Taking compact NMR to monitoring real reactions in large-scale chemical industries—General considerations and learnings from a lab-scale test case

Article abstract of DOI:10.1002/mrc.5061

The considerations for use of compact nuclear magnetic resonance in a large-scale industrial environment clearly differ from those in academic and educational settings and even from those in smaller companies. In the first part of this article, these differences will be discussed along with the additional requirements that need to be fulfilled for successful applicability in different use cases. In the second part of the article, outcomes from different research activities aiming to fulfill these requirements will be presented with a focus on an online reaction-monitoring study on a lab-scale nucleophilic chlorination reaction.

Full text of DOI:10.1002/mrc.5061

Nestle Nikolaus (Orcid ID: 0000-0001-9894-4437)
Taking compact NMR to monitoring real reactions in large-scale chemical
industries – general considerations and learnings from a lab-scale test case
1*
2
2
2
1
Nikolaus Nestle , Zi Jian Lim , Tobias Böhringer , Sarah Abtmeyer , Sven Arenz ,
2
1
3
Felix C. Leinweber , Thomas Weiß , Erik v. Harbou
1
2
3
) BASF SE, Material Physics and Analytics, Carl-Bosch-Straße, 67056
Ludwigshafen am Rhein, Germany
) BASF SE, Center of Expertise for Process Analytical Technology, Carl-Bosch-
Straße, 67056 Ludwigshafen am Rhein, Germany
) BASF SE, Process Reseach and Chemical Engineering, Carl-Bosch-Straße,
6
7056 Ludwigshafen am Rhein, Germany
*
Corresponding author’s email: nikolaus.nestle@basf.com
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1002/mrc.5061
This article is protected by copyright. All rights reserved.

Abstract:
The considerations for use of compact NMR in a large-scale industrial environment
clearly differ from those in academic and educational settings and even from those in
smaller companies. In the first part of this article, these differences will be discussed
along with the additional requirements that need to be fulfilled for successful
applicability in different use cases. In the second part of the article, outcomes from
different research activities aiming to fulfill these requirements will be presented with
a focus on an online reaction-monitoring study on a lab-scale nucleophilic chlorination
reaction.
Introduction:
Over the last ten years, compact NMR spectroscopy equipment has made great
progress from educational demonstrator systems to a serious analytical lab tool [1].
These instruments are based on well-homogenized permanent magnet setups and
usually operate at proton NMR frequencies between 40 and 90 MHz.
Progress has been made especially in signal-to-noise ratio and sensitivity as well as
in the development of adapted data processing routines. Nevertheless, limitations with
respect to the different ratio between couplings and natural line widths and sensitivity
will always persist compared to high-field NMR.
While these differences are only of minor importance in educational applications, they
will need some more detailed consideration in the corporate world. In most large
chemical companies, some internal high-field NMR capabilities are available, usually
at some central research sites. For samples from these sites, transfer times tend to be
short (minutes to half hour). Like that, the use case for compact NMR equipment will
This article is protected by copyright. All rights reserved.

mainly be limited to cases like reaction monitoring or high-throughput screening in
which a large number of samples requiring analysis with minimal transfer time are
produced.
For samples from sites with no NMR, longer transfer times and cost from safety
requirements for shipping will have an influence on the overall assessment of
economic viability: even a few samples per day may already provide a viable case for
installation of a compact NMR system on site. The increased workload for operating
and maintaining the compact NMR should be balanced against the time spent on
packaging samples for shipping and the documentation required to do so.
In all cases, a careful alignment of the spectra from compact NMR with those from
classical high-field NMR is needed and the persons involved in interpretation of the
compact NMR data need to be especially trained in recognizing spectra with possible
peculiarities that suggest the need for a second tier measurement by high-field NMR.
In order to avoid possible ageing effects of the samples during transfer, the alignment
of high-field and compact NMR spectra should be done by simultaneous
measurements wherever possible.
Of the different use scenarios for compact NMR, reaction monitoring is the most
challenging as measurement conditions as close as possible to the actual reaction are
desirable. In many cases, this corresponds to measurements at elevated pressure and
temperature in order to avoid:

physical changes to the sample, such as evaporation and foaming or
precipitation of certain components due to temperature-induced solubility
changes

chemical changes, such as shifts in chemical equilibria or the reaction of
intermediates along different pathways.
This article is protected by copyright. All rights reserved.

