Online Stereochemical Process Monitoring by Molecular Rotational Resonance Spectroscopy

Article abstract of DOI:10.1021/acs.oprd.9b00089

A molecular rotational resonance (MRR) spectrometer designed to monitor the product composition of an asymmetric continuous flow reaction online is presented. The MRR technique is highly sensitive to small changes in molecular structure and, as such, is capable of rapidly quantifying isomers as well as other impurities in a complex mixture, without chromatographic separation or chemometrics. The spectrometer in this study operates by automatically drawing a portion of the reaction solution into a reservoir, volatizing it by heating, and measuring the highly resolved MRR spectra of each of the components of interest in order to determine their relative quantity in the mixture. The reaction under study was the hydrogenation of artemisinic acid, an intermediate step in the semisynthesis of the antimalarial drug artemisinin. Four analytes were characterized in each measurement: the starting material, the product, a diastereomer of the product, and an overreduction byproduct that was not directly quantifiable by either HPLC or NMR methods. The MRR instrument has a measurement cycle time of approximately 17 min for this analysis and can run for several hours without any user interaction.

Full text of DOI:10.1021/acs.oprd.9b00089

Subscriber access provided by Bibliothèque de l'Université Paris-Sud
Full Paper
On-line Stereochemical Process Monitoring by
Molecular Rotational Resonance Spectroscopy
Justin Neill, Yuan Yang, Matt Thomas Muckle, Roger L Reynolds, Luca
Evangelisti, Reilly E. Sonstrom, Brooks H. Pate, and B. Frank Gupton
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.9b00089 • Publication Date (Web): 02 Apr 2019
Downloaded from http://pubs.acs.org on April 2, 2019
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 18
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
On-line Stereochemical Process Monitoring by Molecular
Rotational Resonance Spectroscopy
1*
2
1
1
Justin L. Neill , Yuan Yang , Matt T. Muckle , Roger L. Reynolds , Luca
3
4
4
2
Evangelisti , Reilly E. Sonstrom , Brooks H. Pate , and B. Frank Gupton
1
BrightSpec, Inc., 770 Harris St. Suite 104b, Charlottesville, VA 22903
2
Department of Chemical and Life Science Engineering, Virginia Commonwealth University,
Richmond, VA
3
Dipartimento di Chimica “G. Ciamician” dell’Universita, via Selmi 2, I-40126, Bologna, Italy
Department of Chemistry, University of Virginia, McCormick Rd, Charlottesville, VA 22904
4
*
Corresponding author: Justin L. Neill: justin.neill@brightspec.com, Tel: +1 434 202 2391.
1
ACS Paragon Plus Environment

Organic Process Research & Development
Page 2 of 18
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Table of Contents Graphic
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
ACS Paragon Plus Environment

Page 3 of 18
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Abstract
A molecular rotational resonance (MRR) spectrometer designed to monitor the product
composition of an asymmetric continuous flow reaction on-line is presented. The MRR
technique is highly sensitive to small changes in molecular structure, and as such is capable of
rapidly quantifying isomers as well as other impurities in a complex mixture, without
chromatographic separation or chemometrics. The spectrometer in this study operates by
automatically drawing a portion of the reaction solution into a reservoir, volatizing it by heating,
and measuring the highly resolved MRR spectra of each of the components of interest in order to
determine their relative quantity in the mixture. The reaction under study was the hydrogenation
of artemisinic acid, an intermediate step in the semi-synthesis of the anti-malarial drug
artemisinin. Four analytes were characterized in each measurement: the starting material, the
product, a diastereomer of the product, and an overreduction byproduct that was not directly
quantifiable by either HPLC or NMR methods. The MRR instrument has a measurement cycle
time of approximately 17 minutes for this analysis, and can run for several hours without any
user interaction.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Keywords
Process analytical technology; PAT; reaction monitoring; impurities; on-line spectroscopy;
molecular rotational resonance; MRR; rotational spectroscopy; stereoisomers
Introduction
On-line reaction monitoring techniques are extremely valuable for enabling rapid process
understanding, development, and optimization. In particular, quickly characterizing the yield,
specificity, and impurities generated in a chemical reaction is very important. In contrast to off-
line analyses, which require the user to manually sample, transport, measure, and read results,
on-line instrumentation can run continuously during a chemical reaction, archiving data and
alerting the process chemist when an out-of-specification result is obtained. As a result,
extensive efforts have been undertaken to apply the suite of analytical chemistry instrumentation
to on-line reaction monitoring applications, including FT-IR1-2 and NMR3-6 spectroscopies, mass
7
8-12
spectrometry, and chromatography. Nevertheless, significant challenges remain in
characterizing all of the relevant species present in a chemical process, additional development is
still needed. The United States Food and Drug Administration (FDA), through its Process
3
ACS Paragon Plus Environment

