Development and scale-up of an aqueous ethanolamine scrubber for methyl bromide removal

Article abstract of DOI:10.1021/op025520o

A scrubber system was developed specifically to remove methyl bromide liberated during a demethylation process. On-line mass spectrometry (MS) was implemented and developed as a tool to monitor and quantify the methyl bromide scrubber efficiency during the demethylation reaction for laboratory and pilot-plant campaign runs. The MS technique is relatively simple to interface to existing equipment, requires no direct sample contact, and allows for the sampling from multiple ports. Results of the MS on-line monitoring using ethanolamine for both the laboratory and pilot plant showed scrubber removal efficiency of >99%. In addition to MS, ion chromatography and other gravimetric methods were implemented to confirm the level of methyl bromide consumed by the scrubber.

Full text of DOI:10.1021/op025520o

Organic Process Research & Development 2002, 6, 407−415
Technology Reports
Development and Scale-Up of an Aqueous Ethanolamine Scrubber for Methyl
Bromide Removal
Kevin Hettenbach,* David J. am Ende, Kyle Leeman, Eric Dias, Narasimhan Kasthurikrishnan, Steven J. Brenek, and
Paul Ahlijanian
Pfizer, Inc., Pfizer Global Research and DeVelopment, Chemical Research and DeVelopment, Eastern Point Road,
Groton, Connecticut 06340, U.S.A.
Abstract:
carbon was used to adsorb methyl bromide, followed by
desorption and secondary treatment with sodium hydroxide
solution (to react with methyl bromide).4 Disadvantages of
this method are the disposal of spent carbon and the need
for secondary treatment. In another paper, thermal decom-
position using a propane burner showed promise as a viable
method to reduce methyl bromide emissions;5 however,
highly acidic hydrogen bromide (HBr) was produced as a
reaction byproduct, and therefore secondary treatment is
required.
Alkanolamines in the presence or absence of an alkyl
metal hydroxide were shown in the literature6 to be a viable
medium for the abatement of methyl bromide waste gas.
Advantages of ethanolamine are that it is highly reactive to
methyl bromide and acid vapors (i.e., HBr) and the reaction
byproducts are soluble in water and suitable for wastewater
treatment. The present work extends the applications of
ethanolamine to a scrubber system for a pharmaceutical pilot
plant.
To ensure that methyl bromide was being consumed by
the scrubber, an on-line analytical tool was needed. Mass
spectrometry was chosen in the present work because of (1)
the ability to multiplex several sampling ports, (2) its high
sensitivity, (3) the ability to monitor multiple gases, and (4)
the ease at which it can be interfaced to the piping of a
scrubber system. Previous work has shown the use of on-
line mass spectrometry to monitor compositions in process
streams.7-9 The present work demonstrates for the first time
the use of MS to monitor and quantify scrubber efficiency
in a pharmaceutical pilot plant.
A scrubber system was developed specifically to remove methyl
bromide liberated during a demethylation process. On-line mass
spectrometry (MS) was implemented and developed as a tool
to monitor and quantify the methyl bromide scrubber efficiency
during the demethylation reaction for laboratory and pilot-plant
campaign runs. The MS technique is relatively simple to
interface to existing equipment, requires no direct sample
contact, and allows for the sampling from multiple ports.
Results of the MS on-line monitoring using ethanolamine for
both the laboratory and pilot plant showed scrubber removal
efficiency of >99%. In addition to MS, ion chromatography
and other gravimetric methods were implemented to confirm
the level of methyl bromide consumed by the scrubber.
1. Introduction
Methyl bromide, produced during certain demethylation
reactions, poses environmental and safety concerns and
therefore must be controlled upon scale-up. Methyl bromide
(MW ) 94.94) is a volatile organic compound (VOC) and
is a potential occupational carcinogen.1 The objectives of
this work were the following: (1) to develop an effective
scrubber system to convert methyl bromide, generated from
demethylation reactions (Scheme 1), to a form suitable for
liquid waste disposal, (2) to develop mass spectrometry (MS)
as a tool to monitor scrubber efficiency, (3) to scale-up and
transfer MS into the pilot plant to monitor methyl bromide
and to verify scrubber effectiveness, and (4) to ultimately
mitigate methyl bromide entry to the environment. For
background information and other applications of process
MS in the pharmaceutical industry, see refs 2 and 3.
There are only a few papers in the literature on the topic
of controlling methyl bromide vapors. In one patent, activated
(4) Methyl Bromide Removal Apparatus. Kokai Tokkyo Koho, Japanese Patent
No. 78-138525, 1980.
(5) Wonter-Smith, T. J.; Chakrabarti, B.; Cardwell, S. K.; Branett, G. Reduction
of Emissions of Methyl Bromide from Fumigation Chambers. Int. Pest
Control 1998, 20, 14-17.
* To whom correspondence should be addressed. Pfizer, Inc., Chemical
R&D, MS 8156-69, Groton, CT 06340. E-mail: kevin_w_hettenbach@
groton.pfizer.com.
(6) A Method of Treatment of Methyl Bromide Fumigation Waste Gas. Junkichi
Takahashi, Japanese Patent No. 49-127862, 1973.
(1) The Merck Index, 12th ed.; Budavari, S., Ed.; Merck and Co.: Whitehouse
Station, NJ, 1996; p 6110.
(2) Walsh, M. R.; LaPack, M. A. On-line Measurements Using Mass
Spectrometry. ISA Trans. 1995, 67-85.
(3) am Ende, D. J.; Preigh, M. J.; Hettenbach, K.; Ahlijanian, P.; Ward, H.
W., II. On-Line Monitoring of Vacuum Dryers Using Mass Spectrometry.
Org. Process Res. DeV. 2000, 4, 587-593.
(7) DesJardin, M. A.; Doherty, S. J.; Gilbert, J. R.; LaPack, M. A.; Shao, J.
Better Understanding of Plant and Pilot Plant Operations using On-line
Mass Spectrometry. Process Control Qual. J. 1994, 219-227.
(8) Nicholas, P. Mass Spectrometry in Process Monitoring and Control. Met.
Mater. 1990, 647-650.
(9) Nicholas, P. Process and Environmental Monitoring Using Mass Spec-
trometry. Spectroscopy 1991, 6, 36-46.
10.1021/op025520o CCC: $22.00 © 2002 American Chemical Society
Published on Web 05/15/2002
Vol. 6, No. 4, 2002 / Organic Process Research & Development
•
407

