A Continuous Stirred-Tank Reactor (CSTR) Cascade for Handling Solid-Containing Photochemical Reactions

Article abstract of DOI:10.1021/acs.oprd.9b00378

Visible-light photoredox reactions have been demonstrated to be powerful synthetic tools to access pharmaceutically relevant compounds. However, many photoredox reactions involve insoluble starting materials or products that complicate the use of continuo

Full text of DOI:10.1021/acs.oprd.9b00378

Subscriber access provided by WESTERN SYDNEY U
Full Paper
A Continuous Stirred-Tank Reactor (CSTR) Cascade for
Handling Solid-Containing Photochemical Reactions
Alexander Pomberger, Yiming Mo, Kakasaheb Y Nandiwale, Victor L. Schultz,
Rohit Duvadie, Richard I. Robinson, Erhan I Altinoglu, and Klavs F. Jensen
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.9b00378 • Publication Date (Web): 05 Nov 2019
Downloaded from pubs.acs.org on November 5, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 19
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
A Continuous Stirred-Tank Reactor (CSTR) Cascade for Handling Solid-
Containing Photochemical Reactions
1
1
1
1
Alexander Pomberger , Yiming Mo , Kakasaheb Y. Nandiwale , Victor L. Schultz , Rohit
2
2
3
1
Duvadie , Richard I. Robinson , Erhan I. Altinoglu and Klavs F. Jensen *
1Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA
02139, United States
2Global Discovery Chemistry - Chemical Technology Group, Novartis Institutes for BioMedical
Research, 250 Massachusetts Avenue, Cambridge, MA 02139, United States
3Chemical and Pharmaceutical Profiling, Novartis Global Drug Development, 700 Main Street
South, Cambridge, MA 02139, United States
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 19
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table of Contents
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
For Table of Contents Only)
2
ACS Paragon Plus Environment

Page 3 of 19
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Abstract
Visible-light photoredox reactions have been demonstrated to be powerful synthetic tools to access
pharmaceutically relevant compounds. However, many photoredox reactions involve insoluble
starting materials or products that complicate the use of continuous flow methods. By integrating
a new solid-feeding strategy and a continuous stirred-tank reactor (CSTR) cascade, we realize a
new solid-handling platform for conducting heterogeneous photoredox reactions in flow.
Residence time distributions for single phase and solid particles characterize the hydrodynamics
of the heterogeneous flow in the CSTR cascade. Silyl radical-mediated metallaphotoredox cross-
electrophile coupling reactions with an inorganic base as the insoluble starting material
demonstrate the use of the platform. Gram-scale synthesis is achieved in 13 hours of stable
operation.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Keywords: photoredox catalysis・solid handling ・ flow chemistry ・ multiphase flow systems
3
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 19
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Introduction
Over the past decade, rapid developments in the field of visible-light photoredox catalysis has
enabled an ever-increasing range of synthetically useful transformations under mild conditions.
Utilization of either organometallic or organo photocatalysis opens up safer processing operation
using narrow band visible light sources, when compared to the application of high energy broad
band UV irradiation. Additional benefits are often realized with respect to both substrate and
product stability, particularly in structurally complex molecules such as natural products. Indeed
photoredox reactions increasingly play a strategic role in the selective late stage functionalization
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
of complex pharmaceutically-relevant molecules. In addition, the rapidly expanding area of dual
photoredox catalysis enables access to new reaction pathways by merging photoredox catalysts
3
with traditional metal catalyzed coupling manifolds, such as palladium (C−H arylation) , copper
3
4
3
2
5, 6
(
sp C–N coupling) or a nickel catalyst (C(sp )−C(sp ) coupling).
Continuous flow chemistry has proven to be a reliable and efficient method in organic synthesis,
and features several advantages over traditional batch chemistry, such as enhanced mass transfer,
improved temperature control, easy automation, steady state production, and facile reaction
7
-9
optimization. However, due to the photon flux limited character of many photochemical
reactions, the scaling up of photochemical process is still challenging even with the continuous
flow synthesis.
10, 11
Depending on reactor design, efficiency of light capture, and type of reaction
(e.g. photocatalysis vs direct irradiation), continuous reactors can greatly benefit from increased
photon flux per reactor volume, leading to improved productivity, particularly in a standard
laboratory setting.
