A Small Footprint, High Capacity Flow Reactor for UV Photochemical

Synthesis on the Kilogram Scale

Luke D. Elliott, Malcolm Berry, Bashir Harji, David Klauber, John Leonard and
Kevin I. Booker-Milburn

Supporting Information

General experimental S2
FEP Reactor pictures S3
Experimental procedures S4
Pictures of products S10
Copies of 1H and 13C NMR spectra S12

S1
General Experimental
For column chromatography, Sigma Aldrich technical grade 60 Å silica gel was used. All NMR data
was collected using a Jeol Eclipse 400 MHz or Varian 400-MR instruments. Data was processed
directly using MestReNova (version 8.1). Reference values for residual solvents were taken as δ =
7.26 (CDCl3) and 2.50 ppm (DMSO-d6) for 1 H NMR; δ = 77.26 ppm (CDCl3) and 39.52 ppm (DMSO-
d6) for 13C NMR. Multiplicities for coupled signals were denoted as: s = singlet, d = doublet, t =
triplet, q = quartet, m = multiplet, br. = broad, app. = apparent and dd = double doublet etc.
Coupling constants (J) are given in Hz and are uncorrected. Where appropriate, COSY, DEPT, HMBC,
HMQC and NOE experiments were carried out to aid assignment. Mass spectroscopy data was
collected was carried out by the University of Bristol mass spectrometry service using Fisons
Autospec instruments. Melting points are uncorrected and were recorded on Stuart Scientific
apparatus.

For all photochemical reactions, solvents were “degassed” by evacuating under vacuum and re-filling
with N2. For the duration of the reaction, N2 was bubbled through the solution. When photochemical
reactions were followed for their duration by NMR, 1,3,5-trimethoxybenzene was used as an
internal standard. In all cases, a stock solution of this was added to aliquots of the reaction mixture
prior to NMR sample preparation. For flow reactions, when the full volume of solution had been
taken up, the reactor was flushed out with a reactor volume of neat solvent and the reactor cleaned
with recirculating DMSO/water prior to irradiating the next volume.

Batch reactors, 125 W and 400 W medium pressure mercury lamps were purchased from
Photochemical Reactors Ltd. Reading. The 5 kW variable power supply (JA5000 VPXi) was purchased
from Jenton International, Whitchurch. A Masterflex L/S PTFE tubing peristaltic pump system (Cole-
Parmer) was used for 400 W flow reactions. For the Firefly reactor, an FMI Q-series valveless piston
pump was used along with an FMI pulse dampener (Cole-Parmer). The reactor was cooled with a
Huber Unichiller 025-MPC, purchased from Radleys, Saffron Walden.

S2
Performance of FEP Tubing with High-Power UV Lamps

1. An example of the precipitation of

insoluble material (product or by-product) in
the FEP tubing during reaction. In this
instance, the reactor did not block and the
2.7 mm i.d. FEP enabled the insoluble
residues to successfully exit the reactor. Use
of a narrower bore of tubing would
significantly increase the chance of blockage
and rupture.

Figure S1: Successful clean of an FEP flow reactor

2. An example of a 400 W FEP flow reactor

(left) completely blocked by build-up of
precipitated material. In some cases
introduction of a more polar solvent can
break up the solids and the reactor can be
cleared and reused. In other cases the
reactor has to be re-wrapped with fresh
FEP.

Figure S2: Fully blocked FEP flow reactor

3. Residue build-up on a 400 W FEP flow

reactor after a reaction run. This did not
block the reactor but did lead to a severe
drop off in UV transmission. In some
instances the coloured residues can be
removed by flushing with DMSO, but here
the surface damage to the FEP was
irreversible. The reactor had to be re-
wrapped with fresh FEP.
Figure S3: Irreversibly damaged FEP flow reactor

S3
Prototype Parallel Tube Flow Reactor (PTFR) Trial Reactions

R'
O O
R'
(1.5 eq.)
NR NR
400 W, Pyrex
O MeCN (0.1 M) O

Figure S4: Prototype PTFR with water-jacketed 400 W Hg lamp. Internal volume = 60 ml

General procedure

A solution of the relevant maleimide (R = H, Me) (40 mmol) and alkyne (R’ = nBu, CH2OH) (60 mmol)
in degassed MeCN (0.1 M) was pumped through the prototype PTFR with a peristatic pump. After
200 ml of solution had been passed through the reactor a sample was taken for analysis by 1H-NMR
using 1,3,5 trimethoxybenzene as an internal standard.