If those challenges are mastered, the information from NMR measurements can be
used for automated adjustment of process parameters to increase product yield and
asset utilization [2]. In addition to measuring at reaction conditions, NMR
instrumentation for process analytics needs to come with a sufficiently rugged and
possibly explosion-proofed housing in order to be usable under process site
conditions. Like this, the compact NMR with no need for cryogenic magnets and with
a compact footprint comes with unique advantages over classical NMR in this field of
application.
Furthermore, the NMR instrument needs to be capable of automatic operation and
data processing with minimal human user involvement. On the hardware side, this
requires a flow-through sample cell. Historically, NMR flow-cells have mainly been
used in the context of hyphenated analytical methods, such as HPLC-NMR [3,4] and
electrophoretic separation NMR [5,6]. More recently, some reaction monitoring was
performed on high-field NMR systems with dedicated model process setups installed
in the immediate vicinity of high-field NMR systems [7], often with a focus on
pharmaceutical products [8,9 ]. Flow cells in these studies range from flow-through
systems for use in a conventional NMR probe head [8] to dedicated probe heads with
sophisticated temperature control systems [10].
With the availability of compact NMR setups with better and better spectral
performance, an increasing number of publications on reaction monitoring on lab
setups using compact NMR have been published over the last decades [11, 12]. In
many of these studies, the degree of automation remained rather low and the focus
was on a proof of principle. Such studies may only be considered as intermediate
steps towards NMR systems for 24/7 reaction control in process industry. Later work
includes the optimization of lab-scale organic syntheses [13, 14] and the determination
This article is protected by copyright. All rights reserved.

of chemical engineering parameters, such as distribution coefficients [15]. Newer work
includes the applicability of compact NMR systems to study phase equilibria and
reactions [16] and to monitor dynamic processes, such as batch distillation [17]. Kern
et al. demonstrated that compact NMR facilitates the optimization [18] and control [19]
of complex chemical processes that are relevant for technical or pharmaceutical
applications.
In the present work, a fully automated NMR system based on a prototype compact
NMR spectrometer has been developed which fulfills three of the above-mentioned
requirements (the fourth, explosion protection, has also been fulfilled by building an
Ex p II T4 rated inert-gas flushed box with a flexible barrier around the sample flow
line even allowing measurements on combustible samples together with a local
specialized company, but that development is out of scope of the present article). Its
applicability in real life syntheses and especially the proper collaboration of parts from
many different vendors is shown by online monitoring of the chlorination reaction path
of an alcohol into its alkyl chloride using both NMR’s structural and quantitative
information. The chosen test reaction is a good example for an industrially relevant
synthesis. The results demonstrate that the developed automated NMR system is a
very useful and versatile tool with a broad fiel d of application in industries among
others screening studies (e.g. catalysts, solvents, process parameters) in an early
stage of process development or for fast quality control in production plants.
System design and component validation
In order to use a compact NMR system in process analytics, several changes
compared to a lab instrument need to be made. Some involve the actual NMR
instrument which needs to be equipped with a flow- cell suitable for operation at
This article is protected by copyright. All rights reserved.

