The continuous-flow synthesis of carbazate hydrazones using a simplified computer-vision controlled liquid–liquid extraction system

Article abstract of DOI:10.1016/j.tetlet.2016.10.018

A computer-vision controlled liquid–liquid extraction system was used in the continuous-flow synthesis of a series of carbazate hydrazones. The system uses open-source software components (Python, OpenCV) and is simpler and potentially more economical, in terms of hardware, than one we have described previously.

Full text of DOI:10.1016/j.tetlet.2016.10.018

Tetrahedron Letters 57 (2016) 5188–5191
Contents lists available at ScienceDirect
Tetrahedron Letters
journal homepage: www.elsevier.com/locate/tetlet
The continuous-flow synthesis of carbazate hydrazones using a
simplified computer-vision controlled liquid–liquid extraction
system
⇑
Matthew O’Brien , Dennis A. Cooper, Panashe Mhembere
Lennard-Jones Laboratories, Keele University, Keele, Borough of Newcastle-under-Lyme, Staffordshire ST5 5BG, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
A computer-vision controlled liquid–liquid extraction system was used in the continuous-flow synthesis
of a series of carbazate hydrazones. The system uses open-source software components (Python, OpenCV)
and is simpler and potentially more economical, in terms of hardware, than one we have described
previously.
Received 9 September 2016
Revised 2 October 2016
Accepted 7 October 2016
Available online 10 October 2016
Ó 2016 Published by Elsevier Ltd.
Keywords:
Continuous-flow
Computer-vision
Open-source
Liquid–liquid extraction
Hydrazone
In recent years, the emergence of continuous flow methodology
has created new opportunities for chemical synthesis.1 Compared
with traditional batch processes, flow methods can often offer sig-
nificant safety benefits, particularly for transformations involving
hazardous conditions or reagents.2 Additionally, the small dimen-
sional scales involved lead to the efficient and scale-invariant
interfacial transfer of energy and matter.3 A particularly attractive
aspect of flow chemistry is the ability to incorporate inline purifi-
cation stages. Solid-supported scavengers4 and phase-switching
protocols have been extremely successful in this regard.5 However,
solid-supported chemicals can often be much more expensive than
their solution phase counterparts.6 In addition, they often give rise
to significant and scale-dependent dispersion effects7 and become
depleted over time, thus requiring replacement or regeneration.
This can be a time consuming operation which usually necessitates
halting of the flow process. As liquids can be continuously pumped
through the system, inline liquid–liquid phase separation does not
suffer from this problem and, whilst dispersion cannot be
eliminated, it can be controlled and rendered scale invariant.
One general method of inline liquid–liquid separation used in
continuous flow makes use of the selective wetting of certain
materials, particularly expanded porous PTFE membranes, to
separate aqueous and organic solutions.8 We have been interested,
however, in gravity-based separations of immiscible liquids based
on their densities.9 This is essentially a continuous flow adaptation
of the classical separating funnel. The basic concept is shown in
Figure 1.
A biphasic stream of immiscible liquids with differing densities
will, when passed into a suitable vessel, separate vertically. The
dense phase will exit the vessel through a lower exit and the light
phase will exit the vessel through an upper exit. In Figure 1 the
organic phase is the dense phase and the aqueous extractant is
the light phase, although these roles could be switched if less dense
organic solvents, e.g. diethyl ether, were required. Obviously, it is
important to maintain the interface within the separation vessel,
otherwise liquids may leave through the wrong outlets. This
⇑
Corresponding author.
E-mail address: m.obrien@keele.ac.uk (M. O’Brien).
Figure 1. General schematic for inline separation.
http://dx.doi.org/10.1016/j.tetlet.2016.10.018
0040-4039/Ó 2016 Published by Elsevier Ltd.