Table S1: Comparison of prototype PTFR with FEP flow reactor

Reactor R/R' Flow Rate [mL/min] Yield [%] Productivity [g/h]

3-layer FEP Me/nBu 6 52 3.62
[1] n
400W PTFR Me/ Bu 6 49 3.41
3-layer FEP H/CH2OH 4 64 2.35
[1]
400W PTFR H/CH2OH 4 58 2.13

S4
Kilogram Scale-Up Examples

THPA / 2-Butene-1,4-diol [2+2]

125 W Batch Reaction

A solution of THPA[2] (9.13 g, 60 mmol), cis-2-butene-1,4-diol (7.4 ml, 90 mmol) and

isopropylthioxanthone (153 mg, 0.6 mmol) in degassed EtOAc (0.4 M) was irradiated with a 125 W
medium pressure mercury lamp in a 150 ml batch reactor equipped with a Pyrex immersion well for
2 hours. The mixture was concentrated in vacuo to a slurry which was filtered, washing with EtOAc
and Et2O and dried to give product 2 as a colourless crystalline solid (11.1 g, 77%).

3 kW Flow Reaction

Representative Procedure: A solution of THPA (608 g, 4 mol), cis-2-butene-1,4-diol (500 ml, 6.1 mol)
and isopropylthioxanthone (10.2 g, 40 mmol) in degassed EtOAc (0.4 M) was irradiated with the
firefly reactor fitted with a Pyrex inner filter and lamp at 3 kW at 36 ml/min. The mixture was
concentrated in vacuo to a slurry which was filtered, washing with EtOAc and Et2O and dried to give
product 2 as a colourless crystalline solid (769 g, 80%): m.p. 188 -189°C (MeOH); δH (400 MHz,
DMSO) 12.25 (1H, br. s, OH), 4.45 (1H, br. s, OH), 4.21 (1H, dd, J = 9.2, 4.8 Hz, CHHO), 4.13 (1H, d, J =
9.2 Hz, CHHO), 3.54 – 3.44 (2H, m, CH2OH), 2.83 (1H, dd, J = 8.5, 4.8 Hz, CH), 2.64 (1H, dt, J = 8.5, 7.2
Hz, CH), 1.94 – 1.25 (8H, m, 4×CH2); δC (100 MHz, DMSO) 179.9 (C), 174.4 (C), 70.3 (CH2), 59.4 (CH2),
48.4 (C), 45.7 (C), 42.7 (CH), 37.3 (CH), 26.1 (CH2), 26.1 (CH2), 21.5 (CH2), 20.9 (CH2)
Table S2: Results for THPA / 2-butene-1,4-diol cycloaddition with Pyrex Firefly reactor at 3 kW, 0.4 M, 36 ml/min

VOLUME MOLS TIME ISOLATED YIELD PRODUCTIVITY YIELD

DROP OVER RUN
2.5 L 1 1.16 hr 192 g 80% 691 mmol/hr 0%

7.5 L 3 3.47 hr 577 g 80% 691 mmol/hr 0%

10 L 4 4.63 hr 769 g 80% 691 mmol/hr 0%

Total 20 L 8 9.26 hr 1,538 g 80% 691 mmol/hr

S5
Cookson’s Dione

125 W Batch Reaction

A solution of Diels Alder adduct 3[3] (2.61 g, 15 mmol) in degassed EtOAc (150 ml) was irradiated with
a pre-warmed 125 W medium pressure mercury lamp in a 150 ml batch reactor equipped with a
Pyrex immersion well for 15 min. The solvent was removed in vacuo and chromatography on silica
(40% EtOAc in hexane to 100% EtOAc) yielded Cookson’s dione 4 as an off-white solid (2.38 g, 91%)

