Mass spectrometric directed system for the continuous-flow synthesis and purification of diphenhydramine

Article abstract of DOI:10.1039/c7sc00905d

A highly integrated approach to the development of a process for the continuous synthesis and purification of diphenhydramine is reported. Mass spectrometry (MS) is utilized throughout the system for on-line reaction monitoring, off-line yield quantitation, and as a reaction screening module that exploits reaction acceleration in charged microdroplets for high throughput route screening. This effort has enabled the discovery and optimization of multiple routes to diphenhydramine in glass microreactors using MS as a process analytical tool (PAT). The ability to rapidly screen conditions in charged microdroplets was used to guide optimization of the process in a microfluidic reactor. A quantitative MS method was developed and used to measure the reaction kinetics. Integration of the continuous-flow reactor/on-line MS methodology with a miniaturized crystallization platform for continuous reaction monitoring and controlled crystallization of diphenhydramine was also achieved. Our findings suggest a robust approach for the continuous manufacture of pharmaceutical drug products, exemplified in the particular case of diphenhydramine, and optimized for efficiency and crystal size, and guided by real-Time analytics to produce the agent in a form that is readily adapted to continuous synthesis.

Full text of DOI:10.1039/c7sc00905d

View Article Online
View Journal
Chemical
Science
Accepted Manuscript
This article can be cited before page numbers have been issued, to do this please use: B. P. Loren, M.
Wleklinski, A. Koswara, K. Yamminea, Y. Hu, Z. K. Nagy, D. H. Thompson and R. G. Cooks, Chem. Sci.,
2017, DOI: 10.1039/C7SC00905D.
This is an Accepted Manuscript, which has been through the
Volume
7 Number 1 January 2016 Pages 1–812
Royal Society of Chemistry peer review process and has been
accepted for publication.
Chemical
Science
Accepted Manuscripts are published online shortly after
www.rsc.org/chemicalscience
acceptance, before technical editing, formatting and proof reading.
Using this free service, authors can make their results available
to the community, in citable form, before we publish the edited
article. We will replace this Accepted Manuscript with the edited
and formatted Advance Article as soon as it is available.
You can find more information about Accepted Manuscripts in the
author guidelines.
Please note that technical editing may introduce minor changes
to the text and/or graphics, which may alter content. The journal’s
standard Terms & Conditions and the ethical guidelines, outlined
in our author and reviewer resource centre, still apply. In no
event shall the Royal Society of Chemistry be held responsible
for any errors or omissions in this Accepted Manuscript or any
consequences arising from the use of any information it contains.
ISSN 2041-6539
EDGE ARTICLE
Francesco Ricci et al.
Electronic control of DNA-based nanoswitches and nanodevices
rsc.li/chemical-science

Page 1 of 9
Chemical Science
View Article Online
DOI: 10.1039/C7SC00905D
An integrated system utilizing mass spectrometry for reaction screening, continuous-flow optimization,
and continuous crystallization of diphenhydramine has been developed.

Please do not adjust margins
Chemical Science
Page 2 of 9
View Article Online
DOI: 10.1039/C7SC00905D
Chemical Science
ARTICLE
Mass Spectrometric Directed System for the Continuous-Flow
Synthesis and Purification of Diphenhydramine
Received 00th January 20xx,
Accepted 00th January 20xx
Bradley P. Loren+a, Michael Wleklinski+a, Andy Koswara+b, Kathryn
Yamminea, Yanyang Hua, Zoltan K. Nagy*b, David H. Thompson*a,
and R. Graham Cooks*a
DOI: 10.1039/x0xx00000x
www.rsc.org/
A highly integrated approach to the development of a process for the continuous synthesis and purification of
diphenhydramine is reported. Mass spectrometry (MS) is utilized throughout the system for on-line reaction monitoring,
off-line yield quantitation, and as a reaction screening module that exploits reaction acceleration in charged microdroplets
for high throughput route screening. This effort has enabled the discovery and optimization of multiple routes to
diphenhydramine in glass microreactors using MS as a process analytical tool (PAT). The ability to rapidly screen conditions
in charged microdroplets was used to guide optimization of the process in a microfluidic reactor. A quantitative MS method
was developed and used to measure the reaction kinetics. Integration of the continuous flow reactor/on-line MS
methodology with a miniaturized crystallization platform for continuous reaction monitoring and controlled crystallization
of diphenhydramine was also achieved. Our findings suggest a robust approach for the continuous manufacture of
phamacutical drug products, exemplified in the particular case of diphenhydramine, and optimized for efficiency and crystal
size, and guided by real-time analytics to produce the agent in a form that is readily adapted to continuous synthesis.
