H
H
H
H
H
H
H
Ph
pTol
oTol
2,4-MePh
Mes
p(tBu)Ph
2-CN-Ph
pF-Ph
2,4-F-Ph
3-CF3-Ph
C2H4-Ph
3a
3b
3c
3d
3e
3f
3g
3h
3i
96
87
84
82
86
84
86
82
95
90
23[d]
97
98
94
98
95
90
94
96
95
95
98
28
98
Figure 3: Schematic illustration of the flow setup for the
[2+2] photocyclization reaction; BPR - back pressure
regulator.
During the measurement, the reaction mixture flow stream
passed the IR spectrometer before and right after being irradiated.
The recorded IR data were constantly analyzed by a computer
software which, based on these information, controlled the pump
unit. According to the intensity of the specific absorption of the
carbonyl group (1662 cm-1) of the benzophenone substrate 2a, the
flow rate of the operating HPLC-pump was constantly reduced
until the substrate was completely consumed in the photoreaction
(Figure 4).
9
10
11
12
13
3j
3k
3l
Ph Ph
Ph Ph
3m 77:23[d,e]
75:20[e]
[a] Reaction conditions: 2 (2.73 mmol), furan (20 mL), flow
rate 60 ꢀL/min; Vreactor 5 mL, τ = 83 min [b] Yields after column
chromatography. [c] GC conversions. [d] Calculated rate from
NMR. [e] Rate of substitution at position 1 and 3.
It was shown that both electron donating 3a-f and withdrawing
substituents 3g-j on the aromatic group of the carbonyl
compound were similarly suitable for the cyclization reaction.
Also the substitution pattern of the aromatic group had just a
minor influence on the product yield. Just in case of an aliphatic
carbonyl substituent a lower yield of 23% was isolated (Table 3,
entry 11). This observation can be explained by the application of
a Pyrex-filter and the insufficient irradiation with high energetic
UV-light with wavelength below 300 nm. In a similar reaction of
nonal and furan, Schreiber and Hoveyda15 used a Vycor-glas
batch reactor, which shows comparable transparency to UV-light
as fused silica. Upon performing the Paternó-Büchi reaction with
benzophenone 2a and 2-methylfuran 1b, isomeric products were
isolated. The introduced methyl-group could be located either at
the C-1 or the C-3 position. In conformity with the investigations
by the group of Abe16 the higher substituted oxetane was built
preferably with a 77 to 23 ratio of the isomers.
Figure 4: FTIR spectrum of the characteristic C=O band at
1662 cm-1 showing the conversion of benzophenone 2a.
To further prove the advantage of the developed flow system,
the reaction was also performed in a batch-type reactor. In this
direct comparison study it was clearly shown that under flow
conditions superior yields were obtained after a fraction of the
time which would be necessary under batch conditions (Table
SI1). After the same residence time of about 80 minutes the flow-
mode reaction showed full conversion of the carbonyl compound,
whereas under the batch-mode conditions only minor oxetane
formation could be observed (Table SI1, entries 1 and 5).
In order to make use of the previously mentioned self-
optimizations and our gained expertise in this field, we
envisioned the integration of an optimization algorithm into a
continuous-flow photoreactor. Again, using an in-line ReactIR,
we chose the Paternó-Büchi reaction of benzophenone with furan
as model reaction which is highly suitable due to the
characteristic IR-bands that can be analyzed by FTIR
spectroscopy. Moreover, identification of stationary states can
easily be achieved by this method. Using the IR software (iC
IR™)17 which simultaneously controls the IR machine,
interpretation of the IR spectra was possible. An HPLC pump
With the optimized flow parameters in hand the substrate
scope of the Paternó-Büchi reaction was explored. Applying
various aldehydes and a ketone to the [2+2] photoreaction with
furan and its derivative, the appropriate products were isolated in
good to excellent yields (Table 3). While the conversions have

4
Tetrahedron
(Knauer K-120), valve controls (Upchurch Scientific V-1341)
conversions when the flow rate of benzophenone 2a was
ACCEPTED MANUSCRIPT
and an autosampler (HiTec Zhang AutoSam200) complemented
our reactor setup. The Modified Simplex algorithm (MSIM)18
was integrated in LabVIEW 2010 (National Instruments) with
increased, higher yields were again obtained with an excess ratio
of furan 1a.