1.5 kW Flow Reaction

A solution of Diels Alder adduct 3 (436 g, 2.5 mol) in degassed EtOAc (0.5 M) was irradiated with the
firefly reactor fitted with a Pyrex inner filter and lamp at 1.5 kW at 36 ml/min. The mixture was
concentrated in vacuo to a slurry to which was added hexane. The mixture was filtered, washing
with petroleum ether and the solid dried to give pure Cookson’s dione 4 as an off-white crystalline
solid (387 g, 89%): m.p. 242 - 243°C; δH (400 MHz, CDCl3) 3.19 – 3.14 (2H, m, 2×CH), 2.95 – 2.90 (2H,
m, 2×CH), 2.82 – 2.78 (2H, m, 2×CH), 2.72 – 2.68 (2H, m, 2×CH), 2.04 (1H, app. d, J = 11.4 Hz, CHH),
1.88 (1H, app. d, J = 11.4 Hz, CHH); δC (100 MHz, CDCl3) 212. 2 (2×C), 54.9 (2×CH), 44.8 (2×CH), 43.9
(2×CH), 40.6 (CH2), 38.9 (2×CH)

3.0 kW Flow Reaction

A solution of Diels Alder adduct 3 (436 g, 2.5 mol) in degassed EtOAc (1.0 M) was irradiated with the
firefly reactor fitted with a Pyrex inner filter and lamp at 3.0 kW at 36 ml/min. Purification, as
described above, gave pure Cookson’s dione 4 as an off-white crystalline solid (385 g, 88%)
Table S3: Intramolecular cycloaddition of 3 with Pyrex Firefly reactor at 1.5 kW (0.5 M), and 3.0 kW (1.0 M), 36 ml/min

LAMP VOLUME MOLS TIME ISOLATED YIELD PRODUCTIVITY YIELD

POWER DROP OVER RUN
1.5 kW 5 L 2.5 2.31 hr 387 g 89% 961 mmol/hr 0%

3.0 kW 2.5 L 2.5 1.16 hr 385 g 88% 1,901 mmol/hr 0%

3.0 kW Flow Reaction: 7.5 mol scale-up example

A solution of Diels Alder adduct 3 (1,307 g, 7.5 mol) in degassed solvent (10% MeCN in EtOAc, 1.0 M)
was irradiated with the firefly reactor fitted with a Pyrex inner filter and lamp at 3.0 kW at 36
ml/min. Purification, as described above, gave pure Cookson’s dione 4 as a colourless crystalline
solid (1,165 g, 89%)

S6
Crossed – [2+2] Cycloaddition

400 W Batch Reaction

A solution of enamine 5[4] (105 g, 400 mmol) and isopropylthioxanthone (1.01 g, 4 mmol) in
degassed MeCN (0.4 M) was irradiated with a 400 W medium pressure mercury lamp through a
Pyrex filter in a 1 L batch reactor. Quantitative NMR showed the initial linear rate to be about 17
mmol/hr. Full consumption of starting material was observed at 24 hrs and the solvent was
evaporated to give a slurry (115 g). The mixture was triturated with hexane and the product filtered,
washing with hexane and cold Et2O to give pyrrolidine 6 as a pale yellow crystalline solid (93.4 g,
89%)

3 kW Flow Reaction

Representative Procedure: To a solution of enamine 5 (421 g, 1.6 mol) and isopropylthioxanthone