elevated temperatures and pressures without negative impact on the probe and
magnet of the instrument. The remaining challenges involve the integration of the NMR
instrument with appropriate sampling hardware and the development of an automation
platform able to control sampling and measurement as well as an evaluation of the
NMR signals without a need for human intervention.
The automatic control was implemented in the standard modular lab automation
platform of BASF running on an industrial PC obtained from Beckhoff (Verl, Germany).
The different sampling and analytical modules are connected via RS-232 to this setup.
The sampling system consists of a membrane pump from Fink ChemTech (Leinfelden-
Echterdingen, Germany), a actively-automated 4-way-valve from VICI (Schenkon,
Switzerland), Julabo heating jackets (Julabo, Seelbach, Germany) and PEEK and
Teflon tubings of 2.0 mm ID and 4 mm from IDEX (Oak Harbor, WA, USA). As a
compact NMR spectrometer, a prototype NMReady 60 compact NMR spectrometer
from Nanalysis Inc. (CalgaryAB, Canada) with a 1.4 T Halbach magnet, equipped with
a Dewar with 5.3 mm internal bore diameter and 60 MHz proton resonance frequency
was used. The specially constructed probe with the removable Dewar is sketched in
figure 1. Control of this instrument via the industrial PC was performed using the
NMReady Control interface API provided by Nanalysis.
For the process analytical measurements, the instrument was equipped with a 5 mm
coaxial flowcell purchased from Bruker Biospin (Rheinstetten, Germany) [20,21] in
combination with the InSightMR package. This package consists of a combination of
three software tools from Bruker:Dynamic Center for integration and evaluation of
integration results, TopSpin for automated processing of NMR spectra and IconNMR
for automated conduction of NMR experiments. The software package was adapted
by Bruker to the special needs of this project, c.f. raw data were stored in a special
This article is protected by copyright. All rights reserved.

watch-dog folder, NMR signals were automatically processed and data were saved in
a general Excel format. Additionally, the length of the glass tube of the flow cell was
adapted to reach the active volume of the Nanalysis NMR spectrometer. Further
details of the InsightMR package can be found on the Bruker homepage [22]. The
modified software package only consists of standard commercial products and is now
available on the market as option in standard InsightMR. For visualization of
processed data and spectra, some further standard software packages, such as
MNova (Mestrelab, Santiago de Compostela, Spain) and Origin ( Originlab,
Northampton MA USA) were used.
In order to validate that the modified prototype NMR system is actually capable of
serving the needs of reaction monitoring, several sets of additional experiments were
performed. A range of different tests was conducted to explore the impact of the
sample temperature on the stability of the NMR measurements. In a first set of tests,
both an NMReady with a standard probe and the prototype instrument with the Dewar
were subjected to a mock at-line study with a sample tube filled with n-octanol
preheated to 100 °C. This sample was inserted into the instrument for about 1 min to
conduct a standard measurement. After that, the hot sample was removed and a
water-filled standard shimming sample was inserted and the line width as determined
by the shimming routine was recorded as a function of time. In both cases, two runs
of this experiment were performed. The results are given in figure 2.
As can be seen from the figure, there is thermally introduced magnet inhomogeneity
in both systems but the impact of the exposure to the hot sample is much more
pronounced in the spectrometer with the standard probe. After about 15 min, the
impact of the heat exposure is compensated for both systems and line widths very
This article is protected by copyright. All rights reserved.

close to the original ones reached. (I.e. no long-term damage on the magnet is
inflicted.) Furthermore, it is interesting to note that there is sufficient thermal inertia in
the system to record a single good-quality spectrum on a hot sample every 30 min
(provided the temperature change in the sample during the measurement has no
detrimental impact on the quality of the spectrum).
In a further experiment, the sensitivity of the line width of DMSO circulating at various
temperatures through the flow cell was studied in stopped flow measurements with
different inlet temperatures for the DMSO flow through the cell. The results are shown
in figure 3. As one can see from the figure, there is a pronounced impact even of
moderate temperature differences between magnet (constantly kept at 32°C for
optimal performance) and flowing medium on the line width. Like that, a sample
temperature of 45°C was chosen for the continuous reaction monitoring experiments
shown in the next section of this paper as a compromise between ensuring no
precipitation and achieving as sharp lines as possible.
For experiments at higher sample temperatures, it would be necessary to combine the
Dewar with an air or liquid flow at magnet temperature between the magnet and the
Dewar. While the prototype spectrometer offers this possibility, it was not implemented
yet as it would be costly in terms of instrument footprint and heat load when it comes
to explosion proof housing of the instrument.
The spectrometer settings were unless otherwise noted: single scan mode, 90° pulse
angle, 15 s polarization time before each measurement set by control unit, 12 ppm
spectral width, 8k data points, internal proton lock. The internal proton lock is a
This article is protected by copyright. All rights reserved.

