You are on page 1of 23

Accepted Article

Title: Reactor technology concepts for flow photochemistry

Authors: Thomas Rehm

This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.

To be cited as: ChemPhotoChem 10.1002/cptc.201900247

Link to VoR: http://dx.doi.org/10.1002/cptc.201900247

A Journal of

www.chemphotochem.org
ChemPhotoChem 10.1002/cptc.201900247

REVIEW

Reactor technology concepts for flow photochemistry


Thomas H. Rehm*[a]

In memory of Prof. Carsten Schmuck (1968-2019)

Accepted Manuscript
[a] Title(s), Initial(s), Surname(s) of Author(s) including Corresponding
Author(s)
Department
Institution
Address 1
E-mail:
[b] Title(s), Initial(s), Surname(s) of Author(s)
Department
Institution
Address 2

Supporting information for this article is given via a link at the end of
the document.((Please delete this text if not appropriate))

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
Abstract: Synthetic organic photochemistry has been intensively Continuous flow chemistry is inseparably linked to a higher level
carried out in the 20th century and paved the way to complex of technology than common batch synthesis on lab scale. The
organic molecules, which were not yet accessible via thermal ongoing transfer from flask, stir bar and oil bath to flow chemistry
chemistry. Several photochemical synthesis routes have found equipment leads to the integration of new concepts in mixing or
their way into industrial applications for the production of everyday dosing, heating and processing in general.[7] Other materials than
commodities, but photochemistry was still underutilized as a glass come into focus, e.g. stainless steel or high performance
synthesis method in organic chemistry until recently. With the polymers. This extension in reaction volume determining
advent of novel photocatalytic and photophysical concepts for the materials allows so-called Novel Process Windows with much
use of high power visible light the research field of synthetic broader physical regimes for chemical reactions (e.g. temperature,
organic photochemistry has evolved to a vivid and highly pressure, substrate concentration) than possible for common
recognized technique in the last years. Fortunately, continuous batch chemistry.[8] In the case of photochemistry the contacting
flow technology has become a more and more accepted tool as event between the chemical phases (gas, liquid, solid) and the
well and proved to be an excellent key player for the advancement physical phase of electromagnetic radiation dominates.[6e, 6f, 9]

Accepted Manuscript
of photochemistry in academic and industrial research. This Contrary to batch mode synthesis in static reaction volumes (mL
review shall give an overview on the recent developments of scale!) flow chemistry can take place on the µL scale, but
continuous flow photoreactors and their application to continuously. The reduction of the reaction volume to thin films in
photochemical syntheses under mild and defined process open microchannels or fine streams in capillaries is the major
advantage for photochemical applications in flow equipment as it
conditions.
is prerequisite for defined irradiation time and full transmission of
light through the reaction solution as well as intensive absorption
by molecules in solution or by surfaces with immobilized
1. Introduction catalysts.[6] The regulating variables are in this latter context the
molar extinction coefficient ε and the molar concentration c of the
In the last decade photochemistry experienced a renaissance
light absorbing matter. The relevant dependency is expressed in
with fundamental impact on academic research. Major driving
the Lambert-Beer law with A = log10T = log10(I0/I) = ε × c × d with
force for this rapid development was the application of visible light
A as molar absorption, T as transmission, l0 and I as incident and
to photocatalytically active materials like metalorganic complexes
transmitted light intensity, and d as light path length through the
with e.g. Ru or Ir metal centers, metal-free sensitizers (e.g. Rose
reaction volume.[10] The latter variable d lost its significance on the
Bengal, riboflavin derivatives or perylene diimides) or colored
µ scale whereas the molar concentration c gets more important
electron donor-acceptor complexes.[1] Dual catalysis was applied
due to theoretical and practical reasons: a) The Lambert-Beer law
as well to photochemical reaction pathways enabling precise
is only valid for dilute and homogeneous substrate solutions,
control over the reactivity of combined catalyst systems. [2] All
which compromises the advantage of continuous flow synthesis
efforts in this field stimulated both the development of novel
running reactions with high substrate concentrations; b) Too high
(molecular) photocatalysts with specific control of light absorption
concentrations of substrates can result in precipitating
properties and redox parameters as well as the understanding of
intermediates or products. Under these conditions of solid
photophysical processes of catalyst excitation and subsequent
formation with concentration gradients in the solution the
interaction with the substrate.[3] The use of visible light clearly
Lambert-Beer law is not valid at all.[11] Possibly, clogging of the
broadened the application horizon of photochemistry with milder
flow system might stop any efforts immediately. In our case
reaction conditions available via less energetic light. Now
results derived from the Lambert-Beer law should be seen as
sensitive fine chemicals can be converted by mild visible-light
estimation with a clear view on the physical and chemical
photocatalysis with high (stereo)selectivity.[4] In contrast, less
conditions in the irradiated sample volume. Furthermore, an
reactive aryl carbon-halogen bonds can be selectively cleaved
imprudent concentration increase of e.g. a polyaromatic sensitizer
with consecutively excited photocatalysts, which opens a pathway
can lead to soluble aggregates with different photophysical
to noble metal-free cross coupling reactions.[5] In parallel many
properties and potentially different catalytic activity. [12] Hence,
research groups worldwide started to transfer their results in
new pitfalls clearly follow up to new and enticing possibilities of
photocatalysis from batch mode to continuous flow mode in order
novel technology. But, and this is a most fortune coincidence,
to improve light-matter interaction on smaller length scales. And
photochemistry of the 21st century stimulated not only chemists in
in fact, the recent years and countless publications clearly prove
thinking about chemistry in a novel fashion, but also chemical
that photochemistry benefits from “going to flow”. [6]
engineers in performing chemical processes in continuous flow
with the delicate tool of light. [6a,13] It is indeed amazing how
chemistry and technology development go hand in hand in the
[a] Dr. Thomas H. Rehm recent years for understanding flow photochemistry and
Division Energy & Chemical Technology / Flow Chemistry Group
advancing it towards real applications. In parallel, rapid
Fraunhofer Institute for Microengineering and Microsystems IMM
Carl-Zeiss-Straße 18-20, 55129 Mainz, Germany development of novel light emitting materials boosted the market
*E-mail: thomas.rehm@imm.fraunhofer.de of LED technology as adaptable light source for roughly any
Website: https://www.imm.fraunhofer.de photoreactor architecture and photochemistry application.[14]
Website: https://www.flowphotochemistry.com
Finally, continuous flow technology supports the paradigm

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
change for running chemistry in a sustainable and green e.g. incident photon flux, medium absorbance and photonic
manner.[8b] efficiency.[13] In consequence to the strong interest in
General information about photochemistry and the beneficial use photochemistry (with visible light) and the growing acceptance of
of continuous flow reactors for photochemistry has been flow chemistry in industry, ready-to-use photoreactors and setups
described already before in several reviews and books. [1-6,15] have become commercially available as well.[18] Such complete
Hence, the intention of this review is to summarize recent flow chemistry machines allow easy access to flow synthesis and
developments in reactor designs for flow photochemistry. As it is clearly reduce any fear of contact to this technology. In parallel
usually the case for a review on a specific topic, the selected photochemistry setups have run into further evolution and
examples cannot cover the complete field in photochemistry and combine multiple application scenarios like high-throughput
flow technology development. From a personal view both screening, preparative-scale batch reactions or continuous
literature survey and content selection was a bottomless pit and processing with a single source of light. [19] The following chapters
new papers on novel synthesis routes or reactor concepts and will give an overview about the advancement of already known
characterizations were published frequently per week. Hence, this reactor types as well as rather novel reactor concepts for more

Accepted Manuscript
vast base of information was used to gather advancements in specified applications in photochemistry.
(micro)reactor design and architecture for flow photochemistry.
The review ends with a rather personal conclusion on this vibrant 2.1. Capillary photoreactors
research field looking at the impact on universities, industry and
funding agencies. 2.1.1 Polymer capillary photoreactor for collimating light
sources

Polymer capillary reactors have become the easiest accessible


Thomas Rehm studied chemistry at the
photoreactor systems for continuous flow applications.
Julius Maximilians University in Würzburg,
Chemically inert (co)polymers like fluorinated ethylene propylene
Germany and finished his PhD thesis in 2008
(FEP) or perfluoroalkoxy alkane (PFA) exhibit a very good light
with Carsten Schmuck focusing on
transmission in the ultraviolet and visible light region and allow
supramolecular polymer chemistry in polar
flexible adaptation of the tubings to various reactor
solutions. Afterwards he joined the group of
architectures.[20] Usually polymer capillaries are wrapped around
Frank Würthner at the same university and
a light-transparent annular support that allows central
worked on perylene diimides for DNA/RNA
incorporation of a light source with radial emission, e.g. mercury
recognition. In 2011 he started a new position
vapor lamps. In most cases heat management and control of
at Fraunhofer IMM in Mainz, Germany
emission wavelength can become rather complicated due to less
working on continuous flow chemistry. His
space between light source and capillary. Hence, Telmesani et al.
research is focused on the use of microreaction technology for
developed a photochemistry platform that allows the straight-
photochemical applications in flow and on the integration of benchtop NMR
forward use of an external Hg-Xe light source emitting a
technology as online analysis tool.
collimated light beam of defined irradiance to a capillary
photoreactor.[21] The commercially available high-power light
source provides a homogenous emission of light for irradiating a
planar target. In order to use the benefits of a capillary reactor the
2. Reactor concepts for flow photochemistry
authors had to design a support for the flexible capillary that
imitates a planar surface in view from the light source. This
The focus of this chapter is on reactor concepts for flow
requirement was made possible by using a cone-shaped
photochemistry that emerged in recent years. The main
stainless-steel support with a helical groove on the outer surface.
application range of these photoreactors is the synthesis of fine
The FEP capillary (o.d. 1.59 mm, i.d. 0.79 mm) is wrapped around
chemicals. Reactor concepts for wastewater purification or
the cone and gets stabilized by the cavities of the groove (Figure
degradation of e.g. pharmaceuticals will be not discussed as well
1). The cone itself is hollow and can be continuously flushed with
as water-splitting for hydrogen production. These topics are
a coolant to maintain the desired temperature inside the capillary.
summarized and highlighted in various reviews. [16,17] In the initial
Heat transfer from the reaction solution to the coolant is even
years of flow photochemistry most of the reactors were custom-
promoted by the ideal contact between the stainless-steel
made with equipment easily available from the lab bench and
grooves and the capillary.
shelf. This straightforward strategy is easy to understand from a
chemist’s point of view with a sole focus on running chemical
reactions in flow. But from an engineering point of view this
approach is less reasonable since the physical conditions in
“hand-made” reactors are not as reproducible as necessary for
comparing reactors and processes or for scale-up. Fortunately,
there are meanwhile many activities concerning the investigation
of flow photoreactor equipment on the engineering level, which
make the process parameters easier to understand and deploy,

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
photocatalyst into a regular gas-liquid flow inside a capillary
reactor.[24] The basic idea of this approach is the formation of so-
called serial micro-batch reactors (SMBRs), which are simply
well-ordered liquid phase segments alternating with inert gas
segments inside the capillary reactor. This classical slug flow
regime allows efficient mixing inside the liquid segments and
excellent gas mass transfer at the gas-liquid interface.
Furthermore strong interfacial forces between the gas and the
liquid phase are the prerequisite for constant and efficient
Figure 1. A) Continuous flow photochemistry platform with directional irradiation transport of solid material inside the capillary. Nevertheless, the
by an external light source; B) photoreactor with cone-shaped heat exchanger
as stabilizing element for the capillary. Adapted with permission from ref. [21]
incorporation of equal amounts of solid catalyst into SMBRs is
copyright 2015 Wiley-VCH. mandatory for enabling equal reaction conditions inside each
micro-batch reactor. Pieper et al. solved this problem by the