M. O’Brien et al. / Tetrahedron Letters 57 (2016) 5188–5191
5189
requires some kind of positional feedback system. Although several
methods for determining the position of the interface could be
used (including refractive index10 and electrical impedance),11 in
the growing context of camera-enabled synthesis technologies12
we have focused on the use of computer-vision systems, in
conjunction with inexpensive and readily available web-cams, to
locate a coloured interfacial float. In our previously described sys-
tem, the control computer used the determined interfacial position
to dynamically adjust the flow rate of the extractant-out pump
(Fig. 2A). Although we have used this successfully in a number of
continuous flow reactions, we sought to develop a streamlined
and more cost-efficient system by reducing the overall number
of pumps involved.
stepper motor and required ancillary components can be obtained
for under £50.
For this work we used a simple luer PTFE stopcock which was
actuated using a stepper motor (for technical details and images,
see the ESI).
We were interested in applying this system to the formation of
carbazate hydrazones13 3a–j using a condensation reaction, catal-
ysed by pyridinium toluenesulfonate (PPTS), between tert-butyl
carbazate 1 and a range of aromatic aldehydes 2a–j (Scheme 1).
In this case, excess carbazate was used to ensure reactions went
to completion in a timely manner. Unreacted carbazate material,
due to its basicity, could be extracted from the product stream,
along with the PPTS catalyst, using an inline liquid–liquid extrac-
tion with aqueous phosphoric acid.
In situations where a smoothly varying response to interface
level perturbations is not necessary, the extractant-out pump
could be replaced by a simple motor actuated valve providing a
binary on/off response. If the open valve provides the path of least
resistance to the liquid (which it will if there is any back-pressure
downstream of other outlets) the extractant phase will leave
through the valve. This is shown schematically in Figure 2B. Thus,
if the extractant (i.e. aqueous phase) has a higher volumetric flow
rate than the organic phase, the position of the interface will fall. If
it falls below a lower bound, the valve can be opened, allowing
liquid to leave through the upper (extractant) outlet. When the
interface level reaches an upper bound, the valve can be closed
again which will force liquid to exit via the lower (product) outlet.
Thus, a hysteresis pattern will result, with the interface position
oscillating between the two boundary points. The check-valve
prevents any of the product stream siphoning back in when the
aqueous-out valve is open. As this system uses one less pump than
that shown in Figure 2A, it represents a significant economic
benefit. HPLC pumps typically cost upwards of £1000, whereas a
The flow setup is shown in Figure 3. Solutions of carbazate and
aldehyde/PPTS (cat.) are introduced into the solvent flow streams
(CH2Cl2) via separate injection loops. The carbazate solution was
injected into the flow stream 30 s prior to the aldehyde/PPTS solu-
tion to ensure that the aldehyde was always accompanied by the
carbazate. This corresponds to an overlap of 30 s at the injection
front and 2 min at the tail. These meet at a T-piece and pass into
a residence-loop where they react at room temperature. The reac-
tion flow stream is then met by a flow stream of aqueous phospho-
ric acid. The phases are efficiently mixed together by passing
through a magnetically stirred mixer (several tiny magnetic stirrer
bars inside an omnifit column on the plate of a stirrer/hotplate).14
On exiting the mixer the phases rapidly revert to plug-flow. The
i) PPTS (cat.)
H
N
H
N
t
O
DCM, r.t.
OtBu
OBu
H N
2
R
N
+
R
ii) H3PO4 (aq).
(Figure 3)
O
O
2a-j
3a-j
1
H
N
H
N
OBu
OtBu
t
N
N
O
O
Me
MeO
3b 91%
3a 95%
H
N
H
N
t
OBu
t
OBu
N
N
O
O
Cl
OH
3d 90%
3c 82%
H
N
H
N
t
t
OBu
OBu
N
N
O
O
3f 94%
3e 95%
H
N
t
H
OBu
t
N
OBu
N
N
O
O
NC
F
3h 91%
OBn
3g 87%
H
N
H
N
t
OBu
t
OBu
N
N
O
O
NO2
Cl
3j 93%
3i 89%
Figure 2. (A) HPLC pump regulated extraction system. (B) Motor driven valve
regulated extraction system.
Scheme 1. Results for the synthesis of carbazate hydrazones 3a–j.