software based lock system which uses the highest signal in the spectrum, which
normally is represented by the used solvent, and sets it to a fixed value. In the present
case the highest signal is that of the alkyl chains which was set to 4.8 ppm in the raw
spectra. During processing, it was recalibrated to 1.4 ppm to get chemically
reasonable spectra.
The complete setup is shown in figure 4. The reaction mixture is sampled from the
vessel via a Teflon tube of 2 mm internal diameter into the membrane pump. All
transfer lines to the NMR were heated or isolated to avoid TPPO precipation. The
membrane pump is fully made of Teflon to avoid corrosion of sample contacting parts.
The volumetric flow of the sample transfer was set to 8 ml/min. This leads to a delay
time of roughly 45 s since the dead volume between sampling point and active volume
inside the NMR is about 5.5 ml. All measurements were done in stopped-flow mode
by bypassing the flowcell with the 4-port valve. After flushing the flowcell of about 1
min with new sample, the control unit switches the valve to bypass and sets a waiting
period of 15 s for polarization. Then it starts the NMR experiment and afterwards
transfers the raw spectra to the watch-dog folder. NMR spectra processing and
integration is then done automatically by the InsightMR process control package.
Quantification results were transferred to the Excel file which then can be read by the
control unit for further processing.
Experimental
The practical application was tested with a standard reaction of process chemistry, the
chlorination of alcohols into the respective alkyl chlorides. While the large-scale
chlorination uses phosgene as chlorination agent, in the present study thionyl chloride
This article is protected by copyright. All rights reserved.

was used instead. Thionyl chloride is less harmful, requires less safety precautions
and the reaction is well documented by BASF patents [23]. The reaction scheme
according to standard text books of organic chemistry [24] is shown in figure 5.
1
-octanol, 1-chlorooctane, triphenylphosphinoxide (TPPO) and thionyl chloride were
purchased from Sigma-Aldrich (Germany), glassware was standard material from in-
house stock supply. A motor stirrer on top of the reaction vessel was used (IKA
Eurostar 20, Germany).
The chlorination reaction was executed at 100°C reaction temperature. 0.69 mol
(about 108 ml) of 1-octanol and either 0.014 mol of TPPO catalyst or not were placed
in the reaction vessel, stirred, and heated up to 100°C. Then, the process-NMR setup
was started with a data acquisition rate of 40 spectra per hour. After an initial period
of about 10 min, 0.69 mol of thionyl chloride were added dropwise at about 1 ml/min
(
without catalyst) or at about 0.4 ml/min (with catalyst). All spectra were automatically
processed and integrated directly after their individual acquisition without human
interaction. Peak tracking of the strongly shifting hydroxyl proton worked without
malfunctions
The hydroxyl group is substituted by a chlorine atom and equimolar amounts of sulfur
dioxide and hydrochloric acid will be released. TPPO can be used as a catalyst to
speed up the reaction strongly [23]. HCl and SO generation rises the practical
2
requirements to the system setup in a way that all sample contacting parts have to be
chemically stable. TPPO as a catalyst introduces another obstacle: during reaction the
polarity of the solvent changes and TPPO might precipitate at room temperature and
block the flow cell. Thus, all sampling lines were heated and the temperature of the
This article is protected by copyright. All rights reserved.