(4.1 g, 16 mol) in degassed MeCN (3 L) was added water (380 ml) and further MeCN to make up to 4
L (0.4 M). The solution was irradiated with the firefly reactor fitted with a Pyrex inner filter and lamp
at 3 kW at 9 ml/min. The photolysate was concentrated to a slurry which was triturated with hexane
and filtered, washing with hexane then cold Et2O to give pyrrolidine 6 as a white crystalline granular
solid (367.3 g, 87%): m.p. 101 – 103 °C; δH (400 MHz, CDCl3) 7.63 – 7.03 (10H, m, 10×Ar.H), 3.70 (2H,
br. s, NCH2), 2.93 – 2.89 (1H, m, CH), 2.21 – 2.14 (2H, m, 2×CHH), 2.05 – 1.99 (2H, m, 2×CHH); δC (126
MHz, DMSO, 80°C) 171.8 (C), 139.2 (C), 136.4 (C), 129.6 (CH), 127.5 (2×CH), 127.3 (2×CH), 127.1
(2×CH), 125.9 (CH), 125.7 (2×CH), 73.9 (C), 54.9 (CH2), 42.9 (2×CH2), 34.2 (CH); CI-HRMS m/z
264.1380 (MH+ C18H18NO requires 264.1388)

Table 4 details the results from consecutive 2L (x2) and 4 L (x2) runs. In most cases, no significant
reduction in conversion was observed over the course of each run. For consistency the reactor was
cleaned with DMSO / water in between runs.
Table S4: Results for intramolecular ‘cross’ [2+2] cycloaddition with Pyrex Firefly reactor at 3 kW, 0.4 M, 9 ml/min

VOLUME MOLES TIME ISOLATED YIELD PRODUCTIVITY YIELD

DROP OVER RUN
2L 0.8 3.70 hr 176.2 g 84% 181 mmol/hr 2%

2L 0.8 3.70 hr 177 g 84% 181 mmol/hr 2%

4L 1.6 7.41 hr 361.7 g 86% 186 mmol/hr 0%

4L 1.6 7.41 hr 367.3 g 87% 188 mmol/hr 2%

Total 12 L 4.8 22.22 hr 1,082 g 86% 186 mmol/hr

S7
Sensitizer Screena)

Table S5: Sensitizer screening for intramolecular ‘cross’ [2+2] cycloaddition in 400W batch reactor

Entry Solvent Sensitizer Productivity

(mmol/hr)b)
1 Acetone (0.1 M) - 1.5
2 Acetone (0.1 M) - 8.0 c)
3 MeCN (0.1 M) Xanthone (10%) 5.6
4 MeCN (0.1 M) Xanthone (20%) 6.8
5 Acetone (0.1 M) Xanthone (10%) 5.6
6 MeCN (0.1 M) Benzophenone (10%) 7.6
7 MeCN (0.1 M) Thioxanthone (10%) 19.6
8 MeCN (0.1 M) Thioxanthone (5%) 18.8
9 MeCN (0.1 M) Thioxanthone (3%) 16.8
10 MeCN (0.1 M) Thioxanthone (2%) 15.6
11 MeCN (0.4 M) Isopropylthioxanthone (1%) 17.0
12 EtOAc (0.2 M) Thioxanthone (1%) 8.9
13 MeCN (0.1 M Anthraquinone 2.0
14 MeCN 9-fluorenone -
15 MeCN Dibenzosuberenone -
Footnotes: a) Reactions performed in a 400 ml Pyrex immersion well reactor apart from entry 11 which was a 1
L reactor; b) Productivity determined at incomplete conversion by 1-H NMR using 1,3,5-trimethoxybenzene as
internal standard c) Reaction performed with a quartz immersion well but degradation limited the yield to 50%

The use of an alkylated thioxanthone derivative was significant from a process point of view as
thioxanthone suffers solubility issues. Purification of the product from the sensitizer was easier
when using low loadings of the more soluble isoproplythioxanthone. Diethylthioxanthone also
serves and a convenient alternative.

