Alternately (eg Liu et. al*.) has reported hydrogenation of phenol at pressure and ... solvent, 10 mol% AlCl3, 5 mol% Pd
Development of heterogeneous catalytic processes Transferring batch to continuous flow and including use of scCO2 as solvent
Dr Jasbir Singh
(
[email protected]) HEL Ltd, England
HEL Limited London, England
HEL Inc NJ, USA
HEL China Beijing
HEL Italia Milan
HEL India Mumbai
HEL Products by Research Area Screening
Low Pressure
High Pressure
Process Dev/Optimization
Low Pressure
High Pressure
Scale up/Pilot
Low Pressure
High Pressure
Process Intensification and Clean Synthesis Intensification -
Increased pressure, especially for gas/liquid chemistry
-
Reduce inventory .. minimise hazard (of pressure/inventory)
Clean Synthesis -
Water contamination, pollution, toxicity
-
Potentially use of scCO2
-
Ideal in conjunction with intensification (ie reduced inventory)
better chemistry - faster
Intensification is obvious if switching is possible
better chemistry - faster
Production of Cyclohexanone
Oxidation
Typically at high Temperature, low conversion and selectivity.
Hydrogenation Alternately (eg Liu et. al*.) has reported hydrogenation of phenol at pressure and low temperature, with good results. Batch process. *H. Liu, T. Jiang, B. Han, S. Liang, Y Zhou, Science 326, 1250 (2009).
better chemistry - faster
Hydrogenation of phenol
First step competes with further hydrogenation to produce the alcohol, reducing selectivity. Liu* looked at batch synthesis (1ml scale) using Lewis acid in DCM with good results * H. Liu, T. Jiang, B. Han, S. Liang, Y Zhou, Science 326, 1250 (2009).
better chemistry - faster
HEL Objectives
- Scale up Liu’s (1ml) synthesis to stirred reactors - Optimise conditions for production of cyclohexanone and cyclohexanol (ability to select product)
-
Transfer synthesis to fixed bed, “trickle flow” synthesis
better chemistry - faster
Optimisation in 8 parallel pressure -reactor platform Temperature (ºC)
40
60
80
100
R1
R2
R3
R4
R5
R6
R7
R8
Pressure (bar)
30
40
Figure 1: Diagram illustrating temperature and pressure variation on the HPChemSCAN (using DCM as
Tests in 8 stirred reactors (HP Chemscan)
solvent, 10 mol% AlCl3, 5 mol% Pd/C)
HP Chemscan reactor
10mole % AlCl3 and 5mole % Pd/C in DCM solvent
Phenol Conversion in HP Chemscan at 30 and 40bar
Phenol conversion versus Temperature 102
Phenol Conversion (%)
100 98 96 94 92 Pressure (30 bar) 90 Pressure (40 bar) 88
Literature value at 30 bar 86 30
40
50
60
70
80
Temperature (ºC)
better chemistry - faster
90
100
110
Product Selectivity Cyclohexanone selectivity Cyclohexanone Selectivity ( %)
120 100 80 60
Pressure (30 bar)
40
Pressure (40 bar)
20
Literature value at 30 bar
0 40
50
60 70 80 Temperature (ºC)
90
100
110 Cyclohexanol Selectivity / %
30
Cyclohexanol Selectivity 70
50
Pressure (40 bar)
40 Literature Value at 30 bar
30 20 10
0 -10
better chemistry - faster
Pressure (30 bar)
60
30
40
50
60
70
80
Temperature (ºC)
90
100
110
Summary of Conversion/selectivity Liu et al.
HEL Ltd.
HEL Ltd.