1
2-15
It should however be noted that when comparing experimental setup for
scaling photochemical reactions, the assessment of continuous flow methods verses single batch
process, requires careful consideration.16
A significant limitation in the application of continuous flow reactors towards a wide variety of
useful batch photochemical transformations reported in the literature (particularly in plug flow
reactors - PFR), is the usage or generation of solids during the reaction. This often requires to
change or re-optimizethe original batch conditions in order to achieve a homogeneous system that
is amenable to the continuous flow process. These re-optimization steps not only require tedious
efforts in solubility screening (substrate, by-products and reagents), but also could lead to altered
selectivity of the chemical reaction due to the application or combination of less effective reagents
or solvents. For example, a heterogeneous inorganic base is often replaced by a less compatible
and more expensive homogeneous organic base.11, 17, 18
Controlled handling of heterogeneous reaction mixtures in a continuous flow set-up remains a
major challenge. Recently, a miniaturized cascade of continuously stirred-tank reactors (CSTR)
19
2
0
was proposed as one possible solution for handling reactions involving solid components. The
reliable solid-handling performance of CSTRs was also demonstrated in a multiphase reaction
where a suspension of Pd/C was pumped into the CSTR via a syringe pump equipped with a stir
2
1
bar. Moreover, CSTRs have also been used for mass transfer limited reactions such as biphasic
2
2
Suzuki-Miyaura cross coupling reactions. A recent example of a single stage laser-driven photo
CSTR has been disclosed by a group at Abbvie, capable of delivering multi-kg productivity,
through careful consideration of light attenuation, reactor geometry and catalyst concentration to
flexibly adapt to a given process constraint (in this case applied to homogenous systems).11
4
ACS Paragon Plus Environment

Page 5 of 19
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Herein we seek to address some of the solid handling limitations described above through the
development of a new photo CSTR cascade reactor in combination with a slurry-feeding pump for
conducting heterogeneous visible-light photochemistry in continuous flow. Residence time
distributions (RTDs) for liquid and liquid-solid systems are measured, and the absorbed photon
flux of the reactor is quantified by chemical actinometry. The capability of the platform is
demonstrated with the silyl radical-mediated metallaphotoredox cross-electrophile coupling, a
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
23
photochemical reaction involving an inorganic base as the insoluble reagent. A reaction self-
terminating problem is addressed by solvent and reactor geometry optimization. Finally, gram-
scale synthesis achieved in 13 hours of continuous production demonstrates the stability of the
system.
Results and Discussions
Reactor design and slurry-pump design
A CSTR cascade, consisting of five single chambers for heterogeneous photochemical reactions
2
0
was designed and built based on our previous work (Figure 1a-b). Two CSTR cascades with
different inner diameters (ID) of the connecting tubes were constructed. The first CSTR cascade
(CSTR 1) with 2.5 mm ID connection tubes has good solid-handling capability at expense of back-
mixing between chambers. On the other hand, the second reactor (CSTR 2) has narrower inter
connecting tubes (0.7 mm ID) with reduced back-mixing, but increased risk of plugging.
A slurry pump was engineered to controllably deliver slurry into the CSTR cascade. This pump
consists of a stainless-steel cylinder, in which an inert hydraulic liquid drives the delivery of the
slurry by a piston separating the two fluids (Figure 1c). The slurry is kept well suspended using a
cross-shaped magnetic stir bar. The flowrate of the slurry output equals the flowrate of the inert
hydraulic liquid pumped by a positive displacement pump. The accuracy and stability of slurry
delivering was validated with a suspension of Na CO (8.3 % w/t) in DMF (see Figure S7 in
2
3
Supporting Information).
In order to allow facile translation of heterogeneous photochemical batch conditions into a
continuous flow process, we designed the platform where all starting materials can be fed into the
photo-CSTR via the slurry pump (Figure 2a). With the intention to use gravity to enhance solid
transport and minimize clogging, the slurry pump is positioned ~300 mm above the CSTR. The
system is pressurized via a pressurized fraction collector, where 10 psig argon is applied directly
to the collection chamber to prevent gas evolution in the reactor (Figure 2b).
5
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 19
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1 (a) Exploded-view computer aided design (CAD) model of the CSTR cascade for
handling solid-containing photochemical reactions. From top to bottom, it sequentially contains
the upper aluminum cover, Pyrex glass window, second aluminum cover, polytetrafluoroethylene
(PTFE) reactor block containing five CSTR chambers, and the aluminum housing (see Supporting
Information Figures S1 and S2 with related text) (b) Picture of the assembled CSTR cascade. (c)
CAD model of the slurry pump. (d) Picture of the assembled slurry pump in operation mode. The
outflow is equipped with a vibration motor to enhance solids flow.
Figure 2 (a) Schematic of the photo-CSTR platform including the slurry pump, CSTR cascade, 3
LED light sources (440 nm), focusing lenses and the pressurized fraction collector (for a picture
of the setup see SI Figure S1) (b) Image of the lamps and reactor.