Organic Process Research & Development
Page 4 of 18
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Analytical Technology (PAT) initiative, encourages these developments to further its goals of
improved process control and real-time testing for quality assurance.13
One critical parameter in pharmaceutical manufacturing that is often not monitored
online, due to the absence of suitable methods, is the chiral purity of the product being
synthesized. The number of new small-molecule drugs with at least one chiral center has
increased considerably over the past few decades, with most of these now manufactured as single
enantiomers.14,15 In addition, many drugs contain multiple chiral centers, leading to additional
stereoisomers known as diastereomers. A variety of chirally-sensitive chromatography and
spectroscopy techniques have been developed for resolution of enantiomers, while for
diastereomers – which are chemically similar but have different three-dimensional structures –
techniques with high structural selectivity are needed.16-18 Techniques for distinguishing each of
the possible stereoisomers of a molecule – both enantiomers and diastereomers – are necessary in
order to obtain full process analysis and control in the synthesis of chiral drugs.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Only a few studies demonstrating on-line reaction monitoring of stereoisomers have been
published to date. Several studies have shown the capability for vibrational circular dichroism
(
VCD) to monitor the enantiomeric excess of a chemical reaction19-21, while fast chiral
chromatography methods have also been demonstrated.22,23 In this paper, we present the
application of molecular rotational resonance (MRR) spectroscopy for on-line reaction
monitoring of a pharmaceutical process. MRR, also widely known as microwave or rotational
spectroscopy, is extremely sensitive to small differences in molecular structure, and as such is
24
well suited for the differentiation of isomers. Due to the high spectral resolution of MRR, a
mixture of components, including isomers (regioisomers, diastereomers, cis-trans isomers, or
isotopologues), can be resolved without the need for separation or chemometrics. Additionally,
4
ACS Paragon Plus Environment

Page 5 of 18
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
while not demonstrated directly in this work, recent advances have enabled the resolution of
2
5
enantiomers by MRR using two methods – chiral tagging and three-wave mixing. Chiral
tagging can be performed to determine enantiomeric purity on the instrument described on this
paper without modification. This work is motivated by recent advances in continuous
manufacturing technology in pharmaceutical processing,26-28 and seeks to demonstrate the
feasibility of MRR as a technique for directly measuring stereoisomeric purity, as well as other
impurities in the reaction, at relevant points along small-molecule synthetic processes.
The reaction under study is the Ru/C catalyzed hydrogenation of artemisinic acid (AA),
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
9
which can be produced from genetically engineered yeast, to dihydroartemisinic acid (DHAA)
in a trickle-bed flow reactor. DHAA can thereafter be converted into artemisinin through photo-
30
oxidation ; the diastereomeric selectivity of the hydrogenation step is critical in achieving high
yield of the final product, artemisinin.31-33
Experimental Methods
Broadband MRR spectroscopy for component identification
In order to determine the MRR transition frequencies for each component in the
hydrogenation reaction, an off-line 2-8 GHz broadband MRR spectrometer was employed. The
design of this instrument has been described elsewhere.34,35 A total of four samples were
measured: commercially obtained crystalline high-purity samples of artemisinic acid (AA) and
dihydroartemisinic acid (DHAA); an intentionally over-reduced sample in order to produce the
overreduction byproduct, tetrahydroartemisinic acid (THAA); and finally, a crude reaction
mixture containing AA, DHAA, THAA, and a diastereomeric impurity of DHAA (epimerized at
5
ACS Paragon Plus Environment