(30 bar)
(30 bar)
(40 bar)
Phenol conversion
>99.9%
99.4%
99.6%
Max Cyclohexanone select
95.1%
95.4%
94.1%
Max Cyclohexanol select
4.9%
94.6%
95.9%
Values reported with 0.1 equiv. AlCl3 and 0.0.05 equiv. Pd/C, for 120 minutes. Results at different temperatures that gave maximum selectivity/conversion
Stirred Reactors allows choice between ketone and alcohol
better chemistry - faster
Transfer to min- flow reactor (flowCAT reactor system)
better chemistry - faster
Reactor packing suggestion Active catalyst
Feed
Glass (or metal) beads (400 micron dia)
Preheat zone
Catalyst typically 1/5th to 1/10th dia. of reactor (say 100 micron +)
Reaction zone
better chemistry - faster
Transfer DCM Process to Continuous flow - problematic
Phenol hydrogenation using Lewis Acid and DCM solvent proved to be very difficult: -
-
Pump feed was erratic Sample valve became blocked periodically Back-pressure regulator became blocked periodically
Process not ideal for flow operation
better chemistry - faster
Physical properties give clue to difficulties Melting Point (ºC)
Boiling Point (ºC)
DCM
-96.7
39.6
THF
-108
66
Phenol
40.5
181.7
Cyclohexanone
-16.4
155.6
Cyclohexanol
25.9
160.8
Problems caused by volatility of feeds/products plus … solubility in DCM
Switch solvent to THF (without Lewis Acid) (Liu reported no conversion with this combination)
better chemistry - faster
Back to HP Chemscan – effect of Solvent Change Hydrogen gas uptake in DCM and THF
Liu also reported no conversion in THF !! better chemistry - faster
Reaction Progress tracking by Sampling at intervals (100C, 60bar)
Sample time
R1
R2
R3
R4
45
90
135
180
R5
R6
R7
R8
225
270
315
360
(min):
Sample time
(min):
Stirring and gas feed stopped, cooled, then manually sampled better chemistry - faster
Phenol converted but not to desired product!! Peak at ~4.71
2.40
Peak at ~ 4.86 Peak at ~5.91 (phenol) Peak at ~ 15.29
Peak Area
1.90
Peak at ~ 16.08
P and T too high (60 bar/100C) and leads to very poor selectivity
1.40
0.90
0.40
-0.10 0
45
90
135
180
225
270
Reaction Time (minute)
better chemistry - faster
315
360
Phenol conversion in stirred reactors (H P Chemscan) (using THF, 5 mol% Pd/C)
Focus switched to lower Pressure/temperature .. Improved selectivity (at expense of lower conversion) 20bar and 100C (@30 minute intervals)
Phenol Conversion (%)
30 25 20 15 10 5 0 0
30
60
90
120
Time (mins)
better chemistry - faster
150
180
210
240
Product Slate at highest phenol conversion (in THF, 5 mol% Pd/C, sampling at 30 minute intervals). Pressure 20bar (100C) gives highest selectivity
120
Reaction mixture (%)
100
~24% conversion 80 CYCLOHEXANOL 60 CYCLOHEXANONE PHENOL
40
~20% maximum 20
0 0
30
60
90
120
Time (min)
better chemistry - faster
150
180
210
240
Conclusion – THF and DCM solvent Comparison of results in 16 ml stirred reactors
Max Phenol Conversion (%) Cyclohexanone select (%) Cyclohexanol select (%)
DCM 100 88.7 11.3
Max Cyclohexanone Select (%) Phenol Conversion (%) Cyclohexanol Select(%)
THF 27.3 75.9 24.1 DCM 94.5 88.2 5.5
THF 94.7 19.9 5.3
DCM
THF
Max Cyclohexanol Select. (%)
62.3
46.2
Phenol Conversion (%) Cyclohexanone Select(%)
100 37.7
6.9 53.8
Return to flow reactor with THF solvent
Range of Conditions: -
10% palladium on carbon ~ 3.5 ml catalyst bed (~2.8ml void) mix of catalyst and glass beads (1:4 weight ratio) liquid flow : 0.1 to 4ml/min Corresponding to residence time : 0.7 to 28 min Phenol: hydrogen (molar rate) = 1:9 to 1: 90 (2.5 to 25 excess H2)