10-14
in microreactors.5,
In addition to the increasing diversity of
Introduction
chemistry reported for traditional solution-phase reactions in
microreactors, this format also offers the flexibility for enabling
There is an extremely rich and rapidly growing knowledge base in the
literature for organic transformations; however, the vast majority of
these are based on batch processes. Continuous-flow reactor
technology has generated significant interest due to the variety of
advantages it offers over traditional batch processes such as much
lower process costs and increased process efficiency.1 These
advantages arise from the rapid mixing, superior heat transfer, safe
operation at high temperatures and pressures, and control and
automation that flow reactions can offer.2-4 These features provide
the potential to execute organic reactions and purifications at rates
and in yields not achievable in a batch system.5 Additionally, there is
an inherent scalability with continuous flow processes compared to
batch processes, whereby process quality does not diminish as
process scale increases.6, 7
technologies to be developed to cover other types of reactions, such
16
as gas-liquid reactions15,
,
photochemical reactions17, and
electrochemical reactions18, followed by different types of flow
purifications, including extraction19, crystallization20 and
chromatography.21
All of the technologies described above have the potential for
rapid optimization by means of on-line reaction monitoring.22 Real-
time analysis provides the opportunity for rapid and automated
process optimization.23-25 We are working to develop a continuous
synthesis system that is driven by on-line analytics. Techniques that
have previously been used to monitor microfluidics processes
include UV-Vis26, Raman27, IR28, 29, NMR25, and MS.30-32 The potential
for rapid analysis times, as well as the detailed product and
byproduct information that can be obtained from MS make it
particularly attractive for the near real-time monitoring of reaction
mixtures. Mass spectrometry has extensively been used to monitor
both batch and continuous reactions, including, for example,
Sonogashira and Negishi couplings.33-36 We report here the novel use
These advantages have led to a growing interest in the recent
literature in microfluidic technologies, as is exemplified by review
articles2-4, 8, 9, and the expanding scope of chemistry being employed
of N2(g) segmented droplets and an inductive electrospray ionization
to provide MS feedback in under 30 seconds with
minimal waste production.
a.Department of Chemistry, Purdue University, West Lafayette, IN 47907, USA. E-
Mail: davethom@purdue.edu; cooks@purdue.edu
37
method35,
b.Department of Chemical Engineering, Purdue University, West Lafayette, IN
47907 E-Mail: zknagy@purdue.edu
+ These authors contributed equally
Another significant aspect of this work is the utilization of MS as
a reaction screening module by accelerating reactions in charged
Electronic Supplementary Information (ESI) available NMR spectra of selected
product, mass spectra of selected products, crystallization information, and
experimental procedures are supplied. See DOI: 10.1039/x0xx00000x
This journal is © The Royal Society of Chemistry 20xx
Chem. Sci., 2013, 00, 1-3 | 1
Please do not adjust margins

Please do not adjust margins
Page 3 of 9
Chemical Science
ARTICLE
Journal Name
View Article Online
DOI: 10.1039/C7SC00905D
Scheme 1 Synthesis of Diphenhydramine from
chloro/bromo diphneylmethane
segmented flow crystallizer to continuously produce solid products
without clogging.
Results
Continuous-Flow Synthesis of DPH
Microfluidic reactors have the potential to rapidly mix materials by
diffusion, and the addition of micromixers can enhance the process
by convection.43 A 19.5 L Chemtrix reactor chip equipped with
staggered oriented ridge (SOR) mixing sections was used for this. The
SOR sections of the chip create turbulent flow immediately upon
mixing two reagents to provide rapid and efficient mixing. The outlet
of the reactor chip is fed into a pressure sensor and back pressure
regulator (BPR) that was used to run reactions at pressures ranging
from 10-20 bar. The first routes explored with this reactor system
used either bromo- or chlorodiphenylmethane reacting with
dimethylaminoethanol (DMAE) in ACN, a route previously reported
in flow (Scheme 1).10 Reaction screening using accelerated reactions
in microdroplets was attempted but proved unsuccessful due to the
Figure 1 Representation of (A) accelerated droplet reactor system,
and (B) microfluidic reactor system. The droplet reactor is used to
rapidly screen routes, whilst the microfluidic reactor is used to
optimize favourable routes. Two explored routes are illustrated.
39
microdroplets for high throughput reaction screening.38,
The
acceleration factors vary from 103-106 allowing accelerated access to
slow reactions; reaction acceleration is due to increases in reagent
concentration as solvent evaporates, pH effects, and interfacial
effects.38, 40-42 It is important to note that MS under the appropriate
conditions can be used to either monitor or accelerate reactions. In
this paper, both types of experiments are performed to help in the
optimization of the synthesis of diphenhydramine. While
microfluidics provides the opportunity to screen conditions more
quickly than batch processes, the use of the droplet reactor as a
screening tool prior to microfluidic optimization has the potential to
significantly reduce the time required for route selection and,
subsequently, reaction optimization. The droplet and microreactor
systems are represented in Figure 1. The continuous-flow reactors
used for this system are glass microreactor chips from Chemtrix.®
formation
of
a
gas
phase
hetero-dimer
between
dimethylaminoethanol and the dibenzylic carbocation which has the
same nominal mass as diphenhydramine. This represents a case in
which accelerated reaction screening failed; however, due to the
immense number of synthetic routes it is still a powerful technique
as will be shown later in this paper.