Accepted Manuscript
preparation of a photocatalyst suspension in a viscous solvent,
Telmesani et al. applied this photoreactor system to the [2+2] e.g. an ionic liquid like [Bmim]BF4. Sedimentation of solids is
photocycloaddition of methyl cinnamate 1 (0.8 M in acetonitrile) to strongly hindered in such a solvent and can be even more
methyl truxinate 2. Unfortunately, neither reactor volume nor flow reduced to a negligible rate by active mixing with a magnetic stir
rates were given in this work, but variations in temperature, bar inside the dosing syringe. The authors exemplified their
electrical light source power and emission cut-off are described. approach for the heterogeneously catalyzed decarboxylative
A maximum conversion of 29% were obtained for 2 (d.r. = 1:1) at fluorination of phenoxy acetic acid 4 using CMB-C3N4 as
0 °C, 500 W power and with an irradiation greater 305 nm. Batch photocatalyst derived from cyanuric acid, melamine and barbituric
tests resulted in a low conversion of 14%. A strong increase in acid (Figure 3).
conversion and diastereoselectivity towards the δ-truxinate 2 was
achieved by the use of a thiourea organocatalyst 3 (10 mol%).
Pre-coordination via strong hydrogen-bonding boosted the
conversion to 76% with a diastereomeric ratio of 3:1 in favour of
δ-truxinate (Figure 2).
Just recently Telmesani et al. could improve this process by
switching to a biphasic liquid-liquid process with ethyl acetate and
water as solvent system (1:10).[22,23] The conventional capillary
photoreactor was used with an inner volume of 11 mL at a net Figure 3. Continuous flow synthesis of (fluoromethoxy)benzene 5 with carbon
nitride as solid photocatalyst in a triphasic slug flow system.
flow rate of 100 µL min-1 (τres = 110 min). The resulting water slugs
inside the organic phase led to a better circulation thus mixing of
the organic reaction solution. Furthermore, thin film formation
The preparation of the triphasic system was done by mixing at first
around the water slugs resulted in an improved interaction with
the liquid phase (acetonitrile-water, 1:1, fl = 0.5 mL min-1)
UV light. Both advantages resulted in a maximum conversion of
containing 4 (0.2 mol L-1) and SelectFluor® (3 eq) with nitrogen
87% for methyl truxinate 2 with a diastereomeric ratio of 3:1 (δ:β).
gas (fg = 0.3 mL min-1) via Y-shaped mixer into a FEP capillary
Without slug flow recirculation was necessary for 8 hours to reach
(i.d.: 1.6 mm). The stabilized slug flow was then contacted with
a yield of 75%.
the solid catalyst in a T-piece connected to a syringe filled with
the photocatalyst suspension of 85 mg CMB-C3N4 in 4.2 mL
[Bmim]BF4 with 0.8 mL H2O (Figure 4). A flow rate of fsus = 50 µL
min-1 was applied here and led to the generation of the triphasic
mixture, which was finally transferred into the capillary
photoreactor (FEP, τres = 14 min, 17 mL) and irradiated with an
LED array at 420 nm (12 W). Under these conditions the
decarboxylative fluorination proceeded smoothly with full
conversion and 94% NMR yield. Importantly, any CO2 generated
during the reaction can easily diffuse out of the liquid slugs without
disturbing the stability of the solid-liquid SMBRs by extended gas
Figure 2. Synthesis of methyl δ-truxinate 2 via [2+2] photocycloaddition of slugs. Finally, a first approach for the scale-up of this process was
methyl cinnamate 1 using thiourea organocatalyst 3. done as well. The authors integrated a second suspension dosing
device prior to the photoreactor, which allowed the on-the-fly
exchange of empty syringes for the CMB-C3N4 suspension. Within
2.1.2 Triphasic micro-batch reactors in flow an overall run time of 390 min approx. 1.6 g of 5 was produced.
This work by Pieper et al. clearly sets the stage for future
Slug flow conditions can be also extended to a triphasic system applications of many solid reagents and catalysts to be used in
as demonstrated by Pieper et al. for the incorporation of a solid continuous flow reactors. Gas-liquid and liquid-solid phase

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
separation can be easily done by filtration of the solid content form the same process conditions (photolysate concentration, UV lamp
the reaction stream. Depending on the character of the solid power, emission wavelength) either in batch mode or in capillary
material a re-use may be possible as well, which is beneficial for photoreactors with one or three layers. Based on this work the
the overall (cost) efficiency of the process. authors revealed three very interesting trends for synthetic
photochemistry processes: a) At full conversion the yields are
basically the same in batch and continuous flow mode; b) FEP
photoreactors with three capillary layers have on average a 20%
higher productivity compared to the same batch end point; c) In
reverse one-layer FEP photoreactors have on average a 20%
lower productivity compared to the same batch end point. The
authors state that the lower productivity of the one-layer
photoreactor can be attributed to the shorter light path length of
2.7 mm in the capillary against 8.7 mm in the batch vessel. In

Accepted Manuscript
reverse, the three-layer photoreactor offers superior productivity
due to a strongly increased residence time over three capillary
lengths and the possibility to absorb remaining photons in the
second and third capillary layer after passing the first layer. But
finally Elliott et al. also conclude that the superiority of the FEP
capillary photoreactor technology is not yet fully understood.
Parameters like gradients in concentration and light absorption
throughout the complete capillary system must be investigated as
well as light scattering and refraction effects.
Just recently, Wriedt at al. published their work on the
experimental determination of photon fluxes in multilayer capillary
photoreactors.[13g] The authors applied the photochemical
decomposition of [FeIII(C2O4)3]3- to [FeII(C2O4)2]2-, carbon dioxide
and oxalate (ferrioxalate actinometer) to a three-layer FEP
photoreactor with a medium pressure Hg lamp (150 W) as central
light source (Figure 5).

Figure 4. A) Suspension dosing device for triphasic mixing (inset: schematic


representation of gas-liquid-solid contacting); B) Resulting slug flow with solid
photocatalyst in the liquid phase; C) FEP capillary reactor with SMBRs under
steady state conditions. Adapted with permission from ref. [24] copyright 2018
Wiley-VCH.

2.1.3 Multilayer capillary photoreactor

The next step in the evolution of capillary photoreactors goes Figure 5. A) Self-made three-layer FEP capillary photoreactor used for the
experimental determination of photon fluxes in separate capillaries; B)
ahead to multilayer capillary designs. In general, this approach is
schematic representation of the capillaries wrapped around the central Hg lamp.
attractive due to two reasons: a) Residual photons, which were Adapted with permission from ref. [13g] copyright 2018 Wiley-VCH.
not absorbed in the first layer can be further used in a second or
third layer; b) Any additional capillary layer can be used either for
increasing the residence time or for running different or cascade Interestingly, Wriedt et al. could show that the on average external
reactions in one reactor. In 2014 Elliott et al. published a work photonic efficiency as ratio between the absorbed and available
about the question which process mode is more productive for photons is 11% for the inner layer, 3% for the middle layer and
photochemistry: batch or continuous flow. [25] The authors finally 1% for the outer layer. The data obtained in this work show
performed and compared several photochemical reactions under an exponential decay of the amount of absorbed photons with

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
increasing number of capillary layers, which is in accordance with since light leaks pretty easily out of the waveguide. To overcome
the Lambert-Beer law. This behavior points to the fact that each this problem photonic crystal fibers have been developed, which
additional capillary layer acts with its cross section as extended consist of a waveguide core surrounded by a periodic array of
optical path length for any photons that were able to pass the holes ranging over the complete length of the fiber. [28,29] This
foregoing layer by scattering or refraction. architecture acts as a trap for the light inside the central hollow
From a technological point of view our group started to set up core with a highly reflective cladding structure. One example is
capillary photoreactors with permanently integrated and water- the so-called kagomé hollow core PCF whose name originates
chilled LED light sources and FEP capillaries wrapped around a from the specific arrangement of triangles and hexagons around
quartz glass cylinder.[26] While working with this first version of the the central hollow core (Figure 7A and B). This architecture allows
photoreactor it became obvious that a separate temperature a broadband spectral transmission through the liquid filled fiber
management for capillaries and LED arrays is necessary as well for sensing or photochemical experiments with sub-µL samples
as an useful opportunity to switch to defined wavelengths (Figure 7C).
depending on the reaction to be performed. Hence, the revised

Accepted Manuscript
version is a two-layer capillary photoreactor (0.8 or 1.6 mm i.d.
with 2 × 15 mL or 2 × 30 mL inner volume) with exclusive cooling
for the capillaries in a double cylinder chamber (Figure 6). The
LED arrays are exchangeable and are placed on a separately
chilled cooling finger in the center of the reactor assembly.
Emission wavelengths can be chosen from the UV to visible light
region (six LED emitters per circuit board fixed on a hexagonal
aluminum support). A maximum of 110 W electrical power is
applicable to such an LED array with 36 emitters. Thanks to the
rapidly developing LED market high power LEDs with an emission
wavelength below 340 nm will hopefully become available in near
future. With this device our current research is focused on the
areas of dye-sensitized singlet oxygen formation, cyanation
reactions, diazonium salt chemistry and photochemical
nanoparticle formation.

Figure 7. A) Schematic representation of a kagomé hollow core PCF with a core


diameter of 20 µm; B) SEM image of a kagomé hollow core PCF; C) Schematic
representation of continuous flow setup with a PCF as microreactor on the µL
scale. Adapted with permission from ref. [27] copyright 2013 Wiley-VCH and ref.
[28] copyright 2013 Royal Society of Chemistry.

Figure 6. A) Compact two-layer capillary photoreactor with exchangeable LED


Of course, PCFs used as “capillary” photoreactors cannot be
arrays and separate capillary and LED cooling; B) Assembled photoreactor with
LED array for green light emission at 520 nm; C) Holder for LED arrays compared with their larger FEP relatives, but they offer several
facilitating an easy switch to other emission wavelengths. Copyright 2018 very interesting properties, which complement the application
Fraunhofer IMM. portfolio of common photoreactors: a) The optofluidic channels of
PCFs exhibit an extremely reduced sample volume; b) Low
transmission losses inside PCFs allow the detection of extremely
2.1.4 Photonic crystal fibers as ultra-low volume low concentrated analytes within a light path length of up to
photoreactors several meters; c) Low quantum yield photochemical processes
can be measured and analyzed due to the strongly enhanced
The last example in this chapter deals with the use of photonic light-matter interaction inside the PCF core. [28] Despite these
crystal fibers (PCFs) as an advanced version of conventional solid obvious advantages PCF-based flow set-ups can be also
optical fibers and hollow capillaries. [27] The latter ones can be combined with classical analytical tools of mass spectrometry or
already used as light-guiding capillaries loaded with a chemical in situ spectroscopy. Finally, the inner surface of PCFs can be
sample for intense interaction with the injected light. Unfortunately, chemically modified and used as support for e.g. catalytically
the interaction length is most often restricted to some centimeters active particles.[30]

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
2.1.5 Polymer-based waveguides as photoreactors Coiled quartz glass tubes exhibit an excellent transmission even
in the deep UV light region. Hence, they have been used as
Optofluidic PCF microreactors are limited to either liquid or integral part of continuous flow photoreactors, but they never
gaseous phase inside the hollow fiber. Bubble formation in the gained so much attention compared to their relatives made of
liquid stream or even gas-liquid slug flow destroys the light flexible polymer materials, although such solid glass coils are
guidance inside the fiber and hinders efficient light-matter commercially available and can be implemented in complete
interaction.[31] Solvents with good gas solubility can be photoreactor systems.[33] On the contrary, straight quartz glass
presaturated with the appropriate gas and pumped into the PCF, tubes have meanwhile made their way into several technological
but continuous dosing of gas into the reaction solution is not applications for photochemical processes. Recent examples will
possible as well. To circumvent this problem Ponce et al. used an be summarized in this chapter, whereby conventional glass chip
amorphous fluoropolymer (Teflon® AF-2400), which allows gas reactors with chemically etched or mechanically machined
permeation through chemically inert polymer membrane. Most microchannels will not be discussed here. These rather small
importantly this fluoropolymer has a very low refractive index versions of glass reactors have been already described in detail

Accepted Manuscript
<1.30.[32] Capillaries made of this material can be filled with a as well as their applications in photochemistry. [34]
solvent (e.g. water) and used as liquid core waveguide membrane
(LCWM) due to total internal reflection. In analogy to the setup 2.2.1 Quartz glass photoreactor for UV-C photochemistry
depicted in Figure 7C Ponce et al. used this concept to design a
continuous flow photoreactor system with water as solvent.[31] The In the first example DeLaney et al. describe an annular continuous
authors could show that in the visible light region optical losses flow reactor combined with a set of excimer lamp light sources for
are fairly low at wavelengths <575 nm (1.67 ± 0.70 dB m -1). At the irradiation with defined wavelengths (KrCl: 222 nm; XeBr: 282
longer wavelengths the optical loss increases due to the stronger nm; XeCl: 308).[35] The reactor itself consists of a stainless steel
absorption of light by the solvent water (Figure 8A). The UV-Vis cooling finger (o.d. 12 mm), which is placed inside a quartz glass
spectroscopy performance of the system was then evaluated by tube (o.d. 15.6 mm). The outlet of a 1/16” stainless steel capillary
comparing absorption spectra of dye solutions containing Azure leaves the cooling finger at its end inside the quartz tube and is
A (AA) or Acid Orange (AO) with those obtained in standard used as supply for the reaction solution, which flows back as thin
cuvettes. As depicted in Figure 8B an excellent accordance is film of approx. 120 µm thickness between the inner glass wall and
obtained with the reference measured in the cuvette. the outer cooling finger wall (Figure 9). The assembled reactor is
then placed in the excimer lamp housing with an overall irradiation
volume of 1.37 mL.