Matlab
7 (MathWorks), which controlled all equipment
components via RS-232 connections. By analyzing the obtained
parameter values based on the peak corresponding to the reaction
conversion with consecutive auto-export function, a real-time (by
the LabVIEW controlled) optimization program was established.
Having reached the steady state (after a minimum of three
hydrodynamic residence times and a standard deviation of peak
height difference below ≤ 7•10-4) of the system the next set of
parameters is calculated. Benzophenone dissolved in furan as
well as neat furan was pumped from a reservoir into the
photoreactor, previously mixed by a T-shaped micromixer for the
2-HPLC-pump system. Having passed the IR flow cell, a
computer assisted valve control allows collection of the product
in vessels, as all components and the resulting reaction
parameters were controlled, and optimized by the computer
software. Starting values were chosen within parameter ranges
that allowed efficient progress in the consecutive optimization
process.18
To further improve the presented self-optimizing reaction
system and analyze the influence of reactant dilutions, an
additional parameter should be installed. So benzophenone
dissolved in furan and neat furan were pumped individually into
the photochamber thus allowing the system to identify the best
concentration for both substrates. Therefore, two separately
controlled HPLC pumps were installed to the previous system
and mixed by a T-shaped mixer that finally gives a triangle with
two variables in the given parameter space (Figure 5).
Figure 6: Parameter space of self-optimized [2+2] photo-
cyclization
Table 4: Excerpt of values for self-optimization using two
HPLC-pumps and parameter space of self-optimized [2+2]
photo-cyclization reaction.
τ
Flow rate
BP
Flow rate
Furan
Conv.
point
(µLmin-1)
(µLmin-1)
(min)
63
(%)a
16
1
2
3
4
5
6
80
80
106
117
135
114
95
45
45
40
39
46
117
106
117
140
123
12
10
15
19
30
40
46
54
10
11
12
141
106
71
110
111
114
33
33
35
50
53
16
17
89
70
111
117
51
59
51
53
56
57
53
21
22
23
24
25
81
75
65
61
72
114
115
115
114
115
90
95
95
86
90
Figure 5: Self-optimizing photo-flow reaction setup with two
HPLC pumps and on-line ReactIR.
The parameter space was chosen to be 60 – 150 µLmin-1. This
area roughly covers the previously identified optimum with
regards to its possible variation towards lower concentrations that
have been observed in earlier experiments. After an initial
orientation phase of 10 points (Figure 6, Table 4, for full table
see SI), optimal regions towards low furan flow rates could be
identified, finally resulting in 95% conversions at flow rates of
65 µLmin-1 for benzophenone 2a and 115 µLmin-1 for furan 1a.
aConversions determined by GC; BP: benzophenone.
For parameter point 14 no clear IR spectrum was recorded due
to formation of a hot spot in the tube that led to evaporation of
furan. Nevertheless, the algorithm was able to continue its
optimization also showing the robustness of the optimization
method towards certain casual errors. Having kept the furan flow
rate constant and further lowering the amount of furan, the
algorithm was able to identify an optimal region for the reaction
Starting from 16% conversion for the first parameter point at
80 µLmin-1 for both components,
a
steady increase of
conversions can be observed. After receiving a block of lower

5
system after 25 parameter points. The total optimization was
achieved within 48 hours showing its high efficiency that could
not have been obtained without computer assistance (Table 5).
increasing the number of parameters, an even more detailed
ACCEPTED MANUSCRIPT
“map” of the reaction space could be drawn allowing the user to
choose from a variety of high-yielding parameter sets. The entire
optimization was achieved without human interaction as all
necessary manual interactions, such as probe collection,
recording of spectra and adjustment of flow rates were conducted
by integrated algorithms and control entities.
It becomes evident that there is a correlation between
conversion and benzophenone concentration. Usually, the lower
the benzophenone concentration, the higher the conversion. This
was either obtained by increasing the furan or by lowering the
benzophenone flow rate.
Furthermore, comparing the productivity of both optimal
parameter points from both flow setups, 0.37 mmolh-1 of product
3a for the 2-pump system and 0.35 mmolh-1 for the 1-pump
system could be obtained (according to eq. 1). At the same time,
a 36% reduction in irradiation time was achieved for the 2-pump
system (53 min) compared to the 1-pump system (83 min).
Acknowledgments
Technical assistance by Cornelia Vermeeren is very gratefully
acknowledged. The research leading to these results has received
funding from the European Research Council under the European
Union’s Seventh Framework Programme (FP/2007-2013)/ ERC
Grant Agreement No. 617044 (SunCatChem).