Organic Process Research & Development
Page 6 of 18
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
the chiral center created by the hydrogenation). The latter two samples were provided as oils
(approximately 70 mg each).
Solid samples were transferred directly into a 0.9 mm Parker Hannefin Series 9 pinhole
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
nozzle, modified to incorporate a sample reservoir that can be heated to volatilize liquid and
solid samples. Oil samples were first dissolved in ethanol (at a concentration of approximately
1
00 mg/mL) in order to allow for their transfer by syringe into the nozzles. The sample was split
among 3 nozzles, oriented in parallel, run simultaneously in order to increase measurement
sensitivity. Samples were heated to approximately 160°C to produce suitable vapor pressure (~1
Torr) for analysis. High-grade (>99.999%) neon was used as the carrier gas.
Process monitoring MRR spectrometer
The process MRR spectrometer employed in this study is presented in Figure 1. The
instrument design is based on the Fabry-Perot microwave cavity spectrometer originally
36
37
developed by Balle and Flygare and later miniaturized by Suenram et al. The dimensions of
this spectrometer are the same as those described by Suenram et al. A pair of spherical aluminum
-6
mirrors, held under high vacuum (~10 Torr), are coupled to a microwave generator that
produces pulses with a frequency that is resonant with a mode of the cavity. The vacuum is
generated by a 6-inch diffusion pump with a rotary vane backing pump. Volatilized molecular
samples are injected into the cavity region in synchronization with the microwave pulses. When
the microwave pulse is resonant with a molecular rotational transition of a component in the
sample, the sample emits a free induction decay (FID) signal, which is coupled out of the cavity,
digitized, and Fourier transformed to yield the MRR spectrum.
6
ACS Paragon Plus Environment

Page 7 of 18
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
In comparison to the broadband spectrometer, this cavity-enhanced MRR spectrometer is
faster and consumes significantly less sample (by a factor of approximately 100) to achieve a
given sensitivity level. However, this instrument can only detect a narrow bandwidth in one
measurement - in most cases, a single molecular transition around one of the resonant cavity
modes - due to the high cavity finesse. One of the mirrors in the cavity spectrometer is motor
tunable, so that its position can be changed in about 5 seconds to measure a different transition
within the spectrometer bandwidth (9-18 GHz). Transitions from each of the components of
interest are measured sequentially during each measurement cycle.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 1: Photograph of the MRR spectrometer used for the measurements presented in this study.
The amount of sample required for analysis by MRR - both in the initial broadband
characterization, and the process measurements - depends on the level of sensitivity needed. This
is because the sample, after it is volatilized and injected into the MRR spectrometer, is evacuated
through the system’s vacuum system. To this point, we have not undertaken any attempts to
recover samples. In the case of the reaction studied here, the strongest components in the mixture
7
ACS Paragon Plus Environment

Organic Process Research & Development
Page 8 of 18
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
can be observed in a broadband measurement with 10 mg or less of sample – or with
approximately 100 g of sample in a targeted analysis (where the MRR spectrum of the analytes
is already known). However, since the signal-to-noise ratio of the MRR spectrum depends on the
measurement time, if lower-level impurities are desired, more sample and more measurement
time are needed. The amount of sample required to reach a given sensitivity level also depends
on the molecule’s intrinsic MRR spectral strength – which depends primarily on the permanent
dipole moment and the molecular weight of each analyte.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Sampling interface
A schematic and photographs of the interface between the continuous flow reactor and
the process MRR spectrometer are presented in Figure 2. The output of the flow reactor is
directed into a gas-liquid flow separator, which was designed to give the MRR spectrometer
access to a constantly refreshing reservoir of fresh reaction mixture for analysis while allowing
the hydrogen gas in the line to escape. As more of the product solution comes out of the reactor,
the solution in the reservoir overflows and exits the separator. We estimate that this reservoir,
with a volume of about 2 mL, turns over approximately every 2 minutes under typical reaction
flow rates.
When the MRR spectrometer is ready to begin a new analysis, a solenoid dosing pump is
used to deliver a fixed volume of the reaction mixture through a valving manifold and into a
heated reservoir nozzle. This nozzle is the same design as for the broadband MRR measurements
described in the previous section, except it was fitted with two tube connections on the sides of
the nozzle to allow direct introduction of solutions. This nozzle is inserted into the back of the
stationary cavity mirror. The reservoir is first heated to a temperature slightly below the boiling
8
ACS Paragon Plus Environment