5190
M. O’Brien et al. / Tetrahedron Letters 57 (2016) 5188–5191
Figure 4. Portion of the 1H NMR spectra of (A) product following extraction.
(B) Product bypassing extraction.
the same flow setup was used for all these reactions with no dis-
mantling or cleaning and no cross-contamination of the products
was observed.
In conclusion, a simplified computer-vision controlled inline
liquid–liquid extraction system has been developed for use in
continuous flow synthesis. Requiring fewer pump units than our
previously reported system, this new configuration may provide
significant cost benefits in cases where uniform flow of the outlet
stream is not required. The system was used in the synthesis of a
series of carbazate hydrazones.
The excess carbazate and PPTS catalyst were efficiently
extracted by an aqueous phosphoric acid stream affording the
products in excellent yield and high purity. We are currently work-
ing on the application of this system to a range of other reactions.
We are also investigating the process properties of the system,
including quantification of dispersion effects, and will report our
findings in due course.
Supplementary data
Figure 3. Flow apparatus for the hydrazone formation.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.tetlet.2016.10.
018.
biphasic flow stream then enters the separating vessel where the
two phases settle under gravity according to density. A plastic
float, whose density is between that of the two liquid phases,9a is
placed in the vessel to locate the vertical position of the liquid–
liquid interface using the webcam.15
The webcam stream is fed to the control computer running a
Python script incorporating several open-source software compo-
nents. In addition to the critical OpenCV computer vision library,16
other key components included NumPy17 and PySerial18
(for details of the scripts see the ESI).
References and notes
1. For reviews see: (a) Hessel, V.; Kralisch, D.; Kockmann, N.; Noel, T.; Wang, Q.
ChemSusChem 2013, 6, 746–789; (b) Gutmann, B.; Cantillo, D.; Kappe, C. O.
Angew. Chem., Int. Ed. 2015, 54, 6688–6728; (c) McQuade, D. T.; Seeberger, P. H.
J. Org. Chem. 2013, 78, 6384–6389; (d) Wiles, C.; Watts, P. Chem. Commun. 2011,
47, 6512–6535; (e) Yoshida, J. I. Chem. Rec. 2010, 10, 332–341; (f) Ley, S. V.
Chem. Rec. 2012, 12, 378–390; (g) Wegner, J.; Ceylan, S.; Kirschning, A. Adv.
Synth. Catal. 2012, 354, 17–57; (h) Webb, D.; Jamison, T. F. Chem. Sci. 2010, 1,
675–680; (i) Pastre, J. C.; Browne, D. L.; Ley, S. V. Chem. Soc. Rev. 2013, 42, 8849–
8869; (j) Anderson, N. G. Org. Proc. Res. Dev. 2012, 16, 852–869; (k) Newman, S.
G.; Jensen, K. F. Green Chem. 2013, 15, 1456–1472; (l)Microreactors in Organic
Chemistry and Catalysis; Wirth, T., Ed., 2nd ed.; Wiley: Hoboken, 2013.
2. (a) Razzaq, T.; Kappe, C. O. Chem.-Asian J. 2010, 5, 1274–1289; (b) Gross, U.;
Koos, P.; O’Brien, M.; Polyzos, A.; Ley, S. V. Eur. J. Org. Chem. 2014, 6418–6430;
(c) Mueller, S. T. R.; Wirth, T. ChemSusChem 2015, 8, 245–250; (d) Maurya, R. A.;
Park, C. P.; Lee, J. H.; Kim, D. P. Angew. Chem., Int. Ed. 2011, 50, 5952–5955; (e)
Delville, M. M. E.; van Hest, J. C. M.; Rutjes, F. P. J. T. Beilstein J. Org. Chem. 2013,
9, 1813–1818; (f) Dallinger, D.; Pinho, V. D.; Gutmann, B.; Kappe, C. O. J. Org.
Chem. 2016, 81, 5814–5823.
3. (a) O’Brien, M.; Baxendale, I. R.; Ley, S. V. Org. Lett. 2010, 12, 1596–1598; (b)
Park, C. P.; Kim, D. P. J. Am. Chem. Soc. 2010, 132, 10102–10106; (c) Fukuyama,
T.; Totoki, T.; Ryu, I. Green Chem. 2014, 16, 2042–2050; (d) Jensen, K. F. Chem.
Eng. Sci. 2001, 56, 293–303; (e) Ahmed-Omer, B.; Brandt, J. C.; Wirth, T. Org.
Biomol. Chem. 2007, 5, 733–740; (f) O’Brien, M.; Taylor, N.; Polyzos, A.;
Baxendale, I. R.; Ley, S. V. Chem. Sci. 2011, 2, 1250–1257; (g) Kobayashi, J.; Mori,
Y.; Okamoto, K.; Akiyama, R.; Ueno, M.; Kitamori, T.; Kobayashi, S. Science 2004,
304, 1305–1308; (h) Knowles, J. P.; Elliott, L. D.; Booker-Milburn, K. I. Beilstein J.
Org. Chem. 2012, 8, 2025–2052; (i) Brzozowski, M.; O’Brien, M.; Ley, S. V.;
Polyzos, A. Acc. Chem. Res. 2015, 48, 349–362.
We began our investigation with p-anisaldehyde (2a). A brief
survey of reaction times revealed that
a residence loop of
14.8 mL (corresponding to 7.4 min reaction time) was sufficient
to ensure complete conversion to the hydrazone 3a. The outlet
stream was collected for 40 min and the product isolated in very
high yield simply by removing the solvent under reduced pressure.
Pleasingly, NMR spectroscopy revealed that the PPTS and excess
carbazate reagent had been completely extracted, affording a pure
product. Figure 4 shows a portion of the spectrum of the product
isolated with (A) and without (B) the liquid–liquid purification
stage.
As can be seen in Figure 4B, if the extraction step is bypassed, in
addition to the excess carbazate, the product is also contaminated
with significant residual aromatic material from the PPTS.
Having established appropriate conditions for the hydrazone
formation and extraction, the flow reaction was carried out on a
range of aldehydes 2a–j, in each case affording products 3a–j in
excellent yield and high purity (1H and 13C NMR spectra of all
the products are provided in the ESI). It should be pointed out that
4. Ley, S. V.; Baxendale, I. R.; Bream, R. N.; Jackson, P. S.; Leach, A. G.; Longbottom,
D. A.; Nesi, M.; Scott, J. S.; Storer, R. I.; Taylor, S. J. J. Chem. Soc., Perkin Trans. 1
2000, 3815–4195.
5. O’Brien, M.; Denton, R.; Ley, S. V. Synthesis 2011, 1157–1192.
6. At time of writing: Sigma-Aldrich: Amberlyst 15 Hydrogen form (06423-1 KG)
costs £ 23.50/mol (GB). Sulfuric acid (320501-1L-D) costs £ 1.10/mol.