6
ACS Paragon Plus Environment

Page 7 of 19
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Residence time distribution (RTD) experiments
Liquid phase RTD. In order to gain knowledge about the mixing characteristics of CSTR cascade
systems, residence time distribution (RTD) measurements were performed by using the pulse
injection method. Deconvolution of the reactor RTD from the inlet and outlet concentration
profiles was done by regression with the exponentially modified Gaussian (EMG) distribution
2
4
model (Eqn. 1, see Figures S11-S16 and associated text in Supporting Information).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
휆
2
휆
2
휇 + 휆훿 ― 푡휆
2
퐸(푡) = exp (
(
2휂 + 휆훿 ― 2푡
)
푒푟푓푐(
)
(1)
2휎
The first reactor that can handle a very large amount of solids (CSTR 1) shows non-ideal CSTR
cascade behavior as a result of severe back-mixing (Figure 3a). In contrast, the RTD of CSTR 2 is
2
5
close to the RTD of an ideal CSTR in series model (Eqn. 2) (Figure 3b). The RTD of CSTR 2 at
three different flow rates can be found in Figure S10b. Consequently, narrow connecting tubes are
desirable as long as the amount and size of solid particles are sufficiently small to avoid clogging.
Otherwise, the design with wide connecting tubes needs to be used with the downside of a less
desirable RTD profile.
푡푛 ― 1
(푛 ― 1)!휏푛
퐸(푡) = (
)푒 ―푡/휏
(2)
Figure 3 (a) RTD of CSTR 1 under three different flowrates (0.1, 0.5, 2 mL/min) and comparison
with ideal CSTR in series model, (b) RTD comparison of CSTR 1, CSTR 2, and ideal CSTR in
series model at 0.5 mL/min.
Solids flow RTD. Translating heterogeneous reactions from batch to continuous flow requires the
understanding of the RTD of each phase in the continuous reactor. For example, insufficient
movement of an insoluble base through a reactor could lead to deviation from the desired base
concentration in the reactor. As demonstrated in the previous section, the liquid RTD can be
controllably tuned with the reactor geometry. In order to determine the RTD of the solid particles,
19
we injected a suspension of K [Fe(CN) ] in acetonitrile into the system as a tracer via a six-way
3
6
valve. After the tracer suspension passed through the reactor, addition of a stream of water
7
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 19
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
dissolved the tracer particles with aid of sonication, giving a homogeneous solution for in-line
UV/Vis spectroscopy (
Figure 4). Two different particle sizes were investigated (Figure 5b) to explore whether size
influenced the RTD.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4 Schematic for solids flow RTD measurement: injection of the solid tracer via six-way
valve, in-line dissolving of the particles (using a sonication bath) after the CSTR using water, and
in-line monitoring with UV/Vis spectroscopy.
In CSTR 2, the tracer with two particle sizes shows a similar RTD profile to RTD of the liquid
phase, reflecting almost ideal CSTR in series behavior (Figure 5a). Similarity of the two phases
suggests that accumulation and local excess of a solid component can be avoided, which helps to
accurately control the local stoichiometry of the solid reagents in the reactor. The tracer particles
are useful as long as their size and flow characteristics are similar to those expected of the intended
solid reagents.
Figure 5 (a) Comparison of the RTD of a solid tracer vs. liquid tracer at 0.5 mL/min using CSTR
2
(b) Particle size distribution of the used solid tracer K [Fe(CN) ].
3
6
8
ACS Paragon Plus Environment

Page 9 of 19
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Chemical Actinometry
Photon flux (photons per unit time), a key parameter of photochemical reactor, was characterized
with chemical actinometry experiments. We used the aerobic oxidation reaction of 9,10-
diphenylanthracene 1 (DPA) to the corresponding peroxide 2, catalyzed by Ru(bpy )Cl as a model
reaction (Scheme 1). With the known quantum yield of this reaction, the absorbed photon flux
can be calculated from the amount of consumed starting material.
3
2
2
6
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 1 Oxidation of 9,10-diphenylantracene (DPA) 1 to the corresponding endoperoxide 2 as
a chemical actinometer.
As shown in Figure 6a, between the highest light blue line (no conversion) and the lowest orange
line (full conversion), the inner four dark blue lines represent different levels of partial conversion.