S8
Photodecarboxylative Cycloaddition

125 W Batch Reaction: A mixture of carboxylic acid 7a[5] (7.0 g, 30 mmol) and K2CO3 (2.07 g, 15
mmol) was sonicated in deionised water (30 ml) until fully dissolved. The solution was diluted to 75
ml with more water then made up to 150 ml with degassed MeCN (0.2 M). The solution was
irradiated with a 125 W medium pressure mercury lamp in a 150 ml batch reactor equipped with a
quartz immersion well for 2 hours. The MeCN was removed in vacuo to give a thick white slurry
which was washed (water), filtered and the aqueous filtrate extracted with CHCl3 (3×25 ml). Into the
combined organic extracts was dissolved the wet residue which was separated, dried (MgSO4),
filtered through Celite and evaporated to give product 8 as white solid (4.6 g, 81%)

3 kW Flow Reaction Representative Procedure: A mixture of carboxylic acid 7a (233 g, 1.0 mol) and
K2CO3 (69 g, 0.5 mol) was sonicated in deionised water (1 L) until fully dissolved. The solution was
diluted to 2.5 L with more water then made up to 5 L with degassed MeCN (0.2 M). The solution was
irradiated with the firefly reactor fitted with a quartz inner filter and lamp at 3 kW at 30 ml/min. The
MeCN was removed in vacuo to give a white slurry which was washed (water), filtered and the
aqueous filtrate extracted with CHCl3 (4×400 ml). Into the combined organic extracts was dissolved
the wet residue which was separated, dried (MgSO4), filtered through Celite and evaporated to give
product 8 as white solid (167 g, 88%) m.p. 128 – 130 °C: δH (400 MHz, DMSO) 7.66 – 7.57 (3H, m,
3×Ar.H), 7.54 – 7.49 (1H, m, Ar.H), 6.35 (1H, d, J = 1.6 Hz, OH), 3.52 (1H, app. dt, J = 11.2, 8.4 Hz,
NCHH), 3.29 (1H, ddd, J = 11.2, 9.2, 2.9 Hz, NCHH), 2.53 – 2.40 (1H, m, CHH), 2.27 – 2.12 (2H, m,
CHH, CHH), 1.49 – 1.38 (1H, m, CHH); δC (100 MHz, DMSO) 169.1 (C), 148.3 (C), 132.5 (CH), 131.5 (C),
129.3 (CH), 122.8 (CH), 122.7 (CH), 95.6 (C), 41.1 (CH2), 35.2 (CH2), 27.4 (CH2)
Table S6: Results for photodecarboxylative cycloaddition with quartz Firefly reactor at 3 kW, 0.2 M, 36 ml/min

VOLUME MOLS TIME ISOLATED YIELD PRODUCTIVITY YIELD

DROP OVER RUN
2L 0.4 1.12 hr 66 g 87% 313 mmol/hr 0%

5L 1 2.78 hr 167 g 88% 317 mmol/hr 0%

5L 1 2.78 hr 147 g 78% 281 mmol/hr 13%*

10 L 2 5.56 hr 317 g 84% 302 mmol/hr 5%

10 L 2 5.56 hr 335 g 89% 320 mmol/hr 0%

Total 32 L 6.4 17.78 hr 1,032 g 85% 306 mmol/hr

* After the previous runs the reactor was flushed briefly with MeOH but this appeared not to have cleaned a small residue
build-up on the quartz tubes and hence this run had a 13% drop in yield over the run. Before the subsequent run 30%
water/DMSO was circulated for ca. 5 min (36 ml/min) and no significant drop in conversion was observed for the next 20 L
of reaction solution.

S9
Pictures of Products

S10
References

[1] L. D. Elliott, J .P. Knowles, P.J. Koovits, K. G. Maskill, M. J. Ralph, G. Lejeune, L. J. Edwards, R.
I. Robinson, I. R. Clemens, B. Cox, D. D. Pascoe, G. Koch, M. Eberle, M. B. Berry, K.I. Booker-Milburn,
Chem. Eur. J. 2014, 20, 15226 – 15232

[2] M. E. Bailey, E. D. Amstutz, J. Am. Chem. Soc. 1956, 78, 3828 – 3830

[3] J. Mao, H. Xie, W. Bao, Org. Lett. 2015, 17, 3678 – 3681

[4] D. W. Piotrowski, Synlett. 1999, 7, 1091 – 1093

[5] E. Guenin, M. Monteil, N. Bouchemal, T. Prange, M. Lecouvey, Eur. J. Org. Chem. 2007, 3380
- 3391

S11
1
H and 13C NMR Spectra

S12
S13
S14
S15