Figure 8. A) Optical losses in the visible light region as measured in Teflon® AF


2400 tube filled with bi-distilled water; B) Absorption spectra of Azure A (AA)
and Acid Orange (AO) obtained from the tube system (LCWM) or from a
standard cuvette. Adapted with permission from ref. [31] copyright 2018 Wiley-
VCH.

The polymer waveguide was supplied with air by placing it into a


Figure 9. Schematic representation of the continuous flow device to be placed
pressurized steel tube. Bubble formation inside the polymer
inside an excimer lamp housing. Adapted with permission from ref. [35]
capillary (i.d. 1.0 mm, o.d. 1.6 mm, length: 52 cm) was copyright 2016 Royal Society of Chemistry.
investigated for a gas pressure range of 0.5 to 8 bar while the
liquid pressure inside the capillary was maintained at 2.8 bar.
Short residence times <5 s allow a high gas pressure of up to 8 The photoreactor set-up was used for the photodecarboxylative
bar, whereas a more reasonable residence time >20 s requires a cyclisation of phthalimid 6 in a 1:1 water-acetone solvent mixture
lower gas pressure of approx. 4 bar to avoid oversaturation and (Figure 10). The best yield for cyclized 7 was obtained with an
bubble formation inside the capillary. The system was finally used irradiation at 308 nm and a substrate concentration of 0.1 M. With
to perform a continuous flow methylene blue photocatalyzed a flow rate of fl = 0.2 mL min-1 nearly 100% yield could be obtained
oxidation of D-glucose. corresponding to a space time yield (STY) of 0.88 mmol h -1 mL-
1 [36]
. A maximum STY of 2.99 mmol h-1 mL-1 was obtained by
2.2 Glass (tube) photoreactors increasing the flow rate to fl = 1.0 mL min-1, which, in reverse,
resulted in a lower yield of 68%. Irradiation at 222 nm led under

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
the same reaction conditions to 37% yield. This low value results
from the smaller number of emitted photons at 222 nm compared
to the irradiance with the XeCl lamp (308 nm) at the same
electrical power level. Finally, the authors summarize that in view
of green chemical reaction technology excimer lamps are favored
against Hg lamps due fewer safety issues. And especially for this
novel reactor concept less stray light is produced by the lamps as
well as less ozone due to the straight fit of the reactor inside the
lamp housing.

Accepted Manuscript
Figure 10. Photodecarboxylative cyclisation of potassium N-phthalimido-
butanoate 6 with UV light in continuous flow.

Figure 11. A) Disassembled vortex reactor with LED arrays, motor and control
2.2.2 Gas-liquid mixing via vortex formation
box; B) LED arrays and mirror blocks assembled around the glass tube; C)
Schematic representation of the vortex reactor showing delivery and removal of
A second reactor concept is based on the generation of Taylor reaction solution and gas intake; D) and E) Vortices formation as modelled via
vortices by a rapidly rotating stainless steel cylinder inside of a CFD and in live operation. Adapted with permission from ref. [37] copyright 2017
American Chemical Society.
transparent Pyrex® jacketed tube with a sealed bottom.[37] Based
on already existing design studies of such a vortex reactor Lee et
al. developed an advanced version for reactions with oxygen gas
in combination with high power LED technology.[38] Intense mixing Lee et al. performed several photooxygenation reactions, which
of gas and liquids is made possible in the annular gap between showed the general trend of higher conversion rates at increased
the rotating cylinder and the inner glass tube (approx. 1 mm; rotational speed. Finally, the authors applied their reactor concept
reactor volume: 8 mL). The reactant solution is pumped via the to the three-step synthesis of artemisinin 11 from
rotating cylinder to the bottom of the glass tube whereas the dihydroartemisinic acid 8 (Figure 12). Variation of the solvent,
oxygen-containing gas phase is directly drawn from the type of sensitizer and temperature resulted at 25 °C in 50% yield
surrounding atmosphere without the necessity of pressurizing it at full conversion with acidified meso-tetraphenylporphyrin (TPP)
(Figure 11). The glass tube itself can be cooled with a standard as dye sensitizer for 1O2 generation. The reaction solution (0.05
cryostat as well as the three LED arrays located in close proximity M 8 acid in toluene) was pumped through the reactor with a flow
to the glass tube (distance: 5 mm). In total, the arrays provide a rate of fl = 0.5 mL min-1, the cylinder was rotated with maximum
luminous flux of 21,000 lm (3 × 5 × 1400 lm, 5,000 K), which is speed of 4,000 rpm. Variation in the amount of trifluoroacetic acid
directed on the glass tube photoreactor. Polished aluminum (TFA) between 0.1 and 1 eq gave identical conversion and yield
blocks located in between the LED arrays are used for reflecting as for 0.5 eq. This result indicates excellent mixing properties in
any residual light scattered or refracted from the inside of the the reactor for catalytic applications.
reactor assembly. CFD studies have been performed with water
as solvent in order to visualize the vortex flow structure in the
annular gap. Already a rotational speed of approx. 955 rpm gave
a Taylor number of about 100,000, which is far above the critical
value of 1,700 for vortices formation. The reactor setup allows a
freely selectable rotational speed up to 4,000 rpm thereby offering
a broad regime for efficient gas-liquid contacting.

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
Figure 12. Schematic overview on the three-step synthesis of artemisinin 11 the present case Ioannou et al. use air or oxygen gas as reactive
with air as oxygen source: photooxigenation with 1O2; acid-catalyzed Hock
carrier phase and methanol as solvent. In consequence to these
cleavage (TFA = trifluoroacetic acid) and rearrangement; oxidation with 3O2
(other hydroperoxides and tautomers are omitted for clarity). different conditions and variation in flow rates the authors also
assume a larger droplet size of up to 60 µm in diameter. However,
this still conservative value for the droplet size implies an
2.2.3 Aerosol droplet formation for gas-liquid mixing exceedingly high surface to volume ratio of 100,000 m 2 m-3, which
is the base for a very efficient gas-liquid mass transfer made
In the previous example the liquid phase was in high excess available by the nebulizer reactor. Ioannou et al. applied their
compared to the gas phase take in by the rapidly rotating steel photoreactor setup to the continuous flow synthesis of ascaridol
cylinder. In the following example the ratio of liquid to gas is 13 (Figure 14). A solution of α-terpinene 12 (1.0 M in MeOH) and
reversed with a larger amount of gas being used to nebulize the Rose Bengal (0.8 mol%) was transferred to the nebulizer with a
liquid reaction solution for efficient gas-liquid mixing in a confined flow rate fl = 0.8 mL min-1 and sprayed with pure oxygen gas into
space. Ioannou et al. use a rather large Pyrex® glass cylinder the glass chamber (pgas = 4.14 bar). Under these conditions nearly

Accepted Manuscript
(length: 370 mm, diameter: 65 mm) with a front inlet for the full conversion was possible with a productivity of 8 g h -1. A
pneumatic nebulizer and a rear outlet for product solution.[39] Air maximum productivity of 14.2 g h-1 was achieved with higher
or oxygen gas can be introduced with a flow rate of fg = 0.7-1.1 L substrate concentration (2.0 M 12) and slightly lower liquid flow
min-1 in a pressure range of 2.76-4.14 bar. For the liquid phase a rate of fl = 0.75 mL min-1. Interestingly, the exchange of oxygen
flow rate of fl = 0.5-1.0 mL min-1 is recommended. White light gas with air resulted under the same process conditions in still a
producing LED stripes (3,800-4,200 K) are wrapped around the high productivity of 12.0 g h-1.
reactor chamber with a distance of 10 mm to the glass wall in
order to prevent heat transfer from the LEDs to the reactor (Figure
13).

Figure 14. Ascaridol synthesis with Rose Bengal as dye sensitizer for 1O2
formation in the aerosol.

2.2.4 Gas-liquid mixing under high pressure

The reactor concept comprises an upright standing high pressure


resistant sapphire glass tube (o.d. 10 mm, length: 120 mm, wall
thickness: 1 mm), which was filled with glass beads having a
diameter of 6 mm.[41] The incorporation of the spheres results in
better mixing of the gas-liquid mixture and a reduced light path d
between the inner wall of the glass tube and the bead surface (1
mm < d < 8 mm). Three high power LED arrays were applied for
irradiation with white light (5,000 K). A back pressure regulator
was installed to maintain the system pressure between 1-20 bar.
Penders et al. used their reactor concept for the dye sensitized
hydroxylation of arylboronic acid 14 in an aqueous ethanol
solution with high excess of air compared to the substrate solution
(Figure 15). The phases were combined in a micromixer and the
dispersion was transferred into the glass tube photoreactor. At
Figure 13. A) Schematic representation of the nebulizer photoreactor; B) Fully atmospheric pressure a conversion rate of 67% was achieved
operating nebulizer photoreactor assembled in a standard fume hood. with a 0.05 M solution of 14 in H2O-EtOH (1:1), Rose Bengal as
Reproduced with permission from ref. [39] copyright 2017 Wiley-VCH.
sensitizer (2 mol%) and Hünig’s base as reductive quencher (2
eq). The liquid flow rate was set to fl = 0.8 mL min-1 with a
corresponding gas flow rate of fg = 53.6 mL min-1 (fl : fg = 1 : 67).
The pneumatic nebulizer, the core element of the reactor, was
Conversion was increased to nearly 100% by increasing the
originally designed as part of trace-level analysis equipment for
system pressure to 20 bar while maintaining the liquid and gas
ICP-MS or ICP-OES analysis of aqueous samples. This device
flow rate at the same level.
allows the production of droplets with a diameter of approx. 6 µm
when operated with argon gas and water-based solutions.[40] In

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
Figure 16. Schematic representation and assembled reactor setup of a modified
lab rotary evaporator as photoreactor for thin liquid film formation in the turning
flask. Adapted with permission from ref. [42] copyright 2016 American Chemical
Society.

2.3 Microchannels in light-converting media

2.3.1 Luminescent solar concentrators with fixed emission


wavelength

Luminescent solar concentrators (LSC) are optical devices for


light harvesting and transportation.[43] Usually they exhibit a large
area for illumination with solar light and are made of glass or

Accepted Manuscript
polymeric material. The rather thin devices are doped with
fluorescent materials like dyes or quantum dots. [43b,44] The energy
of incident light with appropriate wavelength is absorbed by the
luminophores and is ideally emitted at longer wavelength again.
By total internal reflection the emitted light is waveguided inside
Figure 15. Hydroxylation of arylboronic acid 14 with Rose Bengal as dye the device to its thin edges, which leads to bright and
sensitizer for 1O2 formation and Hünig’s base as reductive quencher using a
concentrated refulgence. Based on this effect LSCs have been
high pressure resistant glass tube photoreactor with glass beads. Adapted with
permission from ref. [41] copyright 2015 Royal Society of Chemistry. utilized e.g. as solar light harvesting devices optically coupled to
small-area photovoltaic cells or as switchable windows in the case
of chemical host-guest systems composed of liquid crystals and
2.2.5 Gas-liquid mixing via rotating liquid thin film dyes.[45,46] Recently Cambié et al. applied the concept of LSCs for
harvesting and concentrating (diffuse solar) light at the walls of
In the last example, a semi continuously working photoreactor is microchannels, which were integrated into the LSC device.[47] The
described for gas-liquid reactions. Clark et al. modified a lab rotary emitted light of the dopant luminophore can then be absorbed by
evaporator for the generation of a liquid thin film in rotating glass a sensitizer dissolved in the reaction solution and fed into the
flasks of various volumes (Figure 16).[42] The substrate solution microchannels. In order to yield a perfect match between
and the reactive gas were fed into the glass flask with separate fluorescence emission of the LSC and absorption capability of the
peristaltic pumps. Two LED arrays were located in close proximity sensitizer the authors selected a commercially available perylene
to the flask (approx. 20 mm) providing white light with a luminous diimide dye (Lumogen F red 305) as dopant for the LSC and
flux of either 1,000 lm or 8,000 lm. The authors used this simple methylene blue as sensitizer for the photochemical process.
and versatile photoreactor as well for the synthesis of ascaridol Perylene dyes exhibit a broad absorption spectrum in the visible
13 in a 1,000 mL flask (Figure 14). α-Terpinene 12 (0.1 M in EtOH, region between 400 and 600 nm and offer a strong emission
5.0 mL) and Rose Bengal (1.0 mol%) were transferred into the between 600 and 700 nm, which fits perfectly to the absorption
flask and oxygen gas was bubbled through the solution. A band of methylene blue (Figure 17A). Prior to the fabrication of
custom-made UV-Vis spectrometer was used to analyze the film the photomicroreactor Cambié et al. performed in silico
thickness of the solution inside the rotating flask. With a rotation optimization of the channel dimension and architecture using
speed of 175 rpm a film thickness of approx. 260  40 µm was Monte-Carlo ray-tracing simulations. The optimal conditions were
obtained. The reaction solution was irradiated for 10 s with the found with a 50 × 50 × 3 mm3 device containing six microchannels
high power LEDs at 8,000 lm resulting in 72% yield, which of 500 µm width and 1 mm height (Figure 17B).
corresponds to a productivity of 22.2 g h -1. With this high value, Polydimethylsiloxane (PDMS) was selected as LSC microreactor
the rotary evaporator photoreactor outperforms the already pretty material due to its chemical stability, high transparency and
productive nebulizer photoreactor described above. moderate refractive index of n = 1.41, which is enough for total
internal reflection and minimized reflection on the outer surface. [45]
Various PDMS-dye mixtures were prepared with a dopant share
of 10-200 ppm. The reactor body was fabricated by casting the
dye-doped PDMS in 3D-printed mold with the positive relief of the
intended channel design. A plane dye-doped PDMS slab was
prepared as lid for the microchannel structures. Both slabs were
cured at 60 °C for 2 h and bonded via oxygen plasma treatment.
Residual PDMS was cut off and 1/16” PFA capillaries were press
fit into inlet and outlet of the LSC microreactor (Figure 17B). Now
photons can be collected inside the microchannels directly from
the incident light and by light-guided photons from the LSC reactor
body.