The decreasing peak at 373 nm (see arrow) indicates the increasing conversion of 1 with longer
residence time. Since both LED intensity and residence time influence the conversion of DPA
simultaneously, we performed experiments varying both parameters (Figure 6b). The correlation
between the residence time and the conversion is almost linear. In addition, turning-up input power
of the blue LED (see SI for explanation of the % input power) light source linearly increases the
photon flux into the CSTR cascade, indicating light efficiency of the light source remains as a
constant over a wide range of power input (Figure 6c). We compared chemical actinometry results
of the CSTR cascade and PFR (0.76 mm inner diameter, FEP tubing, positioned on an aluminum
plate, for details see SI Figure S9d and the corresponding paragraph). Remarkably, even though
the PFR has a 13 times higher surface (perpendicular to the light source) to volume ratio than the
CSTR, photon flux into the PFR is only 3.5 times higher (Figure 6d), which implies there is a
reasonable amount of light penetrated and absorbed inside the CSTR chamber compensating the
surface to volume ratio difference between PFR and CSTR. The photon flux per reactor volume
was compared since this ratio would determine the productivity per reactor volume. A simplified
mathematical model based Lambert-Beer Law was built to explain this phenomena in Supporting
11
Information.
9
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 19
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 6 (a) Decreasing spectrum of DPA at 10% LED power: starting material (light blue line),
four different residence times 0.5-1 mL/min (dark blue lines) and full conversion (orange line).
The arrow indicates the decreasing spectrum of DPA (Conc of DPA from 100 % to 0 %) (b) DPA
concentration at three different LED powers and four different residence times (c) Plot of the
absorbed photon flux vs. variation of the LED power (d) Comparison of PFR vs CSTR –
comparison of the absorbed photon flux and the surface to volume ratio.
Metallaphotoredox cross-electrophile coupling in the photo-CSTR cascade platform
Recently developed silyl radical-mediated metallaphotoredox cross-electrophile coupling reaction
represents an attractive pathway for the union of two bench-stable electrophiles (e.g. an alkyl and
aryl halide electrophile) under mild conditions. Photocatalysis offers practical advantages such as
easy modulation of the oxidation states of nickel as well as the generation of reactive radical
23
species which can further interface with nickel catalytic cycles. Due to the increasing interest in
the pharmaceutical industry and the heterogeneous nature of the chemistry, we chose to use this
reaction to demonstrate versatility of our newly developed photo-CSTR platform (Scheme 2).
10
ACS Paragon Plus Environment

Page 11 of 19
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 2. Cross coupling of 4-bromotetrahydropyran 3 and methyl-4-bromobenzoate 4 is
performed in flow using photocatalyst 6, Ni catalyst 7, TTMSS, Na CO and blue light-emitting
2
3
diodes (440 nm) at 35° C in dry DME/DMA under exclusion of oxygen.
We started the development of the process based on optimized conditions for the same substrates
but with a homogeneous organic base instead of Na CO . Following successful batch experiments,
2
3
we adopted a solvent combination of N,N-dimethylacetamide (DMA) and 1,2-dimethoxyethane
(DME) (DMA/DME 3:1). We started the investigation using CSTR 1 since it proved to be reliable
in terms of handling large amounts of solids (tested up to 8.3 % wt.) (Figure 7a). All reagents
except for Na CO were dissolved entirely in a mixture of DMA:DME and fed to the CSTR as a
2
3
homogeneous solution via a syringe pump. The inorganic base was fed as a suspension in DMA
via the slurry pump (Figure 7b). Samples of the outflow were collected over time. Surprisingly,
yield and conversion dropped significantly after three residence times (Figure 7c) even though
batch experiments with exactly the same stock solution led to full conversion and 77% yield for
the same reaction time and light source set-up.
Figure 7 (a) Silyl radical-mediated metallaphotoredox cross-electrophile coupling of 3 and 4 using
a solvent mixture of DMA:DME 3:1 (b) Flowchart of the experiment: The inorganic base is
pumped as a suspension in DMA via the slurry pump, and other reagents are pumped as a
11
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 19
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
homogeneous solution via a syringe pump (c) Reaction yield and conversion over time in
continuous flow. Each data point represents average yield and conversion of the outflow during
the collection time.
We hypothesized that this unexpected decrease in performance could be caused by the formation
of side products potentially inhibiting further reactions. A broad residence time distribution in
continuous reactors could potentially lead to accumulation of generated reaction inhibiting species,
preventing the fresh reagents from reacting. As shown in the RTD results of CSTR 1 (Figure 3a),
the difference of RTD profiles between the flow rates at 0.1 mL/min and 0.5 mL/min is relatively
small. Hence, the RTD of actual applied flow rate of 0.176 mL/min can be expected to lie between
the RTD curves for 0.1 mL/min and 0.5 mL/min. thus can be expected to be non-ideal. Narrowing
the RTD, hence using CSTR 2 with close-to-ideal CSTR in series RTD, would mitigate such a
self-inhibiting effect (Figure 3b). SEM analysis of the precipitated NaBr along with the excess
reagent Na CO in the reaction crude showed an average size of d =15.7 µm (d =7.6 µm d =15.7
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
3
50
10
50
µm d =26.5 µm) (Figure 8b). Within batch experiments (see Supporting Information) it was
9
0
demonstrated that NaBr does not lead to self-inhibition. The mechanism of the inhibition
mechanism is under investigation. Additionally, even though the ID of the connection tube was
reduced from 2.5 mm in CSTR 1 to 0.7 mm in CSTR 2, no clogging problem occurred in CSTR 2
for an extended period of time since sizes of solids used and generated during the reaction were
smaller than the connection tube ID (0.7 mm).