incoming liquid into the flow probe was 45°C, therefore above the precipitation limit of
roughly 35°C.
Prior to conducting the online reaction monitoring a test of the system with non-
reactive mixtures of 1-octanol and 1-chlorooctane reference materials was performed.
The molar fraction of reactant and product was shifted between 0 (not present) and 1
(neat material) and the quantification was done via calculation of the signal area ratio
of the hydroxyl proton of 1-octanol versus the adjacent methylene group protons of 1-
octanol and 1-chlorooctane. The results are given in figure 6. The regression data
show a very good concordance between theoretical and experimental values, the
slope is nearly one and the y-intercept close to zero. Additionally, the hydroxyl proton
signal was proven to be suitable for quantification. The reaction is conducted without
any organic solvent and therefore coming close to industrial conditions. Thus, under
the present conditions the hydroxyl protons lack partners for exchanging protons and
their signals can be used for quantification as any other proton signal. Repeatability
was tested by consecutive analysis of one sample of mean concentration and it
revealed a relative standard deviation of about 1.5 % for integration results.
The NMR spectra of the reaction are shown in figure 7, their quantitative evaluation in
figure 8. Again, the molar ratio is given for each substance. 1-Octanol is quantified via
the hydroxyl proton signal area, the other compounds by the signal area of the
adjacent methylene group protons. All data were weighed by the number of protons.
Reaction time starts upon addition of the first drop of thionyl chloride.
This article is protected by copyright. All rights reserved.

Discussion
Online NMR data reveal that the reaction path appears to be more complicated than
described in the standard text books. Literature study of former work on this type of
reaction offers deep insight into the reaction path. It goes via different steps with two
intermediates, the octyl chlorosulfinate and dioctyl sulfite [25, 26, 27].
The combination of historic literature data, chemical knowledge and present online
NMR information indicates a reaction mechanism following the path as described in
figure 9: in a first reaction step, one molecule of 1-octanol 1 reacts with one molecule
thionyl chloride 2 by loss of one molecule HCl to octyl chlorosulfinate 3. This compound
is described as highly reactive [25] and directly reacts in presence of another molecule
of 1 to the first stable intermediate dioctyl sulfite 4. Intermediate 3 is therefore at that
point in the reaction path not accessible by present NMR spectroscopy. Intermediate
4
is quite stable and has a high activation energy for further conversion with 2 into
octyl chlorosulfinate 3 again. This conversion starts at degrees above 80°C without
catalyst and above 60°C with TPPO catalyst [23].
Present online NMR data support these assumptions. In both experiments, with and
without TPPO catalyst, a formation of octyl chlorosulfinate 3 can only be seen at
concentrations of free 1-octanol approaching zero. In figure 8 a), without TPPO,
reaction progress is only observed after full addition of thionyl chloride 2 and nearly
complete conversion into dioctyl sulfite. Only the excess of 2 and absence of 1 lead to
the formation of 3. The same holds for the reaction with catalyst added.
A similar behavior is found for the final SO elimination reaction which converts the
2
octyl chlorosulfinate 3 into the final product 1-chlorooctane 5. Without catalyst this
reaction step’s kinetics is quite slow leading to significant amount of free octyl
This article is protected by copyright. All rights reserved.

chlorosulfinate in solution and a slow product generation. An opposite behavior is
found when TPPO is added: the last reaction step is strongly accelerated and nearly
no free 3 is found in the reaction solution (see figure 8 b).
The NMR data indicate the benefit of the catalyst TPPO [23]. It strongly accelerates
the cleaving and elimination reactions in the present reaction scheme. The addition of
TPPO might decrease this offset temperature [23] or strongly increases the overall
reaction kinetics when operating at constant temperature. In the reaction without
catalyst, the product formation takes hours and finally ends at 50 % conversion, maybe
due to thionyl chloride undergoing side reactions (figure 8 a) while significant amounts
of both intermediates remain in solution. However, when having TPPO catalyst added,
the cleavage reaction and the elimination of SO are highly accelerated and, as
2
consequence, the practical concentration of octyl chlorosulfinate remains quite low
due to the strong enhancement of the elimination reaction’s kinetics. Fast kinetics
accompanied by much less side reactions let the product yield approach nearly 100 %
in a reasonable amount of time of about 5 hours. The results show that the presented
NMR analysis method enables deep insights into the reaction mechanism (among
others the selectivity to target products, formation of undesired side products) and the
effect of a catalyst on the reaction. Thereby, essential information for the optimization
of the process e.g. by choosing suitable catalysts, solvents or process conditions can
be obtained faster.
This article is protected by copyright. All rights reserved.