References
[1] a) Plutschack, M. B.; Pieber, B.; Gilmore, K.; Seeberger, P. H. Chem.
Rev. 2017, 117, 11796–11893; b) Lima, F.; Kabeshov, M. A.; Tran, D.
N.; Battilocchio, C.; Sedelmeier, J.; Sedelmeier, G.; Schenkel, B.; Ley,
S. V. Angew. Chem. Int. Ed. 2016, 55, 14085–14089; c) Rehm, T. H.
Chem. Eng. Technol. 2015, 39, 66–80; d) Su, Y. S.; Straathof, N. J. W.;
Hessel, V.; Noël, T. Chem. Eur. J. 2014, 20, 10562–10589; e) Garlets,
Z. J.; Nguyen, J. D.; Stephenson, C. J. Isr. J. Chem. 2014, 54, 351–
360; f) Oelgemoeller, M.; Chem. Eng. Technol. 2012, 35, 1144–1152;
g) Knowles, J. P.; Elliot, L. D.; Booker-Milburn, K. I. Beilstein J. Org.
Chem. 2012, 8, 2025–2052; h) Oelgemöller, M.; Shvydkiv, O.
Molecules 2011, 16, 7522–7550; i) Matsubara, H.; Hino, Y.; Tokizane,
M.; Ryu, I. Chem. Eng. J. 2011, 167, 567–571; j) Fuse, S.; Tanabe, N.;
Yoshida, M.; Yoshida, H.; Doi, T.; Takahashi, T. Chem. Comm. 2010,
46, 8722–8724; k) Fukuyama, T.; Rahman, T.; Sato, M.; Ryu, I. Synlett
2008, 2008, 151–163; l) Sugimoto, A.; Sumino, Y.; Takagi, M.;
Fukuyama, T.; Ryu, I.; Tetrahedron Lett. 2006, 47, 6197–6200; m)
Andre, J. C.; Viriot, M. L.; Villermaux, J. Pure Appl. Chem. 1986, 58,
907–916.
Table 5: Self-optimization run for photoflow reactor.
Point
Flow rate
BP
Flow rate
Furana
flow rate
total
Conv
(%)
τ
(µL/min)
(µL/min)
(µL/min)
(min)
1
2
3
4
80
117
106
117
65
80
160
223
223
252
180
63
45
45
40
56
16
12
10
15
95
106
117
135
115
opt
[2] Fukuyama, T.; Hino, Y.; Kamata, N.; Ryu, I. Chem. Lett. 2004, 33,
1430–1431.
ac(benzophenone) = 0.1 M in furan; BP: benzophenone;
conversion by GC
[3] Bachollet, S.; Terao, K.; Aida, S.; Nishiyama, Y.; Kakiuchi, K.;
Oelgemöller, M. Beilstein J. Org. Chem. 2013, 9, 2015–2021.
[4] a) Fitzpatrick, D. E.; Battilocchio, C.; Ley, S. V. ACS Cent. Sci. 2016,
2, 131–138; b) Fitzpatrick, D. E.; Battilocchio, C.; Ley S. V. Org.
Process. Res. Dev 2016, 20, 386–394; c) Fabry, D. C.; Sugiono, E.;
Rueping, M. React. Chem. Eng. 2016, 1, 129–133; d) Fabry, D. C.;
Sugiono, E.; Rueping, M. Isr. J. Chem. 2013, 54, 341–350.
[5] McMullen, J. P.; Stone, M. T.; Buchwald, S. L.; Jensen, K. F. Angew.
Chem. Int. Ed. 2010, 49, 7076–7080.
ꢒ
ꢀꢁꢂꢃꢄꢅꢆꢇꢈꢇꢆꢉ [ꢊꢊꢂꢋ ℎꢌꢍ] = ꢅꢎꢏꢐꢎꢑ
∙
ꢔꢕꢖꢔꢗ ∙ ꢛ
(eq. 1)
ꢎꢏꢐꢎꢑ
ꢒ
ꢗꢘꢗꢙꢚ
C
subst : concentration of substrate
ꢜ : flow rate
[6] a) Parrott, A. J.; Bourne, R. A.; Akien, G. R.; Irvine, D. J.; Poliakoff,
M. Angew. Chem. Int. Ed. 2011, 123, 3872–3876; b) Bourne, R. A.;
Skilton, R. A.; Parrott, A. J.; Irvine, D. J.; Poliakoff, M. Org. Process
Res. Dev. 2011, 932–938.