Figure 8 (a) Optimization of the process using CSTR 2 and a solvent mixture of DMA:DME (3:1)
(b) SEM picture of the precipitate in the outflow containing precipitated NaBr and the excess
reagent Na CO .
2
3
Even though the outcome of the reaction could be improved via accurately controlling the reactor
RTD, a stable chemistry without self-inhibiting effect would still be desired to avoid potential
complications for process development. Thus, we sought to minimize the effect of the reaction
inhibiting species via reaction condition optimization. In order to quickly find the desired reaction
conditions, we designed a batch experiment to simulate the scenarios in the CSTR (see Figure S29
in Supporting Information). We found out that the use of pure DME as the solvent not only
minimized the self-inhibiting effect, but also led to high conversion and yield. The majority of the
solid reagents did not dissolve in pure DME which required feeding all starting materials as a
suspension via the slurry pump (See schematic in Figure 2a). In order to decouple the effect of the
12
ACS Paragon Plus Environment

Page 13 of 19
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
optimized solvent from the improvement obtained via narrower RTD, we performed the
experiment in CSTR 1. As shown in Figure 9a, the results indicate that the reaction yield and
conversion remain steady even though the geometry of CSTR 1 causes back mixing.
However, we found it challenging to achieve full conversion with CSTR 1, presumably due to its
broad RTD, and thus proceeded with CSTR 2 along with DME as the solvent. Cross coupling of
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
4
-bromotetrahydropyran 3 (1.5 equiv.) and methyl-4-bromobenzoate 4 (1 equiv.) was performed
in flow using photocatalyst 6 (2 mol%), Ni catalyst 7 (1 mol%), TTMSS (1.3 equiv.), Na CO (2
2
3
equiv.) and blue light-emitting diodes (440 nm) at 35° C in dry DME under exclusion of oxygen.
The stability of the process is demonstrated during 13 hours of production towards 1 gram of 5.
As a result, a yield of 77% and a conversion of 100% at a productivity of 77 mg/h could be
achieved. (Figure 9b). In comparison, the batch experiment yielded full conversion and 80% yield.
These results prove the performance and stability of the slurry pump as well as the feasibility to
handle heterogeneous reaction mixtures in the developed photo-CSTR cascade over extended
reaction times without clogging. The advantages over a similar sized batch reactor include the ease
of reagent handling in flow, a reduced reactor footprint and increased productivity.
Figure 9 (a) Optimization of the process using CSTR 1 and DME as system solvent (b) Gram-scale
synthesis over 13 hours of operation using CSTR 2 and DME as system solvent.
Finally, synthesis of a second pharmaceutically-relevant heterocyclic compound demonstrates the
versatility of the process of silyl radical-mediated metallaphotoredox cross-electrophile coupling
in flow (Scheme 3). The results (58% yield, 99% conversion, 80 mg/h productivity), similar to
batch (100% conversion and 60% yield), were achieved without further condition optimization.
Therefore, our approach demonstrates the direct translation of batch conditions to continuous flow
conditions.
13
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 19
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Scheme 3 Continuous silyl radical-mediated metallaphotoredox cross-electrophile coupling of
tert-butyl 4-(bromomethyl)piperidine-1-carboxylate 8 and methyl 4-bromobenzoate 4 towards
product 9 in the photo-CSTR platform.
Conclusion
This work demonstrates the development of a CSTR platform for handling heterogeneous visible-
light photoredox reactions in flow. The effect of the reactor geometry on particular chemistry
challenges such as potential self-inhibition is shown in a stepwise approach towards the optimized
process parameters. With additional solvent optimization, the reaction yield in steady state
significantly increased from 8% to 77%. Most importantly, the newly developed solid-feeding
technique within the platform offers a strategy towards heterogeneous photo-flow chemistry and
easy translation of batch to continuous flow.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
14
ACS Paragon Plus Environment

Page 15 of 19
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Experimental Methods
Materials
All chemicals used in this work were commercially available and used without further purification
unless stated otherwise. For all experiments dry and degassed solvents were used. Na CO was
sifted through a 53 µm sieve before use. 10 µm particles of K [Fe(CN) ] were bought from Sigma
Aldrich and used directly. 75 µm particles of K [Fe(CN) ] were obtained via grinding of larger
solids and further sieving using a 75 µm sieve. We used a nickel catalyst 7 which was provided by
2
3
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
6
3
6
Novartis Institutes for BioMedical Research.