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW

Figure 18. [4+2] Cycloaddition of 9,10-diphenylanthracene 16 with in situ-


generated singlet oxygen.

Based on this first conceptual paper several other papers were

Accepted Manuscript
published in the following time by the same work group. A
theoretical study was performed by Cambié et al., which
investigates the possible pathways of photons into and inside
microstructured LSC devices: a) top reflection, b) transmission
through the device, c) emission due to insufficient total reflection
inside the device (escape cone), d) emission via the edges of the
device, e) general losses by e.g. matrix absorption, and most
important f) absorption by the reaction media directly from incident
light or from waveguided light.[48] The authors also investigated
the impact of the device size, channel height and channel pattern
Figure 17. A) Absorption (red) and emission (green) bands from the perylene as well as the impact of the photon path length across the LSC
dye as well as the absorption band of methylene blue (blue); B) final LSC
microreactor filled with methylene blue solution; C) Conversion rates by solely
material to the microchannel. This detailed theoretical study was
waveguided photons with light blocked microchannel; D) Conversion rates of done with a Monte-Carlo ray-tracing algorithm designed to model
fully irradiated LSC microreactor. Adapted with permission from ref. [47] LSCs.[49] The results gave important hints for the fabrication of
copyright 2017 Wiley-VCH. future LSC reactors. Large area LSC devices (1 m2) exhibit fewer
losses via edge emission than small devices (25 cm 2). The same
trend is seen for devices with larger channel heights and a greater
Cambié et al. characterized the novel photoreactor by applying it number of integrated channels, ideally as grid with a small gap
to the [4+2] cycloaddition of 9,10-diphenylanthracene 16 and in between the channels. Especially the latter parameter is
situ-generated singlet oxygen with a solar simulator as light important as it directly correlates with the possibility of photon
source (Figure 18). The substrate solution (0.1 mM in acetonitrile) reabsorption by the luminophore in bulk LSC material. For the
was mixed with the sensitizer solution (methylene blue, 0.2 mM in best theoretical device (1 m2 area, grid, 2.5 cm channel distance)
oxygen saturated acetonitrile) prior to the inlet of the LSC the following numbers were obtained for the photon balance: 4%
microreactor. The impact of waveguided photons has been direct reflection, 44% transmission, 4% non-radiative losses, 16%
analyzed by blocking direct irradiation of the microchannels with escape cone, <1% edge emission, 28% absorption of waveguided
a black card board (Figure 17C). For the non-doped LSC a photons in microchannel, and 3% direct absorption of photons in
minimum conversion of approx. 9% was obtained with a residence microchannel.
time of τres = 90 s. The 200 ppm-doped LSC reactor achieved 54% Another project was dedicated to a scale-up approach of LSC
conversion, which is an increase by a factor of 6 solely due to photoreactors via a numbering-up strategy.[50] Zhao et al.
waveguided photons, which were harvested and concentrated in investigated the impact of different zones in the microreactor
the non-blocked part of the reactor. The fully irradiated LSC structure: inlet, distributor, reaction zone, collector and outlet. 3D-
reactor with 200 ppm dopant achieves a conversion as high as printed molds were easily available for rapid prototyping of several
90% with a residence time of τres = 90 s (Figure 17D). The non- LSC reactor devices. As already mentioned in the foregoing
doped version converts approx. 40% of the substrate at the same theoretical paper the interchannel distance is an important
time. These results clearly show the high potential of harvesting parameter, which now has been evaluated in this practical work
and concentrating solar light for photochemical reactions in LSC in greater detail for the [4+2] cycloaddition of 16 with 1O2. An
devices. increase in interchannel distance (1.5  2.5  5 mm) results in
a significant increase in conversion (43%  59%  72%;
residence time for all experiments: τres = 12.5 s). A further increase
to 10 or even 20 mm is less effective for the conversion ( 80%
 95%). Both trends lead to the assumption that with increasing
interchannel distance a greater number of photons is harvested
and concentrated at the microchannels. In contrast to this positive
trend any further increase of bulk LSC material between the

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
channels also leads to reabsorption of waveguided photons and became possible (e.g. aryl amination with NiCl2∙glyme), as well as
to a higher degree of non-radiative losses. In consequence the the application of Eosin Y or Rose Bengal with the green LSC
process gets less efficient with a lower rise in conversion. Zhao et photomicroreactor for benzylamine oxidation or ascaridol
al. could also show that a symmetrical bifurcated distributor leads synthesis.
to an excellent flow distribution over up to 32 channels, whereas
a simple collection chamber overcomes high pressure drops in
the complete reactor.
Furthermore, Zhao et al. developed a real-time reaction control
system for their original LSC microreactor, which allows a more
or less constant photochemical process under fluctuating solar
daylight.[51] The idea behind this approach is the adjustment of the
residence time in the reactor to the incident solar light. In order to
get a measure of incident light on the reactor surface the authors

Accepted Manuscript
integrated a light sensor at the edge of the LSC device, which
detects the photon flux emitted there. This readout value is
proportional to the incident light and can be finally linked by
electronics to the piston pumps for varying the reaction solution’s
flow rate and residence time in the microreactor (Figure 19A).

Figure 20. Advanced LSC photomicroreactors for a broader application range


of light and catalyst classes. Adapted with permission from [52] copyright 2019
Wiley-VCH.

Figure 19. A) Schematic representation of the setup used; B) Conversion as


measure of process stability with reaction control system (red) and without
(green). Adapted with permission from ref. [51] copyright 2018 Royal Society of 2.3.2 Fluorescent fluids for photoreactors with variable
Chemistry. emission wavelength

Contrary to the LSC devices with fixed emission wavelengths


The set-up was applied to the well-known benchmark reaction of Zhang et al. designed a 3D-printable, two-channel reactor
[4+2] cycloaddition of 16 with 1O2. Strong fluctuations in concept, which allows a specific emission wavelength by using
conversion are observed for the non-controlled process with an different fluorescent fluids.[53] These luminophore-containing
amplitude between 55% to 97% (green line in Figure 19B). In the solutions are placed in the so-called light channel, which enfolds
case of the controlled process a significantly more constant the reaction channel. The physical principle is the same as for
conversion rate can be achieved with small fluctuations between LSC devices. Diffuse light is harvested by the dissolved
86% and 93% (red line in Figure 19B). luminophores in the light channel, and emitted as fluorescent light
Just recently, the 1st generation of LSC devices has been of specific wavelength to the reaction channel. In order to analyze
improved according to stability and light emission. [52] In order to and optimize the light illuminance distribution and uniformity on
achieve greater chemical stability Cambié et al. replaced the the reactor channel, the authors performed Monte-Carlo ray-
reactangular microchannels in the PDMS light guide by PFA tracing simulations of several light channel architectures, e.g.
capillaries, and self-made PDMS light guiding material was helical, linear or cylindrical (Figure 21 A). As supposed a full
substituted with commercially available PMMA light guiding enclosure of the reaction channel by a cylindrical light channel
devices (Figure 20). PMMA offers a higher refraction index (n = results in the best illuminance distribution and uniformity. In the
1.49 vs. n = 1.41 for PDMS), which results in consequence in an case of optimizing the reaction channel architecture, numerical
improved photon directing ability inside the light guiding device. In simulation analysis was done with the finite element method. Here,
addition, PMMA is more compatible to other classes of optimal mixing and diffusion was found for a 12-fold helical
luminophores and offers a lower dye degradation over time channel architecture being superior over a 6-fold helix and linear
compared to PDMS. Based on these new PMMA devices blue arrangement (Figure 21 B). Based on these results an appropriate
(430 nm) and green (510 nm) light emitting LSCs are available photomicroreactor was designed via computer-aided design and
now besides the original red one (640 nm). In the case of the blue transferred to a commercially available stereolithographic 3D-
LSC device, photochemistry with common Ru-based dyes printer (Figure 21 C and D).

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
the decreasing overlap between luminophore emission band and
sensitizer absorption band.

2.4 Falling film microreactor

Many examples discussed in the previous chapters are based on


tubular architectures for the transportation of gas and liquid
phases. The diameter of these closed capillaries is usually in the
range of several hundred µm to mm scale. Another concept is the
use of small-sized open microchannels with free access to the gas
phase throughout the complete channel length. [54] These open
microchannels are available in different architectures, e.g. as
array of 32 parallel channels with a width of 600 µm and a depth

Accepted Manuscript
of 200 µm wet-chemically etched into a stainless steel reaction
plate with 89 mm in length and 46 mm in width (Figure 22A).
These devices are usually operated in a tilt or even completely
upright manner. Hence, the liquid phase can flow down inside the
open microchannels by gravitational forces. In addition, capillary
forces result in the formation of a very thin liquid film, which lines
the wall of the microchannels (Figure 22B). Depending on the flow
rate and the viscosity of the liquid phase a film thickness below
50 µm is easily possible offering an excellent gas mass transfer
into the liquid phase with surface to volume ratio up to 20,000 m 2
Figure 21. A) Different light channel architectures enfolding a linear reaction m-3.[55] With these advantages combined a so-called falling film
channel. Illuminance distribution is given for the central cross-section (1 mm2)
of the reaction channel; B) Mixing efficiency of linear, 6-fold and 12-fold helical
microreactor (FFMR) was developed with a reaction plate as core
reaction channels; C) Final photomicroreactor with cylindrical light channel and element in vertical position. This reactor concept was primarily
helical reaction channel under UV irradiation; D) Fabrication process of 3D- used for highly exothermic reactions (e.g. hydrogenation of nitro
printed two-channel photomicroreactors as exemplified with a linear reaction groups) or as solvent evaporator due to its integrated heat
channel surrounded by a helical light channel. Adapted with permission from
[53] copyright 2019 Wiley-VCH.
exchanger (Figure 22C). [54,55]
The falling film approach has been also used for photochemical
applications, as exemplified at first by Griesbeck et al. using a
Zhang et al. used their reactor concept for the [4+2] cycloaddition XeCl excimer lamp (3 kW, length: 60 cm) as central light source
of in situ-generated singlet oxygen to 9,10-diphenylanthracene inside a water-cooled glass barrel.[56] The reaction solution was
(16). With Luminogen F Red 305 as luminophore in the light pumped from a reservoir to the top of the falling film glass barrel,
channel and methylene blue as sensitizer in the reaction channel from where it rinsed down along the barrel surface as thin film with
the same couple was used as in the work of Campié et al. (Figure intensive contact to the outer gas phase and with strong
18 and ref. [47]). As light source either a custom-made cylindrical irradiation from the central excimer lamp. In the case of the FFMR
LED array was used for blue light irradiation at 440 nm, or a solar any light must be incorporated through the inspection window at
simulator (1 sun). The 12-fold helical reaction channel has an the front of the microreactor. Apart from the possible use of
internal volume of 1.23 mL and was fed with 16 (0.1 mM in classical light sources, e.g. Xe arc lamps or compact fluorescence
acetonitrile) and methylene blue (0.2 mM in oxygen-saturated lamps,[57] modern high power LED technology allows the
acetonitrile) at equal flow rates. At a maximum luminophore wavelength-specific irradiation of the reactor chamber without
concentration of 400 ppm in the light channel a maximum complex filter equipment and less safety or heat management
conversion of 70 % was obtained within a residence time of τres = issues (Figure 22C). The small emitters can be arranged in arrays
85 s. By using a solar simulator as light source a conversion of and easily adopted to the appropriate reactor design. [58]
approx. 60 % could be achieved within a residence time of τres =
15 s, again with a maximum luminophore concentration of 400
ppm. The standard setup of Cambié et al. gives approx. 40%
conversion, but already at the half of luminophore concentration
(200 ppm; Figure 17D). Zhang et al. also performed the same
reaction with other dyes as luminophores in the light channel. In
the case of Eosin Y and Rhodamin 6G light harvesting and photon
transmission is only half as efficient as for Luminogen F Red 305.
Finally, fluorescein isothiocyanate was tested as well, but this dye
did not achieve any increase in conversion compared to a non-
doped solution in the light channel. Both results are in line with

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
chlorobenzenediazonium tetrafluoroborate 18 with furan (79%),
thiophene (72%) and pyridine (99%) in a single run with a very
short residence time of approx. τres = 9 s. Compared to the batch
tests with a reaction time of several hours a strong process
intensification has been achieved with the FFMR due to its
efficient irradiation of the liquid phase as thin film on the TiO 2
photocatalyst layer.