Y: yield
It cannot be excluded that other combinations of both flow
rates may lead to optimal conversions. Since the algorithm is
only capable of identifying a local optimum, not the global one,
increasing the length of the starting triangle or starting from a
different starting point may result in a different parameter set.
However, since the initial optimization already identified a local
optimum, we were only interested in the question whether a
modification of setup and a more specified performance could
give rise to a more detailed and advanced optimization method.
As both reactor setups identified similar optima independently,
verification of both methods can be assumed. Nevertheless,
reasonable amounts of these biologically active motifs can now
be easily prepared using readily available hardware.
[7] a) Fréneau, M.; Hoffmann, N. J. Photochem. Photobiol. C: Photochem.
Rev. 2017, 33, 83-108; b) Ralph, M.; Ng, S.; Booker-Milburn, K. I.
Org. Lett., 2016, 18, 968–971; c) Nakano, M.; Nishiyama, Y.;
Tanimoto, H.;,Morimoto, T.; Kakiuchi, K. Org. Process Res. Dev.,
2016, 20, 1626–1632; d) Terao, K.; Nishiyama, Y.; Kakiuchi, K. J.
Flow Chem. 2014, 4, 35–39; e) Yavorskyy, A.; Shvydkiv, O.;
Hoffmann, N.; Nolan, K.; Oelgemoeller, M. Org. Lett., 2012, 14,
4342–4345; f) Fukuyama, T.; Kajihara, Y.; Hino, Y.; Ryu, I. J. Flow.
Chem. 2011, 1, 40–45.; g) The [2+2]-photocycloaddition in an
automated and computer controlled microreaction plant has been
reported: Freitag, A.; Dietrich, T.; Scholz, R. Proceedings of the 11th
International Conference on Microreaction Technology (IMRET 11),
Kyoto, Japan, 8–10 March 2010; pp. 24–25.
[8] Paternó, E.; Chieffi, G. Gazz. Chim. Ital. 1909, 39, 341.
[9] Büchi, G.; Inman, C. G.; Lipinsky, E. S. J. Am. Chem. Soc. 1954, 76,
4327–4331.
[10] Schenck, G. O.; Hartmann, W.; Steinmetz, R. Chem. Ber. 1963, 96,
498–508.
In summary, we have illustrated the development of a [2+2]
photocyclization reaction of benzophenone and furan yielding
biologically active and important structural motifs tolerating a
broad scope of different educts. Starting from batch-type reactor
setups, modern flow-setups in combination with advanced self-
optimization methods led to identification of parameter sets
which show excellent conversions and yields for the desired
oxetanes. By changing to a 2-reservoir reactor system and thus
[11] a) Griesbeck, A. G.; Bondock, S.; Cygon, P. J. Am. Chem. Soc. 2003,
125, 9016–9017; b) Griesbeck, A. G.; Bondock, S.; Gudipati, M. S.
Angew. Chem. Int. Ed. 2001, 40, 4684–4687; c) Griesbeck, A. G.;
Buhr, S.; Fiege, M.; Schmickler, H.; Lex, J. J. Org. Chem. 1998, 63,
3847–3854; d) Griesbeck, A. G.; Stadtmüller, S. Chem. Ber. 1990, 123,
357–362.
[12] Schreiber, S. L.; Satake, K. J. Am. Chem. Soc. 1984, 106, 4186–4188.

6
Tetrahedron
ACCEPTED MANUSCRIPT
[13] Hook, B. D. A.; Dohle, W.; Hirst, P. R.; Pickworth, M.; Berry M. B.;
Booker-Milburn, K. I. J. Org. Chem. 2005, 70, 7558–7564.
[14] Castro, G. T.; Blanco, S. E.; Giordano, O. S. Molecules 2000, 5, 424–
425.
[15] Schreiber, S. L.; Hoveyda, A. H. J. Am. Chem. Soc. 1984, 106, 7200–
7202.
[16] Abe, M.; Kawakami, T.; Ohata, S.; Nozaki, K.; Nojima, M. J. Am.
Chem. Soc. 2004, 126, 2838–2846.
[17] http://us.mt.com/us/en/home/products/L1_AutochemProducts/L2_in-
situSpectrocopy.html (11.07.2017).
[18] Morgan, E.; Burton, K. W.; Nickless, G.; Chemom. Intell. Lab. Syst.
1990, 7, 209–222.