A 2.0 M solution of the halo-diphenylmethane was mixed
with neat DMAE in the first SOR section of the chip and acetonitrile
dilution was performed prior to the second SOR section as a reaction
quench. We first explored the temperature dependence of this
reaction by screening temperatures from 60-200 C at a residence
time of 5.0 min. A strong dependence on temperature was observed
with no significant product being formed below 150 C for both
bromo- and chlorodiphenylmethane. While conducting reactions at
temperatures higher than 200 C would be advantageous to
potentially optimize the process further, this was not possible due to
technical limitation of the commercial system.
In addition to conducting organic transformations in
a
continuous fashion, the potential advantages for executing
continuous purifications are similarly important. This is particularly
true for the production of active pharmaceutical ingredients (APIs),
as exemplified by the recent on-demand continuous synthesis
system developed at MIT.11 Liquid-liquid extraction19 and
crystallization20 have both been demonstrated using continuous-
flow technologies. We show here the integration of a miniaturized
pharmaceutical purification platform (MiniPharm) with the
microreactor and on-line inductive ESI-MS monitoring. This system
consists of a mass-flow controller and temperature bath, equipped
with in-line analytics for droplet and crystal detection, with control
and process automation provided by a LabView based GUI.
1
These results are summarized in Table 1 and Figure 2 with HNMR
found in Supporting Information (Figures S1-3).
Diphenhydramine (DPH) is an antihistamine found in several
With the temperature data in hand, we then screened residence
times (RT) at 150 and 200 C, first for bromodiphenylmethane, using
off-line MS qualitatively as a means to monitor product formation,
followed by both MS and 1H NMR analysis to calculate a relative
conversion. The same trends were observed with both techniques.
common medicines, including Benadryl.®
A continuous-flow
synthesis of DPH was published in 2013 by the Jamison group.10 We
now demonstrate the use of both on-line and off-line MS to optimize
the synthesis of diphenhydramine employing multiple routes using a
Chemtrix Labtrix S1 microfluidic device, along with the rapid
screening of reactions accelerated in microdroplets to guide reaction
routes examined further in continuous-flow. The route screening in
microdroplets expands upon earlier work done by Wleklisnki et al.39
We also show the integration of one of the synthetic routes with a
MiniPharm continuous purification system, which implements an air-
2 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Chemical Science
Page 4 of 9
Chemical Science
Table 1. Effect of residence time (RT) and temperature on synthesis
of diphenhydramine from bromo- and chlorodiphenylmethane
View Article Online
DOI: 10.1039/C7SC00905D
Entry
X
RT (min)
3:(1+3)a
3:(1+3)b
Temp. (C)
1
2
3
4
5
6
7
8
Br
Br
Br
Br
Cl
Cl
Cl
Cl
Br
Br
Br
Br
60
5
5
5
5
5
5
5
5
7.5
10
7.5
10
3.70
0.65
20.3
60.6
6.9
42.2
50.2
77.8
96.9
66.3
78.0
72.5
97.3
81.9
92.4
95.0
95.3
100
150
200
60
100
150
200
150
150
200
200
2.6
41.3
90.1
49
91
95
9
10
11
12
98
a3:(1+3) calculated using 1H NMR. b3:(1+3) calculated using MS
Figure 3 Byproduct formation (A) MS/MS of m/z 272 shows
presence of isomeric ions; (B) MS follows the effect of
residence time on byproduct formation
Diphenhydramine is commonly administered orally as
a
hydrochloride salt. We therefore pursued chlorodiphenylmethane as
an alternative starting material for this process to allow us to isolate
diphenhydramine directly as the hydrochloride salt. To accomplish
this optimization rapidly with a minimal amount of material, we
developed an off-line quantitative MS method using serial dilution
and diphenhydramine-D3 for monitoring product yield (Figure S4).
The temperature screen was repeated and a more extensive
residence time screen was conducted. The quantitative MS data
confirmed the temperature dependence previously observed, and
of the reagents. This information is summarized in Figure 3 and
Figure S5.