M. O’Brien et al. / Tetrahedron Letters 57 (2016) 5188–5191
5191
7. Lange, H.; Carter, C. F.; Hopkin, M. D.; Burke, A.; Goode, J. G.; Baxendale, I. R.;
Ley, S. V. Chem. Sci. 2011, 2, 765–769.
8. (a) Kralj, J. G.; Sahoo, H. R.; Jensen, K. F. Lab Chip 2007, 7, 256–263; (b) Castell,
O. K.; Allender, C. J.; Barrow, D. A. Lab Chip 2009, 9, 388–396; (c) Atallah, R. H.;
Ruzicka, J.; Christian, G. D. Anal. Chem. 1987, 59, 2909–2914; (d) Kolehmainen,
E.; Turunen, I. Chem. Eng. Process. 2007, 46, 834–839.
12. (a) Ley, S. V.; Ingham, R. J.; O’Brien, M.; Browne, D. L. Beilstein J. Org. Chem. 2013,
9, 1051–1072; (b) Ingham, R. J.; Battilocchio, C.; Fitzpatrick, D. E.; Sliwinski, E.;
Hawkins, J. M.; Ley, S. V. Angew. Chem., Int. Ed. 2015, 54, 144–148.
13. (a) Einhorn, A. Justus Liebigs Annalen der Chemie 1898, 300, 135–155; (b) Backer,
H. J. Recueil des Travaux Chimiques des Pays-Bas et de la Belgique 1912, 31, 1–29.
14. R. E. Gugger,; S. M. Mozersky, US 4054270 1977.
9. (a) O’Brien, M.; Koos, P.; Browne, D. L.; Ley, S. V. Org. Biomol. Chem. 2012, 10,
7031–7036; (b) Hu, D. X.; O’Brien, M.; Ley, S. V. Org. Lett. 2012, 14, 4246–4249;
(c) O’Brien, M.; Cooper, D. Synlett 2016, 164–168; (d) Bourne, S. L.; O’Brien, M.;
Kasinathan, S.; Koos, P.; Tolstoy, P.; Hu, D. X.; Bates, R. W.; Martin, B.; Schenkel,
B.; Ley, S. V. ChemCatChem 2013, 5, 159–172.
15. In this system, the vertical position of the float is directly related to its vertical
position in the image. For an approach that relates the size of an object’s image
to its distance from the camera see: Wang, T.-H.; Lu, M.-C.; Hsu, C.-C.; Chen,
C.-C.; Tan, J.-D. Measurement 2009, 42, 604–610.
16. (a) Bradski, G. Dr. Dobbs J. 2000, 25, 120–125; (b) http://opencv.org, (last
accessed 6/9/2016).
10. Maslana, E.; Schmitt, R.; Pan, J. J. Autom. Methods Manage. Chem. 2000, 22, 187–
194.
17. http://www.numpy.org, (last accessed 6/9/2016).
11. Sprecher, H.; Payán, M.; Weber, M.; Yilmaz, G.; Wille, G. J. Flow Chem. 2012, 2,
20–23.
18. https://github.com/pyserial/pyserial, (last accessed 6/9/2016).