2412
M. O’Brien et al. / Tetrahedron Letters 58 (2017) 2409–2413
autosampler has placed the needle in the correct vial of starting
material solution, syringe pump 1, which is partly filled with
dichloromethane, takes in the desired volume of liquid (5 mL). Of
the 5 mL of substrate solution which is drawn up from the vial,
4 mL enters the 8 mL holding loop, partly filling it.
The tubing between the autosampler and 3-way-valve 1, which
has a volume of 1 mL, holds the remaining solution; this is flushed
out at the end of the cycle before moving to the next starting mate-
rial. By using a holding loop with a volume in excess of the volume
of substrate being used, syringe pump 1 is kept free of substrate.
Once the holding loop has been loaded with substrate solution,
syringe pump 2 takes in the desired quantity of the pTsOH solution
(7 mL).
As the pTsOH reagent is common to every reaction, there is no
need to use a holding loop to prevent cross contamination and the
solution is able to enter the syringe which has been primed by
partly filling with the pTsOH solution. The 3-way valves are now
switched so that the holding loop is connected upstream to the pis-
ton pump and downstream to T-junction T-1, whilst syringe pump
2 is also connected to T-1. Syringe pump 2 then begins to dispense
the pTsOH solution (at 0.5 mL minÀ1). After 30 s, piston pump 1
begins pumping DCM (also at 0.5 mL minÀ1). The 4 mL plug of
starting material in the holding loop meets the stream of pTsOH
at T-1 before the combined reaction stream enters the reaction
loop. At a combined flow rate of 1.0 mL minÀ1, the volume of the
reaction loop (20 mL) corresponds to a reaction time of 20 min.
Syringe pump 2 dispenses 7 mL of the pTsOH solution, com-
mencing 30 s before piston-pump 1 begins pumping the substrate
solution out of the holding loop. This results in a 0.25 mL ‘overlap’
of the pTsOH solution before the front of the substrate plug and a
2.75 mL overlap at the tail, ensuring that the substrate is always
accompanied by the pTsOH solution. Once syringe pump 2 has dis-
pensed 7 mL of the pTsOH solution, it stops and the speed of piston
pump 1 is increased to 1.0 mL minÀ1, maintaining a 1.0 mL minÀ1
overall flow rate through the reaction loop. Piston pump 1 contin-
ues to pump DCM through the system for a further 60 min, during
which time the organic outlet stream is collected. After this time,
piston pump 1 stops pumping, the valves are switched to their ini-
tial positions and the autosampler moves the needle to the waste
position. Syringe pump 1 then dispenses 5 mL of DCM, returning
it to its original state and also flushing the residual substrate solu-
tion from the tubing between the autosampler and 3-way-valve 1.
The autosampler then moves the needle to the next substrate vial
and the process begins again. An animated schematic diagram,
illustrating the operation of the system upstream of the liquid-liq-
uid extraction step is included in the ESI. It should be noted that, in
this configuration, the product collection flasks were changed
manually. We are currently working on upgrading the system so
that the autosampler also performs the switching between collec-
tion vessels. A total of 10 different tert-butyldimethylsilyl ethers
(Scheme 1) were deprotected in a single automated run. The free
hydroxyl products were isolated in excellent yield and high purity
simply by removal of the solvent and silicon containing by-prod-
ucts under reduced pressure. No cross contamination was
observed between the products. The 1H, 13C and DEPT-135 NMR
spectra of all silyl ethers and deprotection products are provided
in the ESI.
Conclusions
A homemade 3-axis autosampler, constructed using inexpen-
sive and readily available materials, was used in the automated
flow chemical deprotection of a series of tert-butyldimethylsilyl
ethers. The hydroxyl products were formed in high yields and with
excellent levels of purity due to the incorporation of a computer
vision controlled liquid-liquid extraction step. The control script
(the source-code of which is provided in the ESI) was written using
a number of freely available open-source software components
(e.g. Python, OpenCV, PySerial). We are currently investigating
the use of this system in a range of synthetic chemistry applica-
tions. We are also aiming to improve and expand upon the func-
tionality of the system and will report our findings in due course.
Acknowledgements
We gratefully acknowledge funding from the Royal Society of
Chemistry (Research Fund 2015) and support from the European
Union Erasmus+ Study-Abroad programme (LK). We are grateful
to the technical staff of the School of Chemical and Physical
Sciences at Keele University.
A. Supplementary data
Supplementary data associated with this article can be found, in
Scheme 1. Results of automated flow desilylation.