Conclusion
Lab-based compact NMR spectroscopy has been successfully converted into a fully
automated spectroscopic system for reaction monitoring. Former tedious work by hand
in NMR spectroscopy like parameter setting from default, data handling between
different software and computer systems as well as spectra processing has been
automated and can now be run with highly reduced human interaction. The use of flow
cells with improved thermal insulation in combination with the compact NMR
spectrometers’ characteristics allow the investigation of reactions in their neat
solutions without the need of deuterated solvents and without initial calibration. The
results show that even with this comparatively simple set-up valuable information on
a complex reaction network with several intermediates and with the formation of
gaseous species in the liquid phase can be gained. The reduced complexity of the
set-up and the analysis allows their use in standard research laboratories or even in
laboratories of production plants. The method facilitates fast insights into important
key parameters of a complex reaction network, such as information on yield and
selectivity of target components, the occurrence of unwanted side products and
estimation of time constants of reagents. Thus, the method is especially suited for fast
screening studies (e.g. catalysts, solvents, process parameters) in an early stage of
process development or for fast quality control in production plants. If desired the
accuracy of the method used to gain quantitative information of the acquired NMR
spectra can be further improved for example by employing chemometrics or other
advanced analysis methods. In this case, the presented method can be applied for
development and validation of for example kinetic or thermodynamic models (see e.g.
[
15,16]).
This article is protected by copyright. All rights reserved.

Thus, compact NMR spectroscopy opens the way for NMR to be used in various
laboratories and process environments in which NMR with high-field instruments is
inherently prohibited. Compact NMR spectroscopy offers manifold possible
applications in process and chemical industries as it enables to get swiftly valuable
information on complex syntheses.
,
This article is protected by copyright. All rights reserved.

Figure 1 Sketch of the upper part of the prototype NMReady with air sleeve and
removable Dewar
This article is protected by copyright. All rights reserved.

Figure 2 Water line width as determined by the by internal shimming of the NMR
instruments as a function of time after 1 min exposure to an octanol sample
preheated to 100°C. Square symbols: standard instrument, round symbols: prototype
with Dewar
This article is protected by copyright. All rights reserved.

Figure 3 a) Stacked spectra of DMSO at different continuous flow temperature and
b) line width of DMSO peak at different continuous flow temperature
This article is protected by copyright. All rights reserved.

Figure 4 a) Schematic drawing of the process NMR system. Sampling system and the
NMR spectrometer must be installed in vicinity to the respective process, control and
spectra processing computers are connected via Ethernet and are stored in a
watchdog folder. b) Photograph of the test assembly in a hood and photograph of the
prototype NMR probe with Dewar for heated samples.
This article is protected by copyright. All rights reserved.

Figure 5 Reaction scheme for the model reaction studied in this paper
This article is protected by copyright. All rights reserved.

Figure 6 Quantitative evaluation of unreactive mixtures of reagent (1-octanol) and
product (1-chlorooctane) of the chlorination reaction. The molar fraction of 1-octanol
by found amount versus weighed in amount is given. Hydroxyl proton and CH
2
-Cl or
CH -O protons are used for quantification.
2
This article is protected by copyright. All rights reserved.

a)
b)
Figure 7 H-NMR spectra plot at 60 MHz of the nucleophilic chlorination reaction at
1
1
00°C reaction temperature a) without TPPO catalyst and b) with 2 mol-% TPPO
catalyst added. The NMR signal attribution was done on basis of the pure spectra
and based on the reaction mechanism (3.5 ppm: CH -X methylene group of 1-
octanol and 1-chlorooctane, 4.0 ppm: CH -O methylene group of octyl disulfite, 4.5:
2
2
2
CH -O methylene group of octyl chlorosulfinate, 7.5 ppm: TPPO).
This article is protected by copyright. All rights reserved.