Page 9 of 18
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
point of the solvent, with the MRR spectrometer tuned to a solvent transition to monitor the
remaining solvent concentration in the sample. For this reaction, with ethanol (bp 78.4°C) as the
solvent, a temperature of 75°C was chosen. Once the solvent is nearly evaporated, the reservoir
temperature is increased to 160°C to volatilize the analytes. The selected transitions of each
species of interest are then measured, followed by a brief period of reservoir evacuation at a
higher temperature (in this case, 190°C) to evacuate any remaining sample. Once the reservoir is
clean, the nozzle is cooled back to the solvent volatilization temperature to receive the next
sample. The total cycle time to analyze the 4 observed species in the DHAA reaction, including
sample delivery, measurement, and cleaning, is about 17 minutes with the current system. The
amount of sample consumed per measurement is 0.99 mg (330 L of solution mixture with a
total analyte concentration of 3 mg/mL). The reservoir has a usable volume of approximately
500 L; however, methods can be developed to accommodate larger injection volumes if
necessary (if, for example, the analyte concentration in the solution is very low).
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
9
ACS Paragon Plus Environment

Organic Process Research & Development
Page 10 of 18
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2: Images of the interface between the continuous hydrogenation process and the MRR spectrometer for reaction
monitoring. Top left: The flow catalysis reactor. A flow separator is used to collect a reservoir of liquid and allow the hydrogen
gas to escape. Top right: An internal cutaway drawing of the flow separator. Bottom left: Schematic of the valving manifold to
transfer the reaction solution into the nozzle for volatilization. Bottom right: photograph of the valving manifold mounted to the
spectrometer. (The blue cables observed in this figure are the microwave signal input and output.)
Results and Discussion
Broadband spectroscopy of reaction mixture
In Figure 3, the broadband MRR spectrum of the crude DHAA reaction mixture is
presented. Each component in the species is identified by modeling its spectrum to a rotational
Hamiltonian, where the primary derived parameters are the three molecular rotational constants
(proportional to the inverse of the moments of inertia along the three principal axes of the
molecule). Each component in this mixture, has two low-energy conformational isomers,
differing by a 180° rotation of the carboxylic acid group; each of these is detected as a unique
pattern in the spectrum. Compounds are identified by comparison of the experimental rotational
constants of each fit to calculated values from electronic structure theory (see the Supporting
Information for further description). Because the spectrum consists of approximately 100,000
spectral resolution elements, and each analyte spectrum usually consists of between 100-200
10
ACS Paragon Plus Environment

Page 11 of 18
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
detectable transitions, overlaps between the spectral transitions of different species are extremely
rare, as Figure 3 shows. Therefore, we can monitor the concentration of each species without
interference by monitoring a single line known to be assigned to that species. A more complete
analysis of these spectra, including structural parameters of these species as derived from the
molecular moments-of-inertia, will be presented in a future paper.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
We note that the bandwidth of the process monitoring MRR spectrometer (9-18 GHz) did
not overlap with the broadband MRR spectrometer used in this study (2-8 GHz). We used the fits
from the broadband spectrometer to predict the frequencies and intensities of strong transitions in
the bandwidth of the process spectrometer, which were confirmed by subsequent measurement
of trial reaction mixture samples.
11
ACS Paragon Plus Environment

Organic Process Research & Development
Page 12 of 18
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3: Broadband MRR spectrum of a crude reaction mixture from a continuous hydrogenation of artemisinic acid. The black
spectrum is the full reaction mixture, while the colored traces are masked to show the individual contributions of each
component (using the Hamiltonian fits to each species). The bottom panel shows an expanded view of the spectrum to show
that the rotational lines of each component do not overlap with each other.
Process measurements
We next demonstrated the ability of the spectrometer to quantitatively determine the
composition of a two-component mixture, using prepared mixtures of the starting material (AA)
and purified product (DHAA) at five different ratios from 95:5 (DHAA:AA) to 50:50. The
results of this analysis are presented in Figure 4. The instrument was operated under the same
12
ACS Paragon Plus Environment