On-line MS Reaction Monitoring
The use of off-line and slow analytical methods are common
bottlenecks in flow reaction optimization. Therefore, the
development of an on-line reaction monitoring system using MS
represents a significant advance in the capability to rapidly optimize
reaction conditions. We demonstrate here the ability to monitor and
revealed
a strong RT dependence for the formation of
optimize
the
synthesis
of
diphenhydramine
from
diphenhydramine, with the yield increasing with RT up to 60 seconds,
followed by a leveling off at residence times up to 600 seconds
(Figure 2). The highest yield achieved was 0.5 M, which corresponds
to a production rate of 128 mg of DPH per hour. Assuming second
order kinetics, this is roughly eight times faster than the previously
published continuous flow synthesis.10
chlorodiphenylmethane and dimethylaminoethanol using on-line
MS. The setup for these experiments uses the 19.5 L SOR chip with
the outlet connected to a 2 L injection valve that allowed the
product to be collected, sent to waste, or sent to the MS. As the
material emerging from the reactor was approximately 1.0 M, it was
necessary to include an acetonitrile in-line dilution step after the
injector valve. In order to enhance the rate of data acquisition and
reduce the amount of material wasted when collecting MS data,
The detailed structural information that MS and MS/MS can
rapidly provide make MS a powerful tool for reaction monitoring and
process optimization. The formation of multiple N-oxide byproducts
is observed in the mass spectra and their structures have been
confirmed with MS/MS (Figure 3 and S5).44 Furthermore, the
formation of these byproducts is dependent on the residence time
Figure 4 (A) Sample mass spectra collected on-line. (B)
Calculated conversions from MS data collected on-line
demonstrate temperature dependence of the reaction.
S.M. refers to DMAE.
Figure 2 MS quantitation of DPH synthesis from
chlorodiphenylmethane.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 3
Please do not adjust margins

Please do not adjust margins
Page 5 of 9
Chemical Science
ARTICLE
Journal Name
View Article Online
DOI: 10.1039/C7SC00905D
Figure 5 Engineering details for the integrated synthesis and purification system. Orange lines represent a tubular heater (Fig. S8).
N2(g) segmented droplets were created using an electronic pressure chlorodiphenylmethane, we have utilized an air-segmented flow
regulator (EPR) to rapidly deliver droplets from the reactor to the MS. crystallizer in addition to the on-line MS capabilities described above.
This provided L sized droplets that were approximately 0.01M in Crystallization in segmented droplets allows for the continuous
reaction mixture. These droplets traveled a distance of roughly 3 m, crystallization of material at a controlled crystal size without
representing 150 L of tubing dead volume to the mass fouling.20, 45 A full engineering diagram of the experimental setup is
spectrometer. The use of air-segmentation allowed data collection detailed in Figure 5.
within 30 seconds with minimal waste generation.
A video
Again using a 19.5 L reactor chip, the outlet was fed into a valve
capable of splitting the flow between collection, waste, MS, and the
air-segmented flow crystallizer. As shown in Figure 5, the MiniPharm
consists of PFA tubing connected in a T-junction to a syringe pump
filled with EtOAc, the outlet of the microreactor, and a 0.3 m length
of PFA tubing where mixing takes place by diffusion. The EtOAc
served as the anti-solvent prior to droplet formation. At this point,
there is a high potential for crystal formation that could block the
tubing shortly after anti-solvent mixing (Figure S6, S7). We have,
therefore, implemented cut-to-length tubing heaters before and
after the T-junction developed in-house to keep the material
solubilized and prevent this from occurring (Figure S8). This tubing
section is followed by another T-junction, which is connected to the
mass-flow controller and a 1.2 m length of PFA tubing where the
crystallization takes place. The mass-flow controller is used to
modulate the precise flow rate of N2(g) and in turn create a uniform
distribution of N2(g) segmented droplets. Currently, it takes three
residence times for the droplets to reach a uniform distribution with
low variance. (Figure S9). This can be improved upon by using a BPR
microscope along with an optical sensor was used in-line in order to
monitor the droplets. The MS analysis was conducted using an
inductive ionization source (iESI-MS).35, 37 These data are summarized
in Figure 4. At a residence time of 5.0 minutes, we demonstrated the
same temperature dependence that we observed using off-line MS.
No significant product is detected at lower temperatures, and an
increase in product formation is seen at 150 C, followed by a
significant increase at 200 C. The peak at m/z 550 is a gas phase
dimer of the DPH product ([2 DPH + 2H + Cl]+).
Continuous-Flow Crystallization of DPH
In order to develop a completely continuous and automated process,
we designed a system capable of seamlessly transitioning between
synthesis, analytics, and ultimately purification of the final product.
The purification system has been designed as a flexible miniaturized
pharmaceutical platform, so that the appropriate purification
module, such as extraction and/or crystallization, can be
implemented easily and optimized depending on the synthesis being
conducted. For the synthesis of diphenhydramine from
4 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Chemical Science
Page 6 of 9
Chemical Science
View Article Online
DOI: 10.1039/C7SC00905D
Scheme 2 Synthesis of DPH from benzhydrol.