Figure 8 Quantitative NMR data by plotting the molar fraction of each compound
against reaction time at 100°C reaction temperature a) without TPPO catalyst and b)
with 2 mole-% TPPO catalyst added.
This article is protected by copyright. All rights reserved.

Figure 9 Full reaction scheme of the nucleophilic substitution reaction as suggested
by the NMR results in this work
This article is protected by copyright. All rights reserved.

A prototype compact NMR spectrometer equiped with a Dewar for long-time measurements on
samples with elevated temperature and a flow cell was validated for application in reaction
monitoring and a nucleophilic chlorination reaction was studied as an example. Furthermore,
some general considerations on the possible future role of compact NMR in chemical industry
are made.
This article is protected by copyright. All rights reserved.

References
1
B. Blümich, S. Haber-Pohlmeier, Z. Wasif, Compact NMR, de Gruyter, Berlin, 2014.
R.W. Kessler, Prozessanalytik, Wiley-VCH, Weinheim, 2012.
E. Bayer, K. Albert, M. Nieder, E. Grom, Online coupling of high performance liquid chromatography
2
3
and nuclear magentic resonance, J. Chrom., 186 (1979) 497-507.
4
B. Behnke, G. Schlotterbeck, U. Tallarek, S. Strohschein, L.H. Tseng, T. Keller, K. Albert, E. Bayer,
Capillary HPLC-NMR coupling high-resolution NMR spectroscopy in the nanoliter scale, Anal. Chem.,
6
8 (1996) 1110-1115
5
K. Pusecker, J. Schewitz, P. Gfrörer, L.H. Tseng, K. Albert, E. Bayer, On line coupling of capillary
electrochromatography, capillary electrophoresis, and capillary HPLC with nuclear magnetic
resonance spectroscopy, Anal. Chem., 70 (1998) 3280-3285,
6
A.M. Wolters, D.A. Jayawickrama, A.G. Webb, J.W. Sweedler, NMR detection with multiple solenoidal
microcoils for continuous-flow capillary electrophoresis, Anal. Chem., 74 (2002) 5550-5555.
7
M. Maiwald, H.H. Fischer, Y.K. Kim, K. Albert, H. Hasse, Quantitative high-resolution on-line NMR
spectroscopy in reaction and process monitoring, J. Magn. Res., 166 (2004) 135-146.
8
I.M. Clegg, C.M. Gordon, D.S. Smith, R. Alzaga, A. Codina, NMR reaction monitoring during the
development of an active pharmaceutical ingredient, Anal. Meth., 4 (2012) 1498-1506.
9
D.A. Foley, J. Wang, B. Maranzano, M.T. Zell, B.L. Marquez, Y. Xiang, G.L. Reid, Online NMR and
HPLC as a Reaction Monitoring Platform for Pharmaceutical Process Development, Anal. Chem., 85
(
2013) 8928-8932.
1
0
A. Scheithauer, A. Braecher, T. Gruetzner, D. Zollinger, W.R. Thiel, E. von Harbou, H. Hasse, Online
H-1 NMR Spectroscopic Study of the Reaction Kinetics in Mixtures of Acetaldehyde and Water Using
a New Microreactor Probe Head, Ind. Eng. Chem. Res., 53 (2014) 17589-17596.
1
1
S.K. Küster, E. Danieli, B. Blümich, F. Casanova, High-resolution NMR spectroscopy under the
fume hood, PhysChemChemPhys, 13 (2011) 13172-13176
F. Dalitz, L. Kreckel, M. Maiwald, G. Guthausen, Quantitative Medium-Resolution NMR