Page 13 of 18
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
conditions as in the process monitoring measurements, except the samples were injected via
syringe from prepared solutions at each composition. The sum of the AA and DHAA
concentrations was kept constant at 3 mg/mL in each sample, the same concentration as was used
in the flow process.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 4: MRR measurements of prepared solutions of AA and DHAA, diluted to a combined concentration 3 mg/mL in ethanol.
Panels (a) and (b) show the raw spectra of DHAA and AA obtained from each mixture, while panel (c) plots the measured AA
fraction (defined by the ratio of the AA peak signal to the sum of the two signals) against the actual AA fraction. The overlaid
lines show a slope of 1 (red) and a best fit to the data, with a fixed intercept of zero (green).
In quantifying the relative concentration of each species, we account for the conformer
populations (determined directly from the broadband spectrum), calculated line strength, and the
population difference between the two states measured (assuming a typical rotational
temperature of 2 K). We did not correct for any possible differences in the vapor pressure of each
analyte in the mixture at the measurement temperature. In this reaction, the molecular weights
and structures of the different analytes are similar, and based on our observations, any vapor
pressure differences appear to be small. If desired, a calibration mixture could be used for
correction, if available, or benchmarking to another analytical technique.
In Figure 5 we show example data for monitoring the continuous hydrogenation of AA to
DHAA. During this measurement the conditions (flow rate, temperature, and pressure) were
13
ACS Paragon Plus Environment

Organic Process Research & Development
Page 14 of 18
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
intentionally varied during the synthesis in order to initiate significant changes in the product
distribution between samples. All four species of interest are monitored in each measurement.
We note that THAA in particular, because of its saturated nature, is highly challenging for
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
analysis by either HPLC (with UV detection) or H NMR. MRR does not require a chromophore
in the same sense as these other techniques; it requires only a permanent dipole moment, which
is almost always present in chiral molecules. Therefore, MRR can help to fill analytical gaps of
other techniques.
Figure 5: Left: Concentration data for a series of 8 sample measurements performed over the course of 4 hours with the MRR
1
spectrometer connected to the synthesis flow. Right: Offline H NMR analysis of reaction aliquots collected at the same time as
the MRR analysis.
At the same time the MRR spectrometer drew a sample for analysis, a fraction of the
1
reaction mixture was collected for comparative off-line analysis by H NMR (Bruker 600 MHz).
In these measurements, an internal standard (dimethyl sulfone) was added to determine the
absolute product yield. However, due to the lack of a characteristically-shifted proton for THAA,
it was not possible to determine its concentration directly. The MRR analysis instead determines
the product ratios between the four relevant species. This may explain the lower overall DHAA
percentage in the NMR measurements. Regardless, it can be seen that the trends in the
14
ACS Paragon Plus Environment

Page 15 of 18
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
concentrations of the three species monitored in both samples is similar between the two
analytical techniques.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Conclusion
This paper represents the first demonstration of molecular rotational resonance
spectroscopy for characterization of the purity of a pharmaceutical synthetic process. The pivotal
advantage of MRR over other process monitoring techniques is its high resolution and
specificity: different species, including regioisomers, diastereomers, and enantiomers, can be
individually resolved and quantified in a reaction mixture unambiguously. We show that the
MRR technique has extremely good structural selectivity and good quantitative accuracy.
Because advance sample preparation, internal standards, and calibration measurements are not
needed, this represents a significant new capability for reaction monitoring in continuous
synthetic processes. This method is broadly applicable to analyses of stable intermediates and
products in small-molecule pharmaceutical manufacturing.
Acknowledgements
The authors acknowledge support for this project from the Virginia Biosciences Health
Research Corporation. This work at the University of Virginia was supported by the National
Science Foundation (CHE-153193). This work at BrightSpec was supported by the ARMY SBIR
program (contract W31P4Q-15-C-0019).
15
ACS Paragon Plus Environment