better chemistry - faster
Residence time versus concentration of phenol (5 bar, 20 bar and 60 bar all at temperature, 100 ºC)
Phenol concentration (% of feed)
120
100 5 bar 20 bar
80
60 bar 60
40
20
0 0
5
10
15 Residence time (mins)
better chemistry - faster
20
25
30
Effect of Temperature on product slate
Concentration (% of phenol feed)
(at 60bar and 0.2ml/minute liquid flow)
100
80 CYCLOHEXANOL
60 CYCLOHEXANONE PHENOL
40
PHENOL CONVERSION
20
0 40
better chemistry - faster
60
80 Temperature (ºC)
100
20bar
flowCAT Product Range (20 and 60bar , 100C)
80 PHENOL
60
CYCLOHEXANONE CYCLOHEXANOL
40 20 0 0
2
4
6 8 10 12 Residence time (minutes)
14
100 Percentage Product/Reagent (%)
Percentage Product/Reagent (%)
100
16
60bar
80 60
PHENOL CYCLOHEXANONE
40
CYCLOHEXANOL
20 0 0
2
4
6 8 10 12 Residence time (minutes)
14
16
Conclusion – THF solvent without Lewis Acid Comparison of 16ml stirred and trickle bed flow reactors
Max Cyclohexanone Sel (%) Phenol Conversion (%) Catalyst Pd: Phenol ratio Temperature (C) Pressure (bar) Reaction Time (minutes)
Max Cyclohexanol sel (%) Phenol conversion (%) Temperature (C) Pressure (bar) Reaction Time (minutes)
BATCH 94.7 19.9 20% Pd/C type 91 1:20 100 20 120
46.2 6.9 60 40 120
FLOW 95.7 99.8 10% Pd/C type 39 1:2.98 100 20 14
96.8 99.3 100 60 14
Simpler Processes – switching stirred batch to flow Using simpler processes: -
Gives limited success with batch stirred chemistry (100C/20bar gives ~20% conversion to cyclohexanone) Process successfully switched to trickle flow, fixed bed Similar pressure and temperature at both reactor types Conversion to ~80% cyclohexanone achieved Product slate can be “tuned” to give nearly 100% cyclohexanol too, simply by adjusting residence time (feed rate).
Ketone
Select conditions
Alcohol
Not unusual to get this – see Pfizer example better chemistry - faster
Diastereoselectivity and Conversion vs. Pressure 140
100 90
120 80 70 60
80
50 60
40
Increased diastereoselectivity at the higher pressures easily accessible in flow
40
20
0 0
10
20
30
40
50
60
Pressure (bar)
70
80
90
30 20 10 0 100
Conversion (%)
Cis / Trans Ratio
100
Consider scCO2 .... as Solvent inert
non-toxic readily available relatively low Pc (~ 74 bar) and Tc (~ 31C)
better chemistry - faster
Supercritical (scCO2) hydrogenation of phenol in HP Chemscan (total pressure 110 bar; pH2 80 bar; 0.05 equiv. Pd/C; 0.1 equiv. AlCl3; reaction time 120 minutes).
Gas uptake versus reaction time for the reaction at 110 bar
50 45
Partial H2 pressure = 63.3 bar Partial H2 pressure = 95.3 bar Partial H2 pressure = 82.1 bar Partial H2 pressure = 52.4 bar
40 Gas Uptake (ml)
Partial H2 pressure = 82.4 bar
35 30
Partial H2 pressure = 64.8 bar Partial H2 pressure = 52.4 bar
25 20 15
0 70
Reaction
5
Cool Down
10
80
90
100 110 120 130 140 150 160 170 180 190 200 Reaction time (minutes)
Uptake control 1.Uptake 1 (ml) Uptake control 4.Uptake 4 (ml) Uptake control 7.Uptake 7 (ml)
better chemistry - faster
Uptake control 2.Uptake 2 (ml) Uptake control 5.Uptake 5 (ml) Uptake control 8.Uptake 8 (ml)
Uptake control 3.Uptake 3 (ml) Uptake control 6.Uptake 6 (ml)
Supercritical (scCO2) hydrogenation of phenol in HP Chemscan (total pressure 110 bar; pH2 80 bar; 0.05 equiv. Pd/C; 0.1 equiv. AlCl3; reaction time 120 minutes).