A
major component of our analytic-directed continuous
synthesis system is a reaction screening module which takes
advantage of reaction acceleration in charged microdroplets.38
A
representation of this module is shown in Figure 1A. When starting
the synthesis of a new target molecule, there is a vast number of
options not only for routes, but also for solvents and conditions to
conduct each organic transformation. The droplet methodology
provides for rapid screening of reaction pathways and solvents to
guide the subsequent continuous-flow optimization. Droplet
reactions can be conducted either by spraying directly into the MS
(on-line) or by spraying onto a substrate such as glass wool followed
by dissolution and off-line analysis. The off-line method is beneficial
as it allows one to collect more material for analysis, provides the
opportunity to telescope multiple synthetic steps, and avoids solvent
effects during ionization by dissolution into a common solvent for
analysis. We employed this approach to rapidly screen solvents for
the synthesis of diphenhydramine via the mesylation or tosylation of
benzhydrol to form 5, followed by treatment with
dimethylaminoethanol to make DPH (3). Information obtained from
these experiments was then compared to the data obtained from
continuous-flow reactions.
Figure 6 (A) Droplet train from microreactor. (B) Droplet
detection using LED phototransistor, (C) droplet size distribution.
at the crystallizer’s outlet in the future. The controlled and
accelerated formation of N2(g) segmented crystals was subsequently
induced by immersing a section of the tubing into a dry ice bath. The
temperature was monitored and maintained at -20 C at the
interface between the dry ice bath and the tubing. Here, uniform
crystals are formed while fouling is prevented due to precise
supersaturation within each segmented droplet.
Prior to telescoping the two reaction steps for this synthesis, the
mesylation of benzhydrol was first performed in both the droplet
reactor and in a single microfluidic chip. When benzhydrol and
methanesulfonyl chloride were sprayed for 5 minutes in the off-line
droplet reactor, the dibenzylic carbocation (m/z 167) was observed
in positive mode and the two-electron reduction product of the
mesylate (m/z 263) was observed in negative mode. In a 5 L
microfluidic reactor at 200 C, MS analysis showed that significantly
more product was formed at a residence time of 30 seconds than at
a residence time of 60 seconds (Figures S12 and S13). This was likely
due to degradation of the mesylated benzhydrol on the chip after it
was formed. NMR and MS evidence suggested that the mesylate
degraded under these conditions over the course of a day.
Multiple analytical techniques were used as PAT to monitor this
process, and a LabView based GUI was developed for process
integration and automation. In addition to on-line MS, off-line MS
based quantitation was also utilized during crystallization
experiments in order to monitor product yield and ensure effective
crystallization. A phototransistor was also placed in-line to monitor
droplet formation. The use of the mass flow controller resulted in
uniform droplet distribution that allowed us to control crystal size
and prevent fouling of the tubing (Figure 6C, S9). The phototransistor
was also effective, as shown in Figure 6B, in detecting crystal
formation, and a video microscope was incorporated in-line for the
air-segmented flow crystallizer to visualize crystal formation (Figure
6A). The output was collected by filtration and analyzed by nESI-MS,
and 1HNMR (Figures S10, S11).
Telescoping this two-step synthesis in the droplet reactor was
accomplished by spraying benzhydrol and MsCl/TsCl onto glass wool
for 5 minutes, re-dissolving the collected material and then spraying
that solution again with DMAE. These reactions were sprayed in ACN,
DMF, and toluene while the analysis was conducted in ACN for all
cases. For TsCl, the droplet reactor indicated product formation for
all solvents tested with the reaction in toluene having the highest
conversion (Figure S14). However, fouling issues in the microfluidic
reactor caused us to abandon this route. Droplet reactions occur in
an open environment, therefore potential information about
component insolubility is not obtained. However, the large number
of synthetic routes available suggests that a subset of positive
droplet reactions will meet the requirement of providing a successful
flow synthesis.
Droplet and Flow Synthesis of DPH from Benzhydrol
In an effort to develop a process utilizing commodity starting
materials, we sought to replace the halo-diphenylmethane with the
significantly cheaper diphenylmethanol (benzhydrol). Using
benzhydrol as
a
starting material for the synthesis of
diphenhydramine has been explored previously in continuous-flow
through a p-toluene sulfonic acid catalyzed condensation reaction
with dimethylaminoethanol.10 We pursued a different approach and
created a better leaving group by reacting benzhydrol with
methanesulfonyl chloride (MsCl) or p-toluenesulfonyl chloride (TsCl)
(Scheme 2).