Organic Process Research & Development
Page 16 of 18
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Conflict of Interest Disclosure
Authors J.L.N, M.T.M, and R.L.R are employed by BrightSpec, Inc. and have financial
interests in BrightSpec. Author B.F.G. serves on the BrightSpec Scientific Advisory Board.
Author B.H.P. has a financial interest in BrightSpec.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Supporting Information
The structures and MRR fit and calculated parameters of the species characterized in this
study are available as Supporting Information.
References
1. Carter, C.F.; Lange, H.; Ley, S.V.; Baxendale, I.R.; Wittkamp, B.; Goode, J.G.; Gaunt, N.L.
ReactIR flow cell: A new analytical tool for continuous flow chemical processing. Org. Proc.
Res. & Dev. 2010, 14, 393-404.
2
. Moore, J.S.; Jensen, K.F. “Batch” kinetics in flow: Online IR analysis and continuous control.
Angew. Chem. Int. Ed. 2014, 53, 470-473.
3
. Foley, D.A.; Wang, J.; Maranzano, B.; Zell, M.T.; Marquez, B.L.; Xiang, Y.; Reid, G.L. Online
NMR and HPLC as a reaction monitoring platform for pharmaceutical process development.
Anal. Chem. 2013, 85, 8928-8932.
4
. Foley, D.A.; Doechke, C.W.; Buser, J.Y.; Merritt, J.M.; Murphy, L.; Kissane, M.; Collins, S.G.;
Maguire, A.R.; Kaerner, A. ReactNMR and ReactIR as reaction monitoring and mechanistic
elucidation tools: The NCS mediated cascade reaction of a-thioamides to a-thio-b-
chloroacrylamides. J. Org. Chem. 2011, 76, 9630-9640.
1
9
5
6
. Zientek, N.; Laurain, C.; Meyer, K.; Kraume, M.; Guthausen, G.; Maiwald, M. Simultaneous F-
1
H medium resolution NMR spectroscopy for online reaction monitoring. J. Mag. Reson. 2014,
2
49, 53-62.
. Maiwald, M.; Steinhof, O.; Sleigh, C.; Bernstein, M.; Hasse, N. Quantitative high-resolution
online NMR spectroscopy in pharmaceutical reaction and process monitoring. In NMR
Spectroscopy in Pharmaceutical Analysis, 1 Ed.; Holzgrabe, U., Wawer, I., Biehl, B., Eds;
Elsevier: Oxford, 2008; 471-491.
st
7. Browne, D.L.; Wright, S.; Deadman, B.J.; Dunnage, S.; Baxendale, I.R.; Turner, R.M.; Ley, S.V.
Continuous flow reaction monitoring using an on-line miniature mass spectrometer. Comm. Mass
Spectrom. 2012, 26, 1999-2010.
8
9
1
. Zawatsky, K.; Grosser, S.; Welch, C.J. Facile kinetic profiling of chemical reactions using
MISER chromatographic analysis. Tetrahedron 2017, 73, 5048-5053.
. Welch, C.J.; Gong, X.; Cuff, J.; Dolman, S.; Nyrop, J.; Lin, F.; Rogers, H. Online analysis of
flowing streams using microflow HPLC. Org. Proc. Res. & Dev. 2009, 13, 1022-1025.
0. Schafer, W.A.; Hobbs, S.; Rehm, J.; Rakestraw, D.A.; Orella, C.; McLaughlin, M.; Ge, Z.;
Welch. C.J. Mobile tool for HPLC reaction monitoring. Org. Proc. Res. & Dev. 2007, 11, 870-
8
76.
16
ACS Paragon Plus Environment