Reactor design
The CSTR cascade consists of a polytetrafluoroethylene (PTFE) reactor block, an aluminum
housing, a heat-resistant glass cover and two aluminum covers. The cylinder-shaped inner
chambers on the reactor block have a diameter of 13 mm and a depth of 9 mm. The connecting
tubes between the chambers have an inner diameter of 2.5 mm (CSTR 1) and 0.7 mm (CSTR 2).
An O-ring gap for Kalrez 4079 O-Ring surrounds the chamber. The glass cover is heat-resistant
borosilicate glass (Pyrex) with dimensions of 90 mm × 25 mm × 4.8 mm. Both aluminum covers
have dimensions of 112 mm × 46 mm × 4.8 mm and were fabricated using water jet machining
(OMAX MicroMAX JetMachining Center). The PTFE reactor chamber and aluminum housing
were designed with SolidWorks and manufactured by Protolabs. All connection ports have 1/4-28
threads, which can be directly connected using common IDEX fittings (IDEX Health & Science
LLC.) without additional adapters. Commercially available LED lamps (Kessil lightening A160
WE) and focusing lenses (Thorlabs) were used. All other materials for the reactor were bought
from McMaster-Carr Supply Company.
Liquid RTD
Liquid phase RTD profiles were obtained using pulse injection. Deionized (DI) water was used as
carrier phase and methylene blue as liquid tracer (in water). In-line UV-Vis spectroscopy (light
source: Ocean Optics, Inc., DH- 2000-BAL; spectrometer: Ocean Optics, Inc., HR2000+) was
used to determine the concentration profiles of the tracer at the inlet and outlet. A six-way valve
(IDEX Health & Science LLC., MXP7900-000) was used and data collection was done with
LabVIEW.
Solid RTD
Multiphase RTD experiments use acetonitrile as carrier phase and a suspension of K [Fe(CN) ]
3
6
(two different particle sizes) in acetonitrile as solid tracer. Pulse injection via a six-way valve
equipped with an oscillator, for better solids handling, was performed. A stream of water along
with the aid of sonication is used to dissolve the tracer particles. In-line degassers are used to
prevent gas formation in the sonication bath after mixing in the T-junction. Initially, the
measurements were taken at the inlet and further at the outlet of the reactor at the same conditions.
For the inlet measurement the set-up was modified: T-junction and the sonication bath were
15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 19
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
directly connected to the six-way valve (See Figure S18 in Supporting Information). Particle size
distribution analysis was done on a Mastersizer 3000 (Malvern) with Hydro SV accessory. SEM
images were taken on a Phenom G1.
Chemical actinometry
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Due to the light-limiting reaction kinetics of the oxidation of 1,9-diphenylanthracene (DPA) to its
corresponding endoperoxide, accurate measurement of the differences in photon flux is possible.
A mixture of DPA (0.1 mmol/L) and Ru(bpy) Cl (0.2 mmol/L) in acetonitrile was prepared in a
3
2
dark room, bubbled with air and subsequently the solution (kept in a dark box) was pumped
through the reactor. In-line UV-Vis spectroscopy (light source: Ocean Optics, Inc., DH- 2000-
BAL; spectrometer: Ocean Optics, Inc., HR2000+) was used to determine the concentration of
DPA at 373 nm. Next, a PFR was characterized in order to compare the irradiation efficiency. The
PFR consisted of coiled FEP tubing on the top of an aluminium block (See Figure S9 in Supporting
Information). One LED lamp (Kessil lightening A160 WE, 40W) was placed 7 cm away from the
CSTR/PFR.
General procedure for silyl radical-mediated metallaphotoredox cross-electrophile coupling
reactions in flow
All starting materials (substrates, photocatalyst, Ni catalyst, tris(trimethylsilyl)silane, base,
internal standard) were added to the slurry pump and suspended in DME. The reaction mixture
was directly degassed by sparging with argon in the slurry pump. The CSTR cascade was purged
with argon, prefilled with degassed DME and then connected to the slurry pump. The pressurized
fraction collector was purged with argon before the run and then set to 10 psig. An oscillator
(Model No. 306-10H, Precision Microdrives) was installed at the connecting tube between the
slurry pump and the CSTR in order to enhance solids transport. (see Figure S1 Supporting
Information) The slurry outflow was controlled with a hydraulic liquid (acetonitrile) which was
pumped using a positive displacement pump (VICI, Model M6 pump). Reaction progress was
monitored by reversed-phase UPLC-MS (Waters AcQuity system) equipped with photodiode
array (PAD, 200-400 nm), evaporative light scattering detection (Shimadzu ELSD-LT II) and
single quadrupole mass detection (SQD). NMR experiments were conducted on 400 MHz
instruments (Bruker Ultrashield Plus spectrometer) using CDCl3 (99.9% D) as the solvent.