2.5 Dedicated reactor concepts for scale-up

2.5.1 Gas-liquid flow uniformity during numbring-up

The examples of this chapter focus on the developments of

Accepted Manuscript
continuous flow photoreactor concepts for larger scales
compared to the smaller lab devices presented before. The first
reactor system is based on a numbering-up strategy for gas-liquid
reactions in capillary photoreactors. Su et al. investigated a scale-
up methodology for parallelized 2-, 4- and 8-capillary photoreactor
systems applied to the dye sensitized aerobic oxidation of
Figure 22. A) Stainless steel FFMR reaction plate with 32 microchannels (width: thiophenol 20 (0.5 M in ethanol) to diphenyl disulfide 21 as test
600 µm, depth: 200 µm). Copyright Fraunhofer IMM; B) Schematic reaction. Rose Bengal was used as sensitizer (1 mol%) with 1 eq
representation of the falling film approach for microchannels: (S) stainless steel TMEDA as reductive quencher (Figure 24).[61]
microchannel reaction plate, (L) liquid film, (G) gas chamber, (W) microchannel
width and (H) height, (δl) liquid film height and (δg) gas chamber height. Adapted
with permission from ref. [55] copyright 2012 Elsevier; C) FFMR with LED arrays
magnetically fixed at the front plate (inset). Copyright Fraunhofer IMM.

Meanwhile the FFMR technology has been used for several


different photochemical reactions: a) photochlorination of toluene- Figure 24. Dye sensitized oxidation of thiophenol to diphenyl disulphide
2,4-diioscyanate, b) photooxygenation of cyclopentadiene, and c) (TMEDA = tetramethylethylenediamine).
the synthesis of juglone and ascaridole. [57,59] The aforementioned
reactions are gas-liquid reactions with e.g. homogeneously
dissolved sensitizers for the in situ generation of 1O2. The Actually, the focus of this work is on the assessment of the flow
application of a heterogeneous photocatalyst with the FFMR has non-uniformity in parallel channels used for the necessary gas-
been demonstrated recently by our group for the blue-light liquid two-phase flow splitting in up to eight photoreactors (Figure
mediated C-H arylation of heteroarenes with various diazonium 25A, B). The 2n-capillary photoreactor systems (n = 1, 2, 3) were
salts as substrates (Figure 23).[60] As solid catalyst material TiO2 built from standard equipment with PFA capillaries for the
(anatase phase) was used and immobilized as thin catalyst layer connections (i.d. 750 µm) and the reactors (i.d. 500 µm, 2.15 m
on the microchannel walls of the FFMR reaction plate. length, 0.95 mL) as well as T-micromixers (1/16”, PEEK). The
pressure drop in the distributer system was adjusted by PFA
capillaries with smaller diameter (i.d. 250 µm) and backpressure
regulators, respectively. White light-emitting LEDs are placed in
the capillary photoreactors, while pressurized air is used to
maintain the reactor temperature at approx. 22 °C. The ratio of
the gas to liquid flow rate was set to 3:1. The authors define a
standard deviation (SD) between the actual mass flow rate in
Figure 23. Blue-light mediated C-H arylation of pyridine with TiO2 as each capillary reactor and the average total mass flow rate for all
photocatalyst immobilized onto the FFMR microchannel walls. capillaries. The SD is used to compare the various 2n-reactor
systems in dependence to the average liquid flow rate. In the case
of the non-photocatalytical assessment no oxygen gas is
The diazonium salt substrates (0.05 M) were dissolved in a 1:1 consumed. For all 2n-capillary reactor systems a very low SD <5%
mixture of ethanol with furan, thiophene or pyridine. The reaction was achieved at average liquid flow rates of fl = 0.4-0.7 mL min-1
solution was fed into the FFMR with a flow rate of fl = 0.5 mL min- per capillary. These results are a very good improvement
1
while maintaining the temperature inside the reactor at 20 °C. compared to the literature with values >10%.[62] The authors state
The LED array was switched on providing light with an emission that pressure drop management inside the distributer part was
wavelength of 455 nm (Pele  2.4 W). Under these conditions good very important to achieve a stable and uniformly distributed gas-
to excellent yields were obtained for e.g. 4- liquid flow. In the case of the 4-capillary reactor system a

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
backpressure regulator was necessary and set to Δp = 0.34 bar. yield (D) per single capillary. Adapted with permission from ref. [61] copyright
2016 Royal Society of Chemistry.
In the case of the 8-capillary reactor system the intrinsic pressure
drop of the capillaries and the T-micromixers was high enough to
generate a uniform gas-liquid distribution throughout the complete
system without backpressure regulation. Based on these results 2.5.2 Low power light for scale-up
the planned photocatalytic gas-liquid reaction was investigated
with oxygen gas being consumed during the process. The real In the second example Gonçalves et al. describe the multigram
process conditions gave a higher SD, but still under 10%. This synthesis of hypericin 23, a natural product rarely available by
result is the consequence of the quick oxygen gas consumption direct extraction from St. John’s Wort as main herbal source. [63]
and accompanying changes in the hydrodynamic behavior of the The photocyclisation of protohypericin 22 is the final step in the
gas-liquid flow. As seen in Figure 25C and D liquid flow rate and synthesis of 23 yielding a compound with strong light absorption
yield differ for each of the eight capillary reactors, but the relative in the orange region at λmax = 596 nm (Figure 26). In previous
deviations are smaller than 4%, which can be also assumed for synthesis routes high power light sources with multiband

Accepted Manuscript
consecutive runs of a single photocapillary reactor. Hence, the emissions have been used, which always show a strong overlap
work of Su et al. is an impressive example for the scale-up of with the absorption band of hypericin. Gonçalves et al. suggest
photocatalytic gas-liquid processes via the numbering-up strategy that detrimental photobleaching of hypericin is the reason for the
with stable gas-liquid flow distribution over several photoreactors. moderate to bad productivity of the processes reported so far.

Figure 26. Efficient photocyclisation of protohypericin to hypericin with low-


power red light.

Kinetic studies with small batch samples of 22 (2 µL, c = 1.6 × 10-


5
mol L-1) performed with a high power Hg lamp (400 W) resulted
in 79% yield for 23 after seven minutes, whereas after a prolonged
irradiation time of 35 min only 4% yield was remaining. By
switching from multiband light emission to single band emission
with LED arrays the yield for 23 increased dramatically. In the
case of blue light with λem = 463 nm and a low power of 10.2 mW
cm-2 23 was obtained in 97% yield after 20 min, which decreased
to 89% yield after a prolonged irradiation time of 800 min. Finally
the use of lowest power LED arrays for red light (λem = 629 nm,
0.6 mW cm-2) gave quantitative yield after 74 min. This excellent
result can be attributed to the sufficient overlap of red light
emission with the spectral absorption of 22 and a negligible
overlap with the spectral absorption of 23 (Figure 27A). Based on
these findings Gonçalves et al. designed a glass capillary
photoreactor module with a Pyrex® tube in zigzag-format (i.d. 3
mm, length 180 cm) sandwiched in between two glass slabs.
Each side of this flow device was equipped with three LED arrays
carrying 108 diodes for an emission wavelength of 629 nm. In total
four photoreactor modules of this type were stacked above each
other and used in series (Figure 27B). Starting material 22 (5.0 g)
was dissolved in three liters of acetone (c = 3.3 × 10-3 mol L-1) and
fed into the reactor modules with a very high flow rate of approx.
fl = 1.8 L min-1. The light dose was set to 4.0 mW cm-2. Within an
Figure 25. Schematic representation (A) and photograph (B) of the 8-capillary
photoreactor system used for the investigation of flow non-uniformity of gas-
overall reaction time of 13.6 min the reaction solution was
liquid mixed phases in parallel channels. Relative deviation of flow rate (C) and circulated eight times for full conversion and 97% yield after

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
recrystallization (4.85 g of 23). The authors could show that a productivity 1 kg per day and a safe containment of a standard
defined control of irradiation wavelength and power can have a UV light source with an energy consumption of 1-5 kW. With these
tremendous effect on the productivity of a photochemical reaction. specifications at hand Elliott et al. constructed the so-called Firefly
This is especially important for the scale-up of expensive natural reactor.[66] In sum 2 × 12 glass tubes are arranged in two circles
products with costs of approx. 300 € mg-1 (95% purity by around the central light source and are contained in between two
extraction from St. John’s Wort / hypericum perforatum).[64] glass cylinders (Figure 28). The resulting annular cavity formed
by the cylinders can be flooded with coolant to maintain the
desired temperature in the reaction tubes. The top and bottom
plate of the reactor exhibit channels, which interconnect the glass
tubes in a serial manner resulting in an overall reactor volume of
120 mL. The central light source can be easily exchanged and
adapted to the photochemical process. A fan is installed at one
plate of the reactor and is used to remove the hot air and ozone

Accepted Manuscript
from the gap between light source and interior glass cylinder.
Finally, the complete outer glass cylinder is surrounded by a
polished and earthed metal tube, which reflects any residual UV
light back into the reactor and serves as protection of the
experimenter against the intense UV radiation. As planned the
fully assembled Firefly reactor is compact and can be easily used
in a standard fume hood.

Figure 28. A) Schematic representation and photograph of the Firefly


photoreactor as scale-up approach to classical FEP capillary photoreactors.
Adapted with permission from ref. [65] copyright 2016 American Chemical
Society; B) exploded view of the reactor as described in the corresponding
patent application. Adapted from patent application WO2018011550A1, ref. [66].

Figure 27. A) Visible light spectral changes upon 8-times cycling of the reaction
solution through the photoreactor modules (left), and spectral overlap between
Elliott et al. applied their novel reactor to several photochemical
red LED light emission and absorption spectra of 22 (right); B) Schematic
representation of the glass capillary photoreactor setup for the red-light processes and compared the productivity with classical FEP
mediated photocyclisation of protohypericin to hypericin. Adapted with capillary reactors and literature data. The intramolecular [2+2]
permission from ref. [63] copyright 2017 American Chemical Society. photocycloaddition of tricyclic dione 24 has been efficiently
produced with a capillary photoreactor yielding 6.13 g h-1 of
“Cookson’s Dione” 25 based on a 0.1 M substrate solution (Figure
2.5.3 Scale-up for kg production 29).[25] In the case of the Firefly reactor a 0.5 M solution was
irradiated in a single run of 140 min at 1.5 kW light source power
Capillary photoreactors have been intensively used in the last and resulted in 167 g h-1 of product. By doubling both the
years and promoted the field of flow photochemistry especially on substrate concentration to 1.0 M and the light source power to 3
the academic level. Scale-up approaches resulted in multi- kW a doubled productivity of 331 g h-1 was achieved. A kilogram-
capillary systems, which still have the problems related to the scale run was made possible under these optimized conditions
polymeric material FEP during long-term use: a) FEP is not fully within an overall process time of approx. 3.5 h yielding 1,165 g of
UV transparent and can degrade upon strong UV irradiation; b) product, which crystallized in the receiving flask. Just recently,
mechanical stress like abrasion or kinks can result in weak spots Clark et al. also performed this reaction with their modified lab
or even rupture of the tubing and complete replacement. In order rotary evaporator photoreactor in batch mode (introduced in
to circumvent particularly the UV degradation problem Elliott et al. Figure 16) and achieved a productivity of 210.6 g h-1.[67]
designed a novel high capacity photoreactor based on quartz
glass tubes.[65] The new concept was planned for a higher

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
processes. During visits at industrial fairs one can realize the
interest of the chemical industry in novel technology for a
sustainable production. But novel production technology and
synthesis concepts need new hardware and well-trained and
educated people to set-up and run such new plants. Hence,
academic training at universities must also reflect the new needs
in industry to prepare future employees for new challenges in a
Figure 29. Photochemical synthesis of “Cookson’s Dione” 25 via [2+2]
changing industrial market. Students must be educated in flow
photocycloaddition.
(photo)chemistry to get to know the benefits of this technology
and for being able to discover the potential of “old” batch
processes to be translated to flow mode. [68]
Another kilogram-scale process was presented for the cyclisation
The research in visible-light photocatalysis has contributed a lot
of phthalimid 6 exemplified before by Griesbeck and DeLaney,
in translating classic chemical concepts with e.g. costly noble
respectively (introduced in Figure 10).[56,35] Griesbeck et al.