Page 17 of 18
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
1
1
1
1. Malig, T.C.; Koenig, J.D.B; Situ, H.; Chehal, N.K.; Hultin, P.G.; Hein, J.E. Real-time HPLC-MS
reaction progress monitoring using an automated analytical platform. React. Chem. & Eng. 2017,
2
, 309-314.
2. Li, B.; Guo, W.; Chi, H.; Kimura, M.; Ramsey, E.D. Monitoring the progress of a photochemical
reaction performed in supercritical fluid carbon dioxide using a continuously stirred reaction cell
interfaced to on-line SFC. Chromatographia 2017, 80, 1179-1188.
3. Guidance for Industry, PAT – A Framework for Innovative Pharmaceutical Development,
Manufacturing, and Quality Assurance; U.S. Department of Health and Human Services, Food
and Drug Administration: Rockville, MD, September 2004. Available on the Web at
https://www.fda.gov/downloads/drugs/guidances/ucm070305.pdf.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
4. Shen, Z.; Lv, C.; Zeng, S. Significance and challenges of stereoselectivity assessing methods in
drug metabolism. J. Pharm. Anal. 2016, 6, 1-10.
15. Novel Drug Approvals for 2018. U.S. Department of Health and Human Services, Food and Drug
Administration: Rockville, MD. Accessed February 5, 2019. Available on the Web at
https://www.fda.gov/drugs/developmentapprovalprocess/druginnovation/ucm592464.htm.
16. Ebinger, K.; Weller, H.N. Comparison of chromatographic techniques for diastereomer
separation of a diverse set of drug-like compounds. J. Chromatogr. A 2013, 1272, 150-154.
1
7. Liu, W.; Gan, J.J. Separation and analysis of diastereomers and enantiomers of cypermethrin and
cyfluthrin by gas chromatography. J. Agric. Food Chem. 2004, 52, 755-761.
8. Adams, R.W.; Byrne, L.; Kiraly, P.; Foroozandeh, M.; Paudel, L.; Nilsson, M.; Clayden, J.;
Morris, G.A. Diastereomeric ratio determination by high sensitivity band-selective pure shift
NMR spectroscopy. Chem. Commun. 2014, 50, 25122-2514.
9. Guo, C.; Shah, R.D.; Dukor, R.K.; Cao, X.; Freedman, T.B.; Nafie, L.A. Determination of
enantiomeric excess in samples of chiral molecules using Fourier transform vibrational circular
dichroism spectroscopy: simulation of real-time reaction monitoring. Anal. Chem. 2004, 76,
1
1
2
6
956-6966.
0. Guo, G.; Shah, R.D.; Mills, J.; Dukor, R.K.; Cao, X.; Freedman, T.B.; Nafie, L.A. Fourier
transform near-infrared vibrational circular dichroism used for on-line monitoring the
epimerization of 2,2-dimethyl-1,3-dioxolane-4-methanol: A pseudo racemization reaction.
Chirality 2006, 18, 775-782.
21. Mower, M.P.; Blackmond, D.G. In-situ monitoring of enantiomeric excess during a catalytic
kinetic resolution. ACS Catal. 2018, 8, 5977-5982.
22. Shen, Z.; Lv, C.; Zeng, S. Significance and challenges of stereoselectivity assessing methods in
drug metabolism. J. Pharm. Anal. 2016, 6, 1-10.
2
2
2
3. Welch, C.J. Are we approaching a speed limit for the chromatographic separation of
enantiomers? ACS Cent. Sci. 2017, 3, 823-829.
4. Gordy, W.; Cook, L.R. Microwave Molecular Spectra, 3 Ed. John Wiley & Sons Inc.: New
rd
York, 1984.
5. Pate, B.H.; Evangelisti, L.; Caminati, W.; Xu, Y.; Thomas, J.; Patterson, D.; Perez, C.; Schnell,
M. Quantitative chiral analysis by molecular rotational spectroscopy. In Chiral Analysis:
nd
Advances in Spectroscopy, Chromatography, and Emerging Methods, 2 Ed: Polavarapu, P, Ed.;
Elsevier, 2018.
2
2
6. Adamo, A.; Beingessner, R.L.; Behnam, M.; Chen, J.; Jamison, T.F.; Jensen, K.F.; Monbaliu, J.-
C. M.; Myerson, A.S.; Revalor, E.M.; Snead, D.R.; Stelzer, T.; Weeranoppanant, N.; Wong,
S.Y.; Zhang, P. On-demand continuous/ flow production of pharmaceuticals in a compact,
reconfigurable system. Science 2016, 352, 61-67.
7. Cole, K.P.; McClary Groh, J.; Johnson, M.D.; Burcham, C.L.; Campbell, B.M.; Diseroad, W.D.;
Heller, M.R.; Howell, J.R.; Kallman, N.J.; Koenig, T.M.; May, S.A.; Miller, R.D.; Mitchell, D.;
Myers, D.P.; Myers, S.S.; Phillips, J.L.; Polster, C.S.; White, T.D.; Cashman, J.; Hurley, D.;
Moylan, R.; Sheehan, P.; Spencer, R.D.; Desmond, K.; Desmond, P.; Gowran, J. Kilogram-scale
17
ACS Paragon Plus Environment