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 5
Please do not adjust margins

Please do not adjust margins
Page 7 of 9
Chemical Science
ARTICLE
Journal Name
Table 2. Effect of residence time, temperature, stoichiometry, and
solvent on the synthesis of DPH from benzhydrol
The results obtained from the solvent screen that wasVdieownAertwicliethOntlhinee
droplet reactor are consistent with the findings from the microfluidic
DOI: 10.1039/C7SC00905D
reactor. The droplet reactor indicated product formation was
possible with all solvents, but ACN is significantly better than DMF
and toluene. Due to solubility concerns, DMF was initially screened
for this system as it was the most effective at preventing fouling of
the reactor chip. A large number of conditions were, therefore,
explored in DMF, but the presence of byproducts that resulted from
solvent degradation prevented the synthesis of DPH in a yield over
9%. Fouling issues in toluene prevented the screening of many
different conditions, and those conditions where solubility was
retained all had yields of less than 1 %. ACN also had significant issues
with precipitation and clogging of the reactor that prevented a large
screen of conditions. However, at 150 and 200 C and RT of 5.0 and
0.66 minutes in the first and second chips, respectively, the highest
yielding condition of 34.5 % was achieved (Table 2, Entry 4).
Entry
Temp. (°C)
(1/2)
RT (min)
Stoich.
(2:4)
Solvent
Conv.
(%)
Yield
1
2
3
4
5
6
7
8
9
150/200
150/200
150/200
150/200
150/175
150/175
150/175
150/175
150/200
150/200
150/200
150/200
0.50/0.067
1.0/0.13
2.5/0.33
5.0/0.66
2.5/0.33
5.0/0.67
2.5/0.25
5.0/0.5
1.0/0.13
1.0/0.10
1.0/0.08
1.0/0.67
1:1
1:1
1:1
1:1
1:1
1:1
1:2
1:2
1:1
2:1
3:1
4:1
ACN
ACN
ACN
ACN
DMF
DMF
DMF
DMF
Toluene
Toluene
Toluene
Toluene
32.1
20.2
54.9
55.2
47.5
41.5
43.0
49.6
12.0
17.1
6.85
23.1
-
4.85
5.63
34.5
-
5.93
6.79
8.73
-
10
11
12
0.15
0.14
0.29
aQualitative calculation of conversion from MS. bQuantitative calculation of yield
from MS.
With this information in hand, we then sought to optimize the
telescoped two-step synthesis of diphenhydramine from benzhydrol
in a microfluidic reactor. This was first attempted using a single 15 L
reactor chip that had a third injection port at 5 L, effectively
providing a 5 L reactor for the first step and a 10 L reactor for the
second step. A large number of conditions were attempted using this
system, including residence time screens at 150 and 200 C. All of
these conditions failed, however, due to fouling of the chip shortly
after starting the reaction. Triethylamine (TEA) was then
incorporated into the reaction mixture in an attempt to improve the
solubility of the process. While the addition of TEA maintained
solubility throughout the course of the reaction, the conversion to
DPH remained low for all conditions attempted (Figure S15).
Conclusions
Diphenhydramine was synthesized via multiple routes using
both a microfluidic system and an accelerated droplet reactor.
The synthesis from the reaction of bromo- and
chlorodiphenylmethane with dimethylaminoethanol was first
optimized in the microfluidic system by screening temperatures
and residence times. ESI-MS was used as a process analytical
tool (PAT) both qualitatively and quantitatively to facilitate the
optimization process. These studies led to an optimal condition
from chlorodiphenylmethane in acetonitrile at 200 C with a
In an effort to obtain conditions that were compatible with
continuous-flow, five different bases were screened in small batches
for solubility and product formation using both methanesulfonyl
chloride (MsCl) and p-toluenesulfonyl chloride (TsCl). These
experiments were conducted by both adding the base in the first step
of the synthesis as well as during the second step to explore different
possible configurations for the microfluidic setup. While 2,6-lutidine,
DMAP, and DABCO all exhibited conditions where the reactants
appeared to remain soluble while also forming product, fouling of
the reactor chip remained a major issue that prevented any of these
conditions from being optimized in continuous-flow.
residence time of only 1 minute to produce DPH at a rate of 128
mg/hr. The ability to monitor this system in real time using ESI-
MS was also demonstrated using N2(g) segmented droplets and
an inductive ionization. These tools were then integrated with a
MiniPharm platform and DPH was successfully crystallized
continuously in N2(g) segmented droplets without fouling. The
size of the droplets, and therefore the crystals, was effectively
controlled with a mass flow controller and a temperature bath.
The droplets and crystals are monitored using an in-line
phototransistor and a video microscope, which will be
integrated with an image processing algorithm for real-time
droplet size distribution calculation in the future.
The synthesis of DPH from benzhydrol using
methanesulfonyl and p-toluenesulfonyl chloride was also
accomplished. This synthesis was demonstrated in both the
microfluidic system and an accelerated droplet reactor which
established its potential to be used as a rapid reaction screening
tool. A variety of solvents and conditions were screened in both
reactor systems, using both qualitative and quantitative MS to
monitor these processes. The trends in product formation with
varying solvent were established using the droplet reactor, and
these were shown to be consistent with the findings from
microfluidic reactions. Fouling issues required the development
of a two-chip system, and MS quantitation revealed that ACN
was the best solvent for this synthesis.