12
Spectroscopy Under Non-Equilibrium Conditions, Studied on the Example of an Esterification
Reaction, Appl. Magn. Res., 45 (2014) 411-425
13
M. Goldbach, E. Danieli, J. Perlo, B. Kaptein, V.M. Litvinov, B. Blümich, F. Casanova, A.L.L.
Duchateau, Preparation of Grignard reagents from magnesium metal under continuous flow
conditions and on-line monitoring by NMR spectroscopy, Tetrahedron Lett., 57 (2016) 122-125.
1
4
A. Michalik-Onichimowska, S. Kern, J. Riedel, U. Panne, R. King, M. Maiwald, "Click" analytics for
click" chemistry - A simple method for calibration-free evaluation of online NMR spectra, J. Magn.
"
Res., 277 (2017) 154-161.
1
5
A. Friebel, A. Froscher, K. Münnemann, E. von Harbou, H. Hasse, In situ measurement of liquid-
liquid equilibria by medium field nuclear magnetic resonance, Fluid Phase Equilib., 438 (2017) 44-52
A. Friebel, E. von Harbou, K. Mꢀnnemann, H. Hasse. Reaction monitoring by Benchtop NMR
16
spectroscopy using a novel stationary flow reactor setup. Industrial & Engineering Chemistry
Research, 58 (2019), 18125-18133
17
A. Friebel, E. von Harbou, K. Münnemann, H. Hasse. Online process monitoring of a batch
distillation by medium field NMR spectroscopy. Chemical Engineering Science, 115561, (2020), DOI:
1
0.1016/j.ces.2020.115561.
S. Kern, L. Wander, K. Meyer, S. Guhl, A. Mukkula, M. Holtkamp, M. Salge, C. Fleischer, N. Weber,
1
8
R. King, S Engell, A. Paul, M. Pereira Remelhe, M. Maiwald: Flexible automation with compact NMR
spectroscopy for continuous production of pharmaceuticals. Anal Bioanal Chem 411, 3037–3046
(
2019). https://doi.org/10.1007/s00216-019-01752-y
S. Kern, K. Meyer, S. Guhl, P. Gräßer, A. Paul, R. King, M. Maiwald: Online low-field NMR
1
9
spectroscopy for process control of an industrial lithiation reaction—automated data analysis. Anal
Bioanal Chem 410, 3349–3360 (2018). https://doi.org/10.1007/s00216-018-1020-z
2
0
B. Marquez, M. Fey, K. Colson, R. Krull, E. Bez, D. Piroli, W.E. Maas, 2012 US. Patent
0120092013
D. A. Foley, E. Bez, A. Codina,K. L. Colson, M. Fey, R. Krull, D. Piroli, M. T. Zell; B. L. Marquez
2
21
NMR Flow Tube for Online NMR Reaction Monitoring, Anal. Chem. 2014, 86, 24, 12008-12013;
https://doi.org/10.1021/ac502300q
22
https://www.bruker.com/products/mr/nmr/nmr-software/nmr-software/insightmr/overview.html
This article is protected by copyright. All rights reserved.

2
2
2
3
4
5
T. Rohde, O. Huttenloch, F. Oswald, K. Wissel, B. AG, 2007, DE 10 2005 043 141 A1
K. Schwetlick, Organikum, Wiley-VCH, Weinheim, 2015.
W.E. Bissinger, F.E. Kung, A study of the reaction of alcohols with thionyl chloride, JACS, 69 (1947)
158-2163.
2
26
L. Carius, Ueber die Chloride des Schwefels und deren Derivate, Liebigs Ann. Chem., 111 (1859)
3-113.
9
27
A. Stähler, E. Schirm, The question of the existence of Chlorsulfinester, Ber. Dt. Chem. Gesell., 44
(
1911) 319-323
This article is protected by copyright. All rights reserved.