Organic Process Research & Development
Page 18 of 18
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
prexasertib monolactate monohydrate synthesis under continuous-flow CGMP conditions.
Science 2017, 356, 1144-1150.
8. Korwar, S.; Amir, S.; Tosso, P.; Desai, B.K.; Kong, C.J.; Fadnis, S.; Telang, N.S.; Ahmad, S.;
Roper, T.D.; Gupton, B.F. The application of a continuous Grignard reaction in the preparation of
fluconazole. Euro. J. Org. Chem. 2017, 2017, 6495-6498.
9. Ro, D.-K.; Paradise, E.M.; Ouellet, M.; Fisher, K.J.; Newman, K.L.; Ndungu, J.M.; Ho, K.A.;
Eachus, R.A.; Ham, T.S.; Kirby, J.; Chang, M.C.Y.; Withers, S.T.; Shiba, Y.; Sarpong, R.;
Keasling, J.D. Production of the antimalarial drug precursor artemisinic acid in engineered yeast.
Nature 2006, 440, 940-943.
0. Kong, C.J.; Fisher, D.; Desai, B.K.; Yang, Y.; Ahmad, S.; Belecki, K.; Gupton, B.F. High-
throughput photo-oxidations in a packed bed reactor system. Bioorganic Med. Chem. 2017, 25,
6203-6208.
31. Turconi, J.; Griolet, F.; Guevel, R.; Oddon, G.; Villa, R.; Geatti, A.; Hvala, M.; Rossen, K.;
Goller, R.; Burgard, A. Semisynthetic artemisinin, the chemical path to industrial production.
Org. Proc. Res. & Dev. 2014, 18, 417-422.
32. Pieber, B.; Glasnov, T.; Kappe, C.O. Continuous flow reduction of artemisinic acid utilizing
multi-injection strategies—closing the gap towards a fully continuous synthesis of antimalarial
drugs. Chem. Eur. J. 2015, 21, 4368-4376.
3. Feth, M.P.; Rossen, K.; Burgard, A. Pilot plant PAT approach for the diastereoselective diimide
reduction of artemisinic acid. Org. Proc. Res. & Dev. 2013, 17, 282-293.
4. Brown, G.G.; Dian, B.C.; Douglass, K.O.; Geyer, S.M.; Shipman, S.T.; Pate, B.H. A broadband
Fourier transform microwave spectrometer based on chirped pulse excitation. Rev. Sci. Instrum.
2
2
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
3
3
2
008, 79, 053103.
3
5. Perez, C.; Lobsiger, S.; Seifert, N.A.; Zaleski, D.P.; Temelso, B.; Shields, G.C.; Kisiel, Z.; Pate,
B.H. Broadband Fourier transform rotational spectroscopy for structure determination: The water
heptamer. Chem. Phys. Lett. 2013, 571, 1-15.
6. Balle, T.J.; Flygare, W.H. Fabry-Perot cavity pulsed Fourier transform microwave spectrometer
with a pulsed nozzle source. Rev. Sci. Instrum. 1981, 52, 33-45.
7. Suenram, R.D.; Grabow, J.-U.; Zuban, A.; Leonov, I. A portable, pulsed-molecular-beam,
Fourier-transform microwave spectrometer designed for chemical analysis. Rev. Sci. Instrum.
1999, 70, 2127.
3
3
18
ACS Paragon Plus Environment