100
Phenol Conversion (%)
90 80 70 60 50 40 30 20
10 0 0
20
40
60
80
Temperature ºC
better chemistry - faster
100
120
140
160
Hydrogenation of phenol in scCO2 : Product Slate (total pressure 110 bar; pH2 80 bar; 0.05 equiv. Pd/C; 0.1 equiv. AlCl3; reaction time 120 minutes).
120
Product slate (%)
100
80 CYCLOHEXANOL
60 CYCLOHEXANONE PHENOL
40
20
0 0
20
40
60
80
100
Temperature ºC
better chemistry - faster
120
140
160
Conclusion – Stirred 16ml parallel Reactors Performance in THF and scCO2
Highest Phenol Conver (%) Cyclohexanone Select (%) Cyclohexanol Select (%)
THF 27.3 75.9 24.1
Max Cyclohexanone Sel (%) Phenol Conversion (%) Cyclohexanol Select(%)
Max Cyclohexanol Sel(%) Phenol Conversion (%) Cyclohexanone Selectivity (%)
scCO 78.7 15.5 84.5 THF 94.7 19.9 5.3
scCO 94.0 14.0 6.0 THF 46.2 6.9 53.8
scCO 86.0 70.8 14.0
Finally … flow Reactors flexibility and analytical choices?
. Range of applications are possible – including liquid/liquid reactions with long reaction times and low pressure Chemistry . Advanced analytical methods, such as on-line FTIR are available and beneficial – as supplement (even replacement) for GC.
better chemistry - faster
Liquid-Liquid reaction in flowCAT: Example
Hydrogenation of nitrotoluene (to aniline) using cyclohexene as donor molecule
Transfer hydrogenation of Nitrobenzene by cyclohexene: percentage concentration of aniline as function of residence time at different temperatures 100
Percentage concentration
T=110oC T=100oC
80
T=90oC 60 T=80oC 40
20
0 0
2
4
6 Residence time (mins)
8
10
12
Aniline yield at different conditions Productivity of reactor as function of temperature and residence time 100 Residence=10mins
Conversion (%)
80
60 Residence=5mins Residence=2mins
40
Residence=1min 20
0 70
80
90
100
Temperature (oC)
better chemistry - faster
110
120
Condensation reaction in PTFE pipe reactor
Benzil + diphenylacetone
Product
Base enhanced reaction 1)Product is an intense purple colour while the starting mixture is very pale green 2) With amount of caustic held constant, several runs were performed at different flows (different reaction times), at 90C
better chemistry - faster
PTFE Reactor demonstrates reaction progress
Batch samples
Time
better chemistry - faster
12min
6min
4min
2min
0 min
On-line FTIR for catalyst degradation phenol hydrogenation example
Phenol concentration increasing
Cyclohexanone and cyclohexanol decreasing
better chemistry - faster
Conclusions Processes developed for batch reactors are not necessarily practical in flow reactors
Modifications are generally possible which will overcome the restrictions. Conversion/selectivity on the same process can often be improved by switching to flow due to the flexibility afforded by flow reactors
Flow operation can also provide a simple way of tuning the product slate by changing operating conditions (flow, temperature, pressure). Switching to scCO2 can be quite easily explored with flexible research for both batch and flow chemistry. Improvement in conversion/selectivity is not guaranteed but is possible.
better chemistry - faster