The system was then reconfigured to allow this synthesis to be
conducted in two separate reactor chips. To accomplish this, a device
of multiple heating units that could be independently controlled was
designed and fabricated (Figure S16). A 5 L chip was employed for
the first step and a 1 L chip was used for the second step. The
purpose behind this design was an attempt to drive the second step
of the synthesis at high temperatures in a short residence time to
prevent fouling of the chip. The synthesis was screened in ACN, DMF,
and toluene at a variety of temperatures, residence times, and
stoichiometries. The concentrations of all reagents were kept at 0.5
M. Off-line MS was again used to monitor this synthesis qualitatively,
and a subset of conditions were examined quantitatively in the same
manner
that
was
done
for
the
synthesis
from
chlorodiphenylmethane. Qualitative conversions were calculated by
considering starting materials, intermediates, and byproducts. The
experiments that were analyzed quantitatively are summarized in
Table 2, and the full dataset is detailed in the supporting information
(Table S1).
This study describes our initial efforts towards the
development of an automated organic synthesis system
directed by mass spectrometry. The design of this system
includes using MS as not only a PAT, but also as a way to
6 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins

Chemical Science
Page 8 of 9
Chemical Science
accelerate reactions for high throughput route screening as well 16
as scalable manufacturing. Future plans include the integration
of a design of experiments approach to allow for automated
optimization of a process. The accomplishments demonstrated 17
here are significant steps towards the development of such a
system. Increasing the speed of both route screening and route 18
optimization, along with increasing the scope of chemistry for
both the accelerated reaction screening and the microfluidic 19
reaction and purification are critical next steps.
U. Gross, P. Koos, M. O'Brien, A. Polyzos and S. V. Ley,
View Article Online
Eur. J. Org. Chem., 2014, DOI: 10D.1O0I0: 210/.e10jo3c9./2C071SC40002980054D,
6418-6430.
T. Fukuyama, Y. Fujita, M. A. Rashid and I. Ryu, Org. Lett.,
2016, DOI: 10.1021/acs.orglett.6b02727.
T. Arai, H. Tateno, K. Nakabayashi, T. Kashiwagi and M.
Atobe, Chem. Commum., 2015, 51, 4891-4894.
J. G. Kralj, H. R. Sahoo and K. F. Jensen, Lab Chip, 2007,
7
, 256-263.
20
P. Neugebauer and J. G. Khinast, Cryst. Growth Des.,
2015, 15, 1089-1095.
Acknowledgements
We gratefully acknowledge the financial support of DARPA
(award no. W911NF-16-2-0020) and the technical assistance
21
J. W. Lee, Z. Horváth, A. G. O’Brien, P. H. Seeberger and
A. Seidel-Morgenstern, Chem. Eng. J., 2014, 251, 355-
370.
provided by Ryan Hilger of the Amy Facility at Purdue University,
Department of Chemistry, and Adam Hollerbach.
22
23
24
V. Sans and L. Cronin, Chem. Soc. Rev., 2016, 45, 2032-
2043.
B. J. Reizman, Y. M. Wang, S. L. Buchwald and K. F.
Jensen, React. Chem. Eng., 2016, 1, 658-666.
Notes and references
J. P. McMullen, M. T. Stone, S. L. Buchwald and K. F.
Jensen, Angew. Chem., Int. Ed. Engl., 2010, 49, 7076-
7080.
V. Sans, L. Porwol, V. Dragone and L. Cronin, Chem. Sci.,
2015, 6, 1258-1264.
1
S. D. Schaber, D. I. Gerogiorgis, R. Ramachandran, J. M.
B. Evans, P. I. Barton and B. L. Trout, Ind. Eng. Chem.
Res., 2011, 50, 10083-10092.
25
26
27
28
H. Lu, M. A. Schmidt and K. F. Jensen, Lab Chip, 2001, 1,
22-28.
2
B. Gutmann, D. Cantillo and C. O. Kappe, Angew. Chem.,
Int. Ed. Engl., 2015, 54, 6688-6728.
T. A. Hamlin and N. E. Leadbeater, Beilstein J. Org.
Chem., 2013, , 1843-1852.
3
K. F. Jensen, B. J. Reizman and S. G. Newman, Lab Chip,
2014, 14, 3206-3212.
9
C. F. Carter, H. Lange, S. V. Ley, I. R. Baxendale, B.
Wittkamp, J. G. Goode and N. L. Gaunt, Org. Process Res.
Dev., 2010, 14, 393-404.
S. Newton, C. F. Carter, C. M. Pearson, L. D. Alves, H.
Lange, P. Thansandote and S. V. Ley, Angew. Chem., Int.
Ed. Engl., 2014, 53, 4915-4920.
J. J. Haven, J. Vandenbergh and T. Junkers, Chem.
Commum., 2015, 51, 4611-4614.