Chemical shifts were referenced to the solvent peak (7.26 ppm for CDCl3) and were reported in
parts per million (ppm). Naphthalene (HPLC) and 1,3,5-trimethoxybenzene (NMR) were used as
internal standards. Preparative reversed-phase HPLC (Waters AcQuity system) equipped with a
C-18 reversed phase column (XBridge BEH C18, 30.0 x 50 mm) was used to afford the product.
HRMS data was recorded on a Waters Xevo G2 Qtof.
CSTR simulating batch experiments
For solvent screening, all reagents for the reaction as well as the solvents to be examined were
added to a reaction vessel and the mixture was irradiated. After completion of the reaction, a
sample was taken in order to confirm the progress. Further, new starting materials were added, the
reaction mixture is irradiated again and the analyzed. The stepwise addition of starting materials
alternated by irradiation of the mixture is intended to simulate the conditions in the CSTR – the
16
ACS Paragon Plus Environment

Page 17 of 19
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
starting materials are pumped into the system and react in a continuous manner (see Figure S29
and related text in Supporting Information). Yield and conversion were determined via HPLC and
confirmed with quantitative H NMR using naphthalene and 1,3,5-trimethoxybenzene as internal
1
standards.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Supporting Information
The Supporting Information is available free of charge on the ACS Publications website at DOI:
xxx.xxx.
Detailed experimental descriptions of experimental setups, residence time distribution
measurement, photon flux measurement, and heterogeneous photochemistry on CSTR platform.
NMR spectra of key compounds are also included. (PDF)
Author Information
Corresponding Author:
*
E-mail: kfjensen@mit.edu.
Notes:
The authors declare no competing financial interest.
Acknowledgment
This work was supported by the Novartis MIT Center for Continuous Manufacturing. A.
Pomberger is grateful for a fellowship from the Austrian Marshall Plan Foundation.
References
1. Shaw, M. H.; Twilton, J.; MacMillan, D. W. C., Photoredox Catalysis in Organic Chemistry.
The Journal of Organic Chemistry 2016, 81, 6898-6926.
2
3
4
. Prier, C. K.; Rankic, D. A.; MacMillan, D. W. C., Visible Light Photoredox Catalysis with
Transition Metal Complexes: Applications in Organic Synthesis. Chemical Reviews 2013, 113,
5
322-5363.
. Kalyani, D.; McMurtrey, K. B.; Neufeldt, S. R.; Sanford, M. S., Room-Temperature C–H
Arylation: Merger of Pd-Catalyzed C–H Functionalization and Visible-Light Photocatalysis.
Journal of the American Chemical Society 2011, 133, 18566-18569.
3
. Liang, Y.; Zhang, X.; MacMillan, D. W. C., Decarboxylative sp C–N coupling via dual copper
and photoredox catalysis. Nature 2018, 559, 83-88.
17
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 19
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
5
6
. Zuo, Z.; Ahneman, D. T.; Chu, L.; Terrett, J. A.; Doyle, A. G.; MacMillan, D. W. C., Merging
3
photoredox with nickel catalysis: Coupling of α-carboxyl sp -carbons with aryl halides.
Science 2014, 345, 437.
. Tóth, B. L.; Tischler, O.; Novák, Z., Recent advances in dual transition metal–visible light
photoredox catalysis. Tetrahedron Letters 2016, 57, 4505-4513.
7. Jensen, K. F., Flow chemistry—Microreaction technology comes of age. AIChE Journal 2017,
3, 858-869.
8
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6
. Plutschack, M. B.; Pieber, B.; Gilmore, K.; Seeberger, P. H., The Hitchhiker’s Guide to Flow
Chemistry. Chemical Reviews 2017, 117, 11796-11893.
. Roberge, D. M.; Ducry, L.; Bieler, N.; Cretton, P.; Zimmermann, B., Microreactor
Technology: A Revolution for the Fine Chemical and Pharmaceutical Industries? Chemical
Engineering & Technology 2005, 28, 318-323.