Accepted Manuscript
metal catalyst systems and rather harsh reaction conditions to
reported a productivity of 7 g h-1 for their falling film reactor,
more green processes with higher selectivity. The last ten years
whereas DeLaney et al. achieved 0.78 g h-1 with their glass tube
proved that photons can be used as a delicate energy source or
photoreactor. Elliott et al. outperformed these values by far with a
“traceless” reagents in synthetic organic chemistry. The on-going
productivity of 57.9 g h-1 (0.2 M, 30 mL min-1, 3 kW). Interestingly,
rapid development of novel synthesis routes and
the authors used acetonitrile instead of acetone as co-solvent to
photocatalytically active materials opens the door for replacing
water, pointing out that acetone is not necessary as triplet
well-known chemistry with new reactivity concepts in bond
sensitizer for this photodecarboxylative cyclisation. These two
activation and formation via direct or sensitized interaction with
examples clearly show the great potential of the Firefly reactor
light. Based on these achievements photochemistry got more and
concept. The authors conclude that the Hg lamp is not an integral
more into the focus of larger public funded programs, especially
of the reactor. It can be replaced either by future high power UV
with the prominent idea of using solar light to run chemical
LEDs or other LEDs to match especially the strongly increasing
reactions.[69] National programs combine photocatalysis with e.g.
number of sensitized photoreactions utilizing visible light.
activation of atmospheric/anthropogenic CO2 and its conversion
to energy carriers or bulk chemicals. [70] International programs
and cooperations foster academic research between countries,
4. Conclusion and perspective whose technology and geographical location can provide
synergies for solar photochemistry. In addition several European
As mentioned already at the beginning of this manuscript
programs clearly call up industrial participation to advance the use
continuous flow photochemistry has run through a major evolution
of solar energy for industrial processes.[71] At this stage a multi-
in the last decade. The selected examples depict an amazing
disciplinary approach is necessary, which combines and aligns
increase in reactor concepts, which have been designed from
the expertise of chemistry, physics, biology, material development
scratch or advanced from already existing architectures. The
and engineering. Finally, scientific long-term programs have been
important fact about this development is clearly the acceptance of
rated just recently as a most important chance to solve global
continuous flow technology and its advantages for
problems of energy consumption and climate change on an
photochemistry. Research groups worldwide are no longer afraid
international level.[72] Solar light and photochemistry play pivotal
to use technology for photochemical synthesis and become
roles in the projected concepts. Therefore, we can all hope that
convinced by the results and advantages of these flow devices.
photochemistry and its adjunct technologies will keep track on
Solid catalysts can be applied meanwhile without immobilization
their way to a conceptual chemical technology with disruptive
as well as challenging reactions with e.g. singlet oxygen can be
impact on sustainable production of energy carriers, everyday
performed under strictly controlled conditions. In parallel
commodities and specialty chemicals.
photoreactors became characterized not only at the chemical
level by running and comparing benchmark reactions. Detailed
investigations were done on the engineering level to understand
the way and flux of light interacting with gas, liquid and solid Acknowledgements
phases inside defined reactor architectures. These results can
help to utilize incident (solar) light even more efficiently inside flow The author would like to thank MSc Christoph Deckers, Dr.
photoreactors and reduce energy costs. Finally, flow Gabriele Menges-Flanagan and Dr. Stefan Kiesewalter (all
photoreactors have also entered the commercial market as Fraunhofer IMM) for proofreading. The author would also like to
optional part of already existing continuous flow “machines” for thank Lisa Pokropp (Fraunhofer IMM) for creating the frontispiece.
thermal chemistry. Hence, it is consequential, that photoreactors This work has been additionally supported by the German Federal
become re-designed to larger scales in order to be promoted for Ministry of Education and Research (BMBF) under the program
industrial applications. Especially any future integration into CO2Plus (Project CarbonCat; grant number: 033RC009A).
chemical industry will stimulate the adaptation of flow
photoreactors to worldwide standardized plant components with Keywords: photocatalysis • continuous flow • microreactor •
an increased compatibility and comparability of photochemical intensification • mixing

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
100–103; c) J. Metternich, R. Gilmour, J. Am. Chem. Soc.
[1] a) Chemical Photocatalysis. (Ed.: Burkhard König), 2016, 138, 1040–1045; d) J. McManus, D. Nicewicz, J. Am.
DeGruyter, Berlin, 2013; b) H. Kisch, Semicondurctor Chem. Soc. 2027, 139, 2880–2883; e) L. Wozniak, G.
Photocatalysis – Principles and Applications. Wiley-VCH, Magagnano, P. Melchiorre, Angew. Chem. 2018, 130, 1080–
Weinheim, 2015; c) Natural and Artificial Photosynthesis – 1084, Angw. Chem. Int. Ed. 2017, 56, 1068–1072; f) I.
Solar Power as an Energy Source. (Ed.: Reza Razeghifard), Ghosh, R. Shaikh, B. König, Angew. Chem. 2017, 129,
Wiley, New Jersey, 2013; d) Visible Light Photocatalysis in 8664–8669, Angew. Chem. Int. Ed. 2017, 56, 8544–8549; g)
Organic Chemistry. (Eds.: C. Stephenson, T. Yoon, D. G. Goti, B. Bieszczad, A. Vega-Peñaloza, P. Melchiorre,
MacMillan), Wiley-VCH, Weinheim, 2018; e) C. Prier, D. Angew. Chem. 2019, 131, 1226–1230, Angew. Chem. Int.
Rankic, D. MacMillan, Chem. Rev. 2013, 113, 5322–5364; Ed. 2018, 57, 1213–1217; h) T. Chisholm, D. Clayton, L.
f) D. Hari, B. König, Chem. Commun. 2014, 50, 6688–6699; Dowman, J. Sayers, R. Payne, J. Am. Chem. Soc. 2018, 140,
g) N. Romero, D. Nicewicz, Chem. Rev. 2016, 116, 10075– 9020–9024; i) C. Kerzig, X. Guo, O. Wenger, J. Am. Chem.
10166; h) M. Kärkäs, J. Porco, C. Stephenson, Chem. Rev. Soc. 2019, 141, 2122–2127; j) N. Ichiishi, J. Caldwell, M. Lin,

Accepted Manuscript
2016, 116, 9683–9747; i) V. Ramamurthy, J. Sivaguru, W. Zhong, X. Zhu, E. Streckfuss, H.-Y. Kim, C. Parish, S.
Chem. Rev. 2016, 116, 9914–9993; j) L. Marzo, S. Pagire, Krska, Chem. Sci. 2018, 9, 4168–4175; k) J. Li, D. Zhu, L.
O. Reiser, B. König, Angew. Chem. 2018, 130, 10188– Lv, C.-J. Li, Chem. Sci. 2018, 9, 5781–5786; l) K. Singh, S.
10228, Angew. Chem. Int. Ed. 2018, 57, 10034–10072; k) Staig, J. Weaver, J. Am. Chem. Soc. 2014, 136, 5275–5278;
D. Ravelli, M. Fagnoni, A. Albini, Chem. Soc. Rev. 2013, 42, m) B. Liu, C.-H. Lim, G. Miyake, J. Am. Chem. Soc. 2017,
97–113; l) R. Brimioulle, D. Lenhart, M. Maturi, T. Bach, 139, 13616–13619; n) D. Rombach, H.-A. Wagenknecht,
Angew. Chem. 2015, 127, 3944–3963, Angew. Chem. Int. ChemCatChem 2018, 10, 2955–2961; o) I. Jurberg, H.
Ed. 2015, 54, 3872–3890; m) J. Beatty, C. Stephenson, Acc. Davies, Chem. Sci. 2018, 9, 5112–5118.
Chem. Res. 2015, 48, 1474–1484; n) D. Staveness, I. [5] a) I. Ghosh, T. Ghosh, J. Bardagi, B. König, Science 2014,
Bosque, C. Stephenson, Acc. Chem. Res. 2016, 49, 2295– 346, 725–728; b) I. Ghosh, B. König, Angew. Chem. 2016,
2306; o) J.-R. Chen, X.-Q. Hu, L.-Q. Lu, W.-J. Xiao, Acc. 128, 7806–7810, Angew. Chem. Int. Ed. 2016, 55, 7676–
Chem. Res. 2016, 49, 1911–1923; p) O. Reiser, Acc. Chem. 7679; c) H.-X. Gong, Z. Cao, M.-H. Li, S.-H. Liao, M.-J. Lin,
Res. 2016, 49, 1990–1996; q) Y.-Q. Zou, F. Hörmann, T. Org. Chem. Front. 2018, 5, 2296–2302; c) J. Haimerl, I.
Bach, Chem. Soc. Rev. 2018, 47, 278–290; r) T. Yoon, M. Ghosh, B. König, J. Vogelsang, J. Lupton, Chem. Sci. 2019,
Ischay, J. Du, Nat. Chemistry 2010, 2, 527–532. 10, 681–687; L. Zeng, T. Liu, C. He, D. Shi, F. Zhang, C.
[2] a) J. Tellis, C. Kelly, D. Primer, M. Jouffroy, N. Patel, G. Duan, J. Am. Chem. Soc. 2016, 138, 3958–3961.
Molander, Acc. Chem. Res. 2016, 49, 1429–1439; b) X. [6] a) Photochemical Processes in Continuous-Flow Reactors –
Lang, J. Zhao, X. Chen, Chem. Soc. Rev. 2016, 45, 3026– From Engineering Principles to Chemical Applications. (Ed.:
3028; c) K. Skubi, T. Blum, T. Yoon, Chem. Rev. 2016, 116, Tmothey Noël), World Scientific, London, 2017; b) Y. Su, N.
10035–10074; d) J. Milligan, J. Phelan, S. Badir, G. Straathof, V. Hessel, T. Noël, Chem. Eur. J. 2014, 20,
Molander, Angew. Chem. Int. Ed. 2019, 58, 6152–6163; d) 10562–10589; c) J. Knowles, L. Elliott, K. Booker-Milburn,
I. Perry, T. Brewer, P. Sarver, D. Schultze, D. DiRocco, D. Beilstein J. Org. Chem. 2012, 8, 2025–2052; d) K. Gilmore,
MacMillan, Nature, 2018, 560, 70–75; e) Q.-Y. Meng, S. P. Seeberger, Chem. Rec. 2014, 14, 410–418; e) D. Cambié,
Wang, B. König, Angew. Chem. 2017, 129, 13611–13615, C. Bottecchia, N. Straathof, V. Hessel, T. Noël, Chem. Rev
Angew. Chem. Int. Ed. 2017, 56, 13426–13430; f) M. 2016, 116, 10276–10341; f) F. Politano, G. Oksdath-
Hopkinson, B. Sahoo, J.-L. Li, F. Glorius, Chem. Eur. J. Mansilla, Org. Process Res. Dev., 2018, 22, 1045–1062; g)
2014, 20, 3874–3886; g) D. Fabry, M. Rueping, Acc. Che. E. Bremus-Köbberling, A. Gillner, F. Avemaria, C. Réthore,
Res. 2016, 49, 1969–1979; h) M. Silvi, P. Melchiorre, Nature, S. Bräse, Beilstein J. Org. Chem. 2012, 8, 1213–1218; h) B.
2018, 554, 41–49; i) B. Mühldorf, R. Wolf, Angew. Chem. Anderson, W. Bauta, W. Cantrell, Org. Process Res. Dev.
2016, 128, 437–441, Angew. Chem. Int. Ed. 2016, 56, 2012, 16, 967–975; i) C. Shen, Y. Wang, J. Hu, G. Luo,
5342–5345; j) C. Seel, A. Králik, M. Hacker, A. Frank, B. Chem. Eng. J. 2015, 277, 48–55; j) M. Sellaro, M. Bellardite,
König, T. Gulder, ChemCatChem, 2018, 10, 3960–3963. A. Brunetti, E. Fontananova, L. Palmisano, E. Drioli, G.
[3] a) M Majek, A. von Wangelin, Acc. Chem. Res. 2016, 49, Barbieri, RSC Adv. 2016, 6, 67418–67427; k) T. Junkers, B.
2316–2327; b) S. Protti, D. Ravelli, M. Fagnoni, Photochem. Wenn, React. Chem. Eng. 2016, 1, 60-64; l) M. Baumann, I.
Photobiol. Sci. 2019, 18, 2094–2101; c) O. Wenger, Chem. Baxendale, React. Chem. Eng. 2016, 1, 147–150; m) B. da
Eur. J. 2019, 25, 6043–6052; d) J.-R. Chen, D.-M. Yan, Q. Costa Filho, A. Araujo, B. Silva, R. Boaventura, M. Dias, J.
Wie, W.-J. Xiao, ChemPhotoChem 2017, 1, 148–158; e) V. Lopes, V. Vilar, Chem. Eng. J. 2017, 310, 331–341; n) Y.
Balzani, G. Bergamini, P. Ceroni, ANgew. Chem. 2015, 127, Pihosh, J. Uemera, I. Turkevych, K. Mawatari, Y. Kazoe, A.
11474–11492, Angew. Chem. Int. Ed. 2015, 54, 2–20. Smirnova, T. Kitamori, Angew. Chem. 2017, 129, 8242–
[4] a) E. Areco, I. Jurberg, A. Álvarez-Fernández, P. Melchiorre, 8245, Angew. Chem. Int. Ed. 2017, 56, 8130–8133; o) A.
Nat. Chemistry 2013, 5, 750–756; b) M. Pirnot, D. Rankic, D. Kouridaki, K. Huvaere, React. Chem. Eng. 2017, 2, 590–
Martin, D. MacMillan, Science 2013, 39, 1593–1596; c) H. 597; p) D. Heggo, S. Ookawara, Chem. Eng. Sci. 2017, 169,
Hoa, X. Shen, C. Wang, L. Zhang, P. Röse, L.-A. Chen, K. 67–77; q) X.-Z. Fan, J.-W. Rong, H.-L. Wu, Q. Zhou, H.-P.
Harms, M. Marsch, G. Hilt, E. Meggers, Science, 2014, 515, Deng, J. Tan, C.-W. Xue, L.-Z. Wu, H.-R. Tao, J. Wu, Angew.