J. S. Mathieson, M. H. Rosnes, V. Sans, P. J. Kitson and L.
4
K. S. Elvira, X. C. I. Solvas, R. C. R. Wootton and A. J.
deMello, Nat. Chem., 2013, 5, 905-915.
5
H. Kim, K. I. Min, K. Inoue, D. J. Im, D. P. Kim and J.
Yoshida, Science, 2016, 352, 691-694.
29
6
N. Kockmann, M. Gottsponer, B. Zimmermann and D. M.
Roberge, Chem. Eur. J., 2008, 14, 7470-7477.
Y. Zhang, S. C. Born and K. F. Jensen, Org. Process Res.
Dev., 2014, 18, 1476-1481.
30
31
32
33
34
35
36
37
38
39
7
8
J. P. McMullen and K. F. Jensen, Annu. Rev. Anal. Chem.,
Cronin, Beilstein J. Nanotechnol., 2013, 4, 285-291.
2010, 3, 19-42.
B. V. Silva, F. A. Violante, A. C. Pinto and L. S. Santos,
Rapid Commun. Mass Spectrom., 2011, 25, 423-428.
K. L. Vikse, M. P. Woods and J. S. McIndoe,
Organometallics, 2010, 29, 6615-6618.
9
P. L. Mills, D. J. Quiram and J. F. Ryley, Chem. Eng. Sci.,
2007, 62, 6992-7010.
10
11
D. R. Snead and T. F. Jamison, Chem. Sci., 2013, 4, 2822-
2827.
A. J. Ingram, C. L. Boeser and R. N. Zare, Chem. Sci., 2016,
A. Adamo, R. L. Beingessner, M. Behnam, J. Chen, T. F.
Jamison, K. F. Jensen, J. C. Monbaliu, A. S. Myerson, E.
M. Revalor, D. R. Snead, T. Stelzer, N. Weeranoppanant,
S. Y. Wong and P. Zhang, Science, 2016, 352, 61-67.
A. R. Bogdan, S. L. Poe, D. C. Kubis, S. J. Broadwater and
7
, 39-55.
X. Yan, E. Sokol, X. Li, G. Li, S. Xu and R. G. Cooks, Angew.
Chem., Int. Ed. Engl., 2014, 126, 6041-6045.
K. L. Vikse, M. A. Henderson, A. G. Oliver and J. S.
McIndoe, Chem. Commum., 2010, 46, 7412-7414.
G. Huang, G. Li and R. G. Cooks, Angew. Chem., Int. Ed.
Engl., 2011, 50, 9907-9910.
12
D. T. McQuade, Angew. Chem., Int. Ed. Engl., 2009, 48
,
8547-8550.
13
14
D. R. Snead and T. F. Jamison, Angew. Chem., Int. Ed.
Engl., 2015, 54, 983-987.
F. Lima, M. A. Kabeshov, D. N. Tran, C. Battilocchio, J.
Sedelmeier, G. Sedelmeier, B. Schenkel and S. V. Ley,
Angew. Chem., Int. Ed. Engl., 2016, 55, 14085-14089.
S. V. F. Hansen, Z. E. Wilson, T. Ulven and S. V. Ley, React.
R. M. Bain, C. J. Pulliam and R. G. Cooks, Chem. Sci.,
2015, 6, 397-401.
M. Wleklinski, C. E. Falcone, B. P. Loren, Z. Jaman, K. Iyer,
H. S. Ewan, S.-H. Hyun, D. H. Thompson and R. G. Cooks,
Eur. J. Org. Chem., 2016, 2016, 5480-5484.
X. Yan, R. M. Bain and R. G. Cooks, Angew. Chem., Int.
Ed. Engl., 2016, 55, 12960-12972.
15
40
Chem. Eng., 2016, 1, 280-287.
This journal is © The Royal Society of Chemistry 20xx
J. Name., 2013, 00, 1-3 | 7
Please do not adjust margins

Please do not adjust margins
Page 9 of 9
Chemical Science
ARTICLE
Journal Name
41
R. M. Bain, C. J. Pulliam, S. T. Ayrton, K. Bain and R. G.
Cooks, Rapid Commun. Mass Spectrom., 2016, 30, 1875-
1878.
View Article Online
DOI: 10.1039/C7SC00905D
42
43
Y. Li, X. Yan and R. G. Cooks, Angew. Chem., Int. Ed. Engl.,
2016, 55, 3433-3437.
C. P. Holvey, D. M. Roberge, M. Gottsponer, N.
Kockmann and A. Macchi, Chem. Eng. Process., 2011, 50
,
1069-1075.
44
45
A. Baldacci, F. Prost and W. Thormann, Electrophoresis,
2004, 25, 1607-1614.
J. Lu, J. D. Litster and Z. K. Nagy, Cryst. Growth Des.,
2015, 15, 3645-3651.
8 | J. Name., 2012, 00, 1-3
This journal is © The Royal Society of Chemistry 20xx
Please do not adjust margins