0. Elliott, L. D.; Berry, M.; Harji, B.; Klauber, D.; Leonard, J.; Booker-Milburn, K. I., A Small-
Footprint, High-Capacity Flow Reactor for UV Photochemical Synthesis on the Kilogram
Scale. Organic Process Research & Development 2016, 20, 1806-1811.
1. Harper, K. C.; Moschetta, E. G.; Bordawekar, S. V.; Wittenberger, S. J., A Laser Driven Flow
Chemistry Platform for Scaling Photochemical Reactions with Visible Light. ACS Central
Science 2019, 5, 109-115.
9
1
1
12. Su, Y.; Hessel, V.; Noël, T., A compact photomicroreactor design for kinetic studies of gas-
liquid photocatalytic transformations. AIChE Journal 2015, 61, 2215-2227.
3. Mohamed, D. K. B.; Yu, X.; Li, J.; Wu, J., Reaction screening in continuous flow reactors.
Tetrahedron Letters 2016, 57, 3965-3977.
4. Tucker, J. W.; Zhang, Y.; Jamison, T. F.; Stephenson, C. R. J., Visible-Light Photoredox
Catalysis in Flow. Angewandte Chemie International Edition 2012, 51, 4144-4147.
1
1
15. Booker-Milburn, K. I.; Noël, T., Flow Photochemistry. ChemPhotoChem 2018, 2, 830-830.
16. Elliott, L. D.; Knowles, J. P.; Koovits, P. J.; Maskill, K. G.; Ralph, M. J.; Lejeune, G.; Edwards,
L. J.; Robinson, R. I.; Clemens, I. R.; Cox, B.; Pascoe, D. D.; Koch, G.; Eberle, M.; Berry, M.
B.; Booker-Milburn, K. I., Batch versus Flow Photochemistry: A Revealing Comparison of
Yield and Productivity. Chemistry – A European Journal 2014, 20, 15226-15232.
1
7. Hsieh, H.-W.; Coley, C. W.; Baumgartner, L. M.; Jensen, K. F.; Robinson, R. I., Photoredox
Iridium–Nickel Dual-Catalyzed Decarboxylative Arylation Cross-Coupling: From Batch to
Continuous Flow via Self-Optimizing Segmented Flow Reactor. Organic Process Research &
Development 2018, 22, 542-550.
8. Tucker, J. W.; Zhang, Y.; Jamison, T. F.; Stephenson, C. R. J., Visible-light photoredox
catalysis in flow. Angewandte Chemie (International ed. in English) 2012, 51, 4144-4147.
9. Hartman, R. L., Managing Solids in Microreactors for the Upstream Continuous Processing of
Fine Chemicals. Organic Process Research & Development 2012, 16, 870-887.
1
1
20. Mo, Y.; Jensen, K. F., A miniature CSTR cascade for continuous flow of reactions containing
solids. Reaction Chemistry & Engineering 2016, 1, 501-507.
2
1. Chapman, M. R.; Kwan, M. H. T.; King, G.; Jolley, K. E.; Hussain, M.; Hussain, S.; Salama,
I. E.; González Niño, C.; Thompson, L. A.; Bayana, M. E.; Clayton, A. D.; Nguyen, B. N.;
Turner, N. J.; Kapur, N.; Blacker, A. J., Simple and Versatile Laboratory Scale CSTR for
Multiphasic Continuous-Flow Chemistry and Long Residence Times. Organic Process
Research & Development 2017, 21, 1294-1301.
22. Mo, Y.; Lin, H.; Jensen, K. F., High-performance miniature CSTR for biphasic C–C bond-
forming reactions. Chemical Engineering Journal 2018, 335, 936-944.
18
ACS Paragon Plus Environment

Page 19 of 19
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
2
2
3. Zhang, P.; Le, C. C.; MacMillan, D. W., Silyl Radical Activation of Alkyl Halides in
Metallaphotoredox Catalysis: A Unique Pathway for Cross-Electrophile Coupling. J Am Chem
Soc 2016, 138, 8084-7.
4. Yang, Q.; Drak, A.; Hasson, D.; Semiat, R., RO module RTD analyses based on directly
processing conductivity signals. Journal of Membrane Science 2007, 306, 355-364.
25. Levenspiel, O., Chemical reaction engineering. Wiley: New York, 1999.
2
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
6. Pitre, S. P.; McTiernan, C. D.; Vine, W.; DiPucchio, R.; Grenier, M.; Scaiano, J. C., Visible-
Light Actinometry and Intermittent Illumination as Convenient Tools to Study Ru(bpy) Cl
3
2
Mediated Photoredox Transformations. Scientific Reports 2015, 5, 16397.
19
ACS Paragon Plus Environment