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
Chem. 2018, 130, 8650–8654, Angew. Chem. Int. Ed. 2018, Photobiol. Sci., 2016, 15, 1103–1114; e) F. Würthner, C.
57, 8514–8515; r) F. Kloss, T. Neuwirth, V. Haensch, C. Saha-Möller, B. Fimmel, S. Ogi, P. Leowanawat, D. Schmidt,
Hertweck, Angew. Chem. 2018, 130, 14684–14689, Angew. Chem. Rev., 2016, 116, 962-1052; f) M. Enoki, R. Katoh,
Chem. Int. Ed. 2018, 57, 14476–14481; s) M. Lesieur, C. Photochem. Photobiol. Sci., 2018, 17, 793–799.
Battilocchio, R. Labes, J. Jacq, C. Genicot, S. Ley, P. Pasau, [13] a) A. Cassano, C. Martín, R. Brandi, O. Alfano, Ind. Eng.
Chem. Eur. J. 2018, 24, 1203–1207; t) M. Czarnecki, P. Chem. Res., 1995, 34, 2155–2201; b) T. van Gerven, G. Mul,
Wessig, Org. Process. Res. Dev. 2018, 22, 1823–1827; u) J. J. Moulijn, A. Stankiewicz, Chem. Eng. Process., 2007, 46,
Williams, M. Nakano, R. Gérady, J. Rincón, Ó. de Frutos, C. 781–789; c) D. Ziegenbalg, B. Wriedt, G. Kreisel, D. Kralisch,
Mateos, J.-C. Monbaliu, O. Kappe, Org. Process Res. Dev. Chem. Eng. Technol., 2016, 39, 123–134; d) K. Loubière, M.
2019, 23, 78–87; v) Z. Hamami, L. Vanoye, P. Fongarland, Oelgemöller, T. Aillet, O. Dechy-Cabaret, L. Prat, Chem. Eng.
C. de Bellefon, A. Favre-Réguillon, ChemPhotoChem 2019, Process., 2016, 104, 120–132; e) M. Sender, D. Ziegenbalg,
3, 122–128; w) D. Cambié, T. Noël, Top. Curr. Chem. 2018, Chem. Ing. Tech., 2017, 89, 1159–1173; f) M. Martín-Sómer,
376, 45; x) L. Lin, K. Xie, M. Beaucamp, N. Job, M. Penhoat, C. Pablos, R. van Grieken, J. Marugán, Appl. Catal. B

Accepted Manuscript
ChemPhotoChem 2019, 3, 198–203; x) J. Schachtner, A. Enironment, 2017, 215, 1–7; g) B. Wriedt, D. Kowalczyk, D.
von Wangelin, Beilstein J. Org. Chem. 2016, 12, 1798–1811; Ziegenbalg, ChemPhotoChem, 2018, 2, 913–921; f) R.
y) Special Issue: Flow Photochemistry in ChemPhotoChem Radjagobalou, J.-F. Blanco, O. Dechy-Carbaret, M.
2018, 2(10), Oelgemöller, K. Loubiere, Chem. Eng. Proc. Process
https://onlinelibrary.wiley.com/toc/23670932/2018/2/10 Intensification, 2018, 130, 214–228; h) E. Moschetta, S.
(website accessed 04.03.2019). Richter, S. Wittenberger, ChemPhotoChem 2017, 1, 539–
[7] a) K. Jähnisch, V. Hessel, H. Löwe, M. Baerns, Angew. 543; i) G. Glotz, O. Kappe, React. Chem. Eng. 2018, 3, 478–
Chem. 2004, 116, 410–451, Angew. Chem. Int. Ed., 2004, 486; j) K. Loponov, J. Lopes, M. Barlog, E. Astrova, A.
43, 406–446; b) Flow Chemistry - Fundamentals (Vol. 1) & Malkov, A. Lapkin, Org. Process Res. Dev. 2014, 18, 1443–
Applications (Vol. 2). (Eds.: F. Darvas, V. Hessel, G. 1454.
Dorman), De Gruyter, Berlin, 2014; c) M. Plutschak, B. [14] a) Y. Sun, Y. Jiang, X. Sun, S. Zhang, S. Chen, Chem. Rec.
Pieber, K. Gilmore, P. Seeberger, Chem. Rev., 2017, 117, 2019, 19, 1l–25; b) S. Li, S.-C. Tan, C. Lee, E.
11796–11893; d) I. Rossetti, M. Compagnoni, Chem. Eng. J. Waffenschmidt, S. Hui, C. Tse, IEEE Trans. Power Electron.
2016, 296, 56–70. 2016, 31, 1503–1516; c) M. Kneissl, T. Kolbe, C. Chua, V.
[8] a) Micro Process Engineering: A Comprehensive Textbook. Kueller, N. Lobo, J. Stellmach, A. Knauer, H. Rodriguez, S.
(Eds.: V. Hessel, A. Renken, J. Schouten, J.-I. Yoshida), Einfeldt, Z. Yang, N. Johnson, M. Weyers, Semicond. Sci.
Wiley-VCH, Weinheim, 2013; b) Novel Process Windows: Technol. 2011, 26, 14036–14042; d) P. Xiao, J. Huang, Y.
Innovative Gates to Intensified and Sustainable Chemical Yu, B. Liu, Molecules, 2019, 24, 151–179; e) L. Quan, P. de
Process. (Eds.: V. Hessel, D. Kralisch, N. Kockmann), Wiley- Qrquer, R. Sabatini, E. Sargent, Adv. Mater. 2018, 30,
VCH, Weinheim, 2015; c) Sustainable Flow Chemistry. (Ed.: 1801996–18022015; f) Y. Wei, Z. Cheng, J. Li, Chem. Soc.
Luigi Vaccaro), Wiley-VCH, Weinheim, 2017; Rev. 2019, 48, 310–350; g) T. Wu, C.-W. Sher, Y. Lin, C.-F.
[9] a) T. H. Rehm, Continuous-flow photochemistry in Lee, S. Liang, Y. Lu, S.-W. Chen, W. Guo, H.-C. Kuo, Z.
microstructured environment. in: Flow Chemistry – Chen, Appl. Sci. 2018, 8, 1557–1574; h) S. Nakamura,
Applications (Eds.: F. Darvas, V. Hessel, G. Dorman), De Angew. Chem.2015, 127, 7880–7899, Angew. Chem. Int. Ed.
Gruyter, Berlin, 2014; b) K. Mizuno, Y. Nishiyama, T. Ogaki, 2015, 54, 7770–7788; i) D. Ziegenbalg, G. Kreisel, D. Weis,
K. Terao, H. Ikeda, K. Kakiuchi, J. Photochem. Photobiol. C: D. Kralisch, Photochem. Photobiol. Sci. 2014, 13, 1005–
Photochem. Reviews, 2016, 29, 107–147. 1015; j) C. Haas, T. Roider, R. Hoffmann, U. Tallarek, React.
[10] P. Atkins, J. de Paula, Atkins’ Physical Chemistry. Oxford Chem. Eng., 2019, DOI: 10.1039/c9re00339h.
University Press, Oxford, 2014. [15] P. Klán, J. Wirz, Photochemistry of Organic Compounds –
[11] a) A. Pistorius, W.DeGrip in Encyclopedia of Spectroscopy From Concepts to Practice. Wiley, Chichester, 2009.
and Spectrometry, 2nd Edition. (Eds.: J. Lindon, G. Tranter, [16] a) E. Kowalska, S. Rau, Recent Patents Eng., 2010, 4, 242–
D. Koppenaal), Academic Press, Amsterdam, 2010, p.142– 266; b) M. Chong, B. Jin, C. Chow, C. Saint, Water Res.,
152; b) A. Rodger, K. Sanders in Encyclopedia of 2010, 44, 2997–3027; c) A. Deegan, B. Shaik, K. Nolan, K.
Spectroscopy and Spectrometry, 2nd Edition. (Eds.: J. Lindon, Urell, M. Oelgemöller, J. Tobin, A. Morrissey, Int. J. Environ.
G. Tranter, D. Koppenaal), Academic Press, Amsterdam, Sci. Tech., 2011, 8, 649–666; d) R. Portela, R. Tessinari, S.
2010, p.166–173; c) T. Mayerhöfer, H. Mutschke, J. Popp, Suárez, S. Rasmussen, M. Hernández-Alonso, M. Canela,
ChemPhysChem 2016, 17, 1948–1955; d) T. Mayerhöfer, J. P. Ávila, B. Sanchez, Environ. Sci. Technol., 2012, 46,
Popp, ChemPhysChem 2019, 20, 511–515. 5040–5048; e) I. Pibiri, S. Buscemi, A. Piccionello, A. Pace,
[12] a) D. Xu, D. Neckers, J. Photochem. Photobiol. A: Chemistry, ChemPhotoChem 2018, 2, 535–547; f) D. García-
1987, 40, 361–370; b) O. Valdes-Aguilera, D. Neckers, Acc. Fresnadillo, ChemPhotoChem 2018, 2, 512–534.
Chem. Res., 1989, 22, 171–177; c) V. Buss, L. Eggers in [17] a) L. Minguu, W. Daud, M. Kassim, Int. J. Hydrogen Ener.,
Encyclopedia of Spectroscopy and Spectrometry, 2nd Edition. 2010, 35, 5233–5244; b) Z. Xing, J. Pan, L. Wang, Chem.
(Eds.: J. Lindon, G. Tranter, D. Koppenaal), Academic Press, Eng. Sci., 2013, 104, 125–146; c) C. Acar, I. Dincer, G.
Amsterdam, 2010, p.442–450; d) B. Heyne, Photochem. Naterer, Int. J. Energy Res., 2016, 40, 1449–1473; d) Ü.

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
Taştan, D. Ziegenbalg, Chem. Eur. J., 2016, 22, 18824– [33] a) B. Shen, M. Bedore, A. Sniady, T. Jamison, Chem.
18832; e) B. Cecconi, N. Manfredi, T. Montini, P. Fornasiero, Commun., 2012, 48, 7444–7446; b)
A. Abbotto, Eur. J. Org. Chem. 2016, 5194–5215; K. https://www.aceglass.com/html/3dissues/Photochemistry-
Christoforidis, P. Fornasiero, ChemCatChem, 2017, 9, Photobiology/index.html (website accessed 07.02.2019).
1523–1544. [34] a) P. Suryawanshi, S. Gumfekar, B. Bhanvase, S.
[18] Suppliers of complete systems for flow photochemistry: Sonawnae, M. Pimplapure, Chem. Eng. Sci., 2018, 189,
Vapourtec Ltd. (www.vapourtec.com); FutureChemistry 431–448; b) K. Mizuno, Y. Nishiyama, T. Ogaki, K. Terao,
Holding BV (www.futurechemistry.com); Creaflow H. Ikeda, K. Kakiuchi, J. Photochem. Photobiol. C, 2016, 29,
(www.creaflow.be); Peschl Ultraviolet GmbH (www.peschl- 107–147.
ultraviolet.com); Corning Inc. (www.corning.com). [35] E. DeLaney, D. Lee, L. Elliott, J. Jin, K. Booker-Milburn, M.
[19] H. Bonfield, K. Mercer, A. Diaz-Rodriguez, G. Cook, B. Poliakoff, M. George, Green Chem., 2017, 19, 1431–1438.
McKay, P. Slade, G. Taylor, W. Ooi, J. Williams, J. Roberts, [36] Space time yield (STY) is determined by [(substrate
J. Murphy, L. Schmermund, W. Kroutil, T. Mielke, J. concentration × flow rate × yield) / irradiation volume] and

Accepted Manuscript
Cartwright, G. Grogan, L. Edwards, ChemPhotoChem, 2019, can be used for the description of the productivity in
DOI: 10.1002/cptc.201900203. dependence to the reactor volume. In conjunction with
[20] a) S. Ebenesajjad, P. Khaladkar: Fluoropolymers microreactors this figure of merit also describes the
applications in chemical processing industries. William effectiveness of miniaturization compared to conventional
Andrew Inc., Norwich, 2005; b) S. Protti, D. Ravelli, M. batch reactors.
Fagnoni, Design consideration of continuous flow [37] D. Lee, Z. Amara, C. Clark, Z. Xu, B. Kakimpa, H. Morvan,
photoreactors. in: Photochemical processes in continuous S. Pickering, M. Poliakoff, M. George, Org. Process Res.
flow reactors. (Ed.: Timothy Noël), World Scientific, Dev., 2017, 21, 1042–1050.
Singapore, 2017. [38] Literature on previous versions of the vortex reactor can be
[21] R. Telmesani, S. Park, T. Lynch-Colameta, A. Beeler, Angew. found in references [52-60] in ref. [37].
Chem. 2015, 127, 11683–11687, Angew. Chem. Int. Ed., [39] G. Ioannou, T. Montagnon, D. Kalaitzakis, Sp. Pergantis, G.
2015, 54, 11521–11525. Vaissilikogiannakis, ChemPhotoChem, 2017, 1, 173–177.
[22] R. Telmesani, J. White, A. Beeler, ChemPhotoChem, 2018, [40] http://www.geicp.com/cgi-
2, 865–869. bin/site/wrapper.pl?c1=Products_nebs_bytype_seaspray
[23] This approach was already described for Paternò–Büchi (website accessed 08.04.2019).
type reaction by the groups of Nishiyama and Kakiuchi: a) K. [41] I. Penders, Z. Amara, R. Horvath, K. Rossen, M. Poliakoff,
Terao, Y. Nishiyama, K. Kakiuchi, J. Flow Chem., 2014, 4, M. George, RSC Adv., 2015, 5, 6501–6504.
35–39 ; b) M. Nakano, Y. Nishiyama, H. Tanimoto, T. [42] C. Clark, D. Lee, S. Pickering, M. Poliakoff, M. George, Org.
Morimoto, K. Kakiuchi, Org. Process Res. Dev., 2016, 20, Process Res. Dev., 2016, 20, 1792–1798.
1626–1632. [43] a) A. M. Hermann, Sol. Energy 1982, 29, 323–329; b) M.
[24] B. Pieper, M. Shalom, M. Antonietti, P. Seeberger, K. Debije, P. Verbunt, Adv. Energy Mater. 2012, 2, 12–32.
Gilmore, Angew. Chem. 2018, 130, 10127–10131, Angew. [44] A. Bozdemir, S. Erbas-Cakmak, O. Ekiz, A. Dana, E.
Chem. Int. Ed., 2018, 57, 9976–9979. Akkaya, Angew. Chem. 2011, 123, 11099–11104,
[25] L. Elliott, J. Knowles, P. Koovits, K. Maskill, M. Ralph, G. Angew.Chem. Int. Ed. 2011, 50, 10907–10912.
Lejeune, L. Edwards, R. Robinson, I. Clemens, B. Cox, D. [45] M. Zettl, O. Mayer, E. Klampaftis, B. Richards, Energy
Pascoe, G. Koch, M. Eberle, M. Berry, K. Booker-Milburn, Technol. 2017, 5, 1037–1044.
Chem. Eur. J., 2014, 20, 15226–15232. [46] M. Debije, Ad. Funct. Mater. 2012, 20, 1498–1502.
[26] T. H. Rehm, C. Hofmann, D. Reinhard, H.-J. Kost, P. Löb, M. [47] D. Cambié, F. Zhao, V. Hessel, M. Debije, T. Noël, Angew.
Besold, K. Welzel, J. Barten, A. Didenko, D. Sevenard, B. Chem. 2016, 129, 1070–1074, Angew. Chem. Int. Ed. 2017,
Lix, A. Hillson, S. Riegel, React. Chem. Eng., 2017, 2, 315– 56, 1050–1054.
323. [48] D. Cambie, F. Zhao, V. Hessel, M. Debije, T. Noël, React.
[27] M. Schmidt, A. Cubillas, N. Taccardi, T. Euser, T. Cremer, F. Chem. Eng. 2017, 2, 561–566.
Maier, H.-P. Steinbrück, P. Russell, P. Wasserscheid, B. [49] a) D. Farrell, PhD Thesis, Department of Physics, Imperial
Etzold, ChemCatChem, 2013, 5, 641–650. College London, 2009; PvTrace on GitHub.com:
[28] A. Cubillas, S. Unterkofler, T. Euser, B. Etzold, A. Jones, P. https://github.com/danieljfarrell/pvtrace (website accessed
Sadler, P. Wasserscheid, P. Russell, Chem. Soc. Rev., 2013, 21.02.2019).
42, 8629–8649. [50] F. Zhao, D. Cambié, J. Janse, E. Wieland, K. Kuijpers, V.
[29] J. Knight, Nature, 2003, 424, 847–851. Hessel, M. Debije, T. Noël, ACS Sustainable Chem. Eng.
[30] Y. Han, S. Tan, M. Oo, D. Pristinski, S. Skhishvili, H. Du, 2018, 6, 422–429.
Adv. Mater., 2010, 22, 2647–2651. [51] F. Zhao, D. Cambié, V. Hessel, M. Debije, T. Noël, Green
[31] S. Ponce, H. Christians, A. Drochner, B. Etzold, Chem. Ing. Chem. 2018, 20, 2459–2464.
Tech., 2018, 11, 1855–1863. [52] D. Cambié, J. Dobbelaar, P. Riente, J. Vanderspikken, C.
[32] TeflonTM Products LINK (website accessed 18.11.2019). Shen, P. Seeberger, K. Gilmore, M. Debije, T. Noël, Angew.

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW
Chem. 2019, 131, 14512–14516, Angew. Chem. Int. Ed., [69] International Call “Solar-Driven Chemistry 2019/2020” of the
2019, 58, 14374–14378. German Research Foundation,
[53] L. Zhang, Z. Thu, B. Liu, C. Li, Y. Yu, S. Tao, T. Li, Adv. Sci., http://www.dfg.de/en/research_funding/announcements_pr
2019, 6, 1900583. oposals/2018/info_wissenschaft_18_94/index.html
[54] C. Wille, PhD thesis, TU Clausthal, 2000. (website accessed 04.03.2019).
[55] M. Al-Rawashdeh, A. Cantu-Perez, D. Ziegenbalg, P. Löb, [70] a) National Funding Programs of the German Federal
A. Gavriilidis, V. Hessel, F. Schönfeld, Chem. Eng. J. 2012, Ministry of Education and Research: a) “CO2Plus - Stoffliche
179, 318–329. Nutzung von CO2 zur Verbreiterung der Rohstoffbasis”,
[56] A. Griesbeck, N. Maptune, S. Bondock, M. Oelgemöller, http://www.chemieundco2.de/en/ (website accessed
Photochem. Photobiol. Sci. 2003, 2, 450–451. 04.03.2019); b) “CO2WIN - CO2 als nachhaltige
[57] a) H. Ehrlich, D. Linke, K. Morgenschweiss, M. Baerns, K. Kohlenstoffquelle – Wege zur industriellen Nutzung”,
Jähnisch, Chimia 2002, 56, 647–653; b) K. Jähnisch, U. https://www.fona.de/en/fundingmeasures (website
Dingerdissen, Chem. Eng. Technol. 2005, 63, 426–427; c) accessed 04.03.2019); b) National Funding Program of the

Accepted Manuscript
O. Shvydkiv, C. Limburg, K. Nolan, M. Oelgemöller, J. Flow German Federal Ministry of Economic Affairs and Energy:
Chem. 2012, 2, 52–55; d) O. Shvydkiv, K. Jähnisch, N. 7th Energy Research Programme of the Federal
Steinfeldt, A. Yavorskyy, M. Oelgemöller, Catal. Today 2018, Government,
308, 102–118. https://www.bmwi.de/Redaktion/EN/Publikationen/Energie/
[58] a) W.-K. Jo, R. Tayade, Ind. Eng. Chem. Res. 2014, 53, 7th-energy-research-programme-of-the-federal-
2073–2084; b) M. Sener, D. Ziegenbalg, Chem. Ing. Tech. government.html;
2017, 89, 1159–1173. https://www.ptj.de/projektfoerderung/angewandte-
[59] T. H. Rehm, S. Gros, P. Löb, A. Renken, React. Chem. Eng. energieforschung; (websites accessed 11.10.2019).
2016, 1, 636–648. [71] EU Horizon2020 Calls: a) “Photocatalytic synthesis (CE-
[60] D. Fabry, Y. Ho, R. Zapf, W. Tremel, M. Panthöfer, M. NMBP-25-2019)” LINK; b) “Converting Sunlight to storable
Rueping, T. H. Rehm, Green Chem. 2017, 19, 1911–1918. chemical energy (LC-SC3-RES-29-2019)” LINK; “Solar
[61] Y. Su, K. Kuijpers, V. Hessel, T. Noël, React. Chem. Eng. Energy in Industrial Processes (LC-SC3-RES-7-2019)”
2016, 1, 73–81. LINK; “International Cooperation with USA and/or China on
[62] For details, see references [44,45] in ref. [61]. alternative renewable fuels from sunlight for energy,
[63] R. Gonçalves, B. Rabello, G. César, P. Pereira, M. Ribeiro, transport and chemical storage (LC-SC3-RES-3-2020)”
E. Meurer, N. Hioka, C. Nakamura, M. Bruschi, W. Caetano, LINK; “Developing the next generation of renewable energy
Org. Process. Res. Dev. 2017, 21, 2025–2031. technologies (LC-SC3-RES-1-2019-2020)” LINK; (websites
[64] Hypericin from hypericum perforatum: accessed 11.10.2019).
https://www.sigmaaldrich.com/catalog/product/sigma/5669 [72] a) Coordinating & Support Actions to Large-Scale Research
0?lang=de&region=DE (website accessed 19.02.2019). Initiatives of the European Union: a) SunRise,
[65] L. Elliott, M. Berry, B. Haji, D. Klauber, J. Leonard, K. https://www.sunriseaction.com; b) Energy-X,
Booker-Milburn, Org. Process Res. Dev. 2016, 20, 1806– https://www.energy-x.eu; b) Mission Innovation, Panel IC5:
1811. Converting Sunlight, http://mission-innovation.net/our-
[66] International patent application: Photochemical Reactor. work/innovation-challenges/converting-sunlight/; (websites
WO2018011550A1 accessed 11.10.2019).
[67] C. Clark, D. Lee, S. Pickering, M. Poliakoff, M. George, Org.
Process Res. Dev. 2018, 22, 595–599.
[68] T. Noël, Chemistry World 2019,
https://www.chemistryworld.com/opinion/flow-into-the-
chemistry-curriculum/4010382.article (website accessed
20.10.2019).

This article is protected by copyright. All rights reserved.


ChemPhotoChem 10.1002/cptc.201900247

REVIEW

REVIEW
Photon2Flow: In the last decade Thomas H. Rehm*
photochemistry experienced a
renaissance with fundamental impact Page No. – Page No.
on academic research. In conjunction,
Reactor technology concepts for flow
the growing acceptance of continuous
photochemistry
flow technology in the chemical
community resulted in a most fruitful
alliance up to date and revolutionized
“thinking and doing chemistry” with

Accepted Manuscript
light as delicate energy source.

This article is protected by copyright. All rights reserved.

You might also like