Scheme 1. Demethylation reaction: compound A, glacial acetic acid, and aqueous hydrobromic acid (48%) are combined in a
stirred vessel and heated to 90-95˚C until reaction completion is indicated by HPLC (reaction time ≈ 24 h); the reaction
liberates primarily methyl bromide during the aqueous hydrobromic acid addition, and a small percentage of methyl chloride
and hydrogen chloride are evolved due to the chloride ion present as the HCl salt of the substrate; methyl halide balance:
x + y ) 1 equiv (balance of HCl remains in solution)
the scrubber. The estimated inlet and outlet flow rates of
methyl bromide were determined by way of calibration
2. Experimental Section
2.1. On-Line Mass Spectrometry Analysis. The instru-
curves of methyl bromide (m/z 94) against argon (m/z 40),
ment used for both laboratory and pilot-plant scrubber
and then the respective flow rates were integrated over time
analysis was an Ametek/Dycor X-proof, multiport quadrupole
to compute the total input and outlet mass. A MS calibration
mass spectrometer (model: Dycor ProMaxion) with a mass
curve of methyl bromide flow versus argon flow is shown
range of 1-300 Da. The Dycor System 2000 software was
used in tandem with the ProMaxion software to collect mass
in Figure 2. Scrubber efficiencies were calculated by
comparing the area under the curve of the flow rates of
spectra and control valve switching between the inlet ports,
methyl bromide measured in and out of the scrubber and
respectively. The inlet capillary lines were polyetherether-
applying eq 1:
ketone (PEEK) tubing 0.030 in. in diameter and 12 ft in
length. The capillary was maintained at ambient temperature
total flow in - total flow out
scrubber efficiency )
× 100%
(
)
and was interfaced to vent piping via swagelok fittings. The
vacuum pressure inside the mass spectrometer was 5 × 10-6
Torr. The pressure difference between the vent line and the
mass spectrometer created the driving force for the introduc-
tion of gas sample through the capillary. As a safety
precaution, the discharge port of the mass spectrometer was
vented to an extraction device.
2.1.a. Laboratory Scrubber Analysis. The dual port
process MS was interfaced to both the inlet and outlet of a
Pyrex gas washing bottle, thus allowing the detection and
analysis of methyl bromide gases (Figure 1). Experiments
were conducted by flowing pure methyl bromide into a 500-
mL Pyrex gas bottle, containing various ethanolamine
(redistilled, 99.5+% (Aldrich)) scrubbing solutions. Mathe-
son flow controllers (model 8280) were used to regulate the
flow of methyl bromide into the Pyrex gas bottle. To obtain
quantitation, an internal standard of argon (MW ) 39.95)
was used with a constant flow rate of 10 mL/min through
total flow in
(1)
Independent verification that methyl bromide was being
consumed by the scrubber and to confirm efficiency included
(1) ion chromatography and (2) silver nitrate titration to
quantify the silver bromide precipitate.
2.1.b. Pilot-Plant Scrubber Analysis. Similar to the
laboratory setup, the dual-port process MS was interfaced
to both the inlet and outlet of the pilot-plant scrubber, thus
allowing the detection and analysis of methyl bromide (m/z
94), methyl chloride (m/z 50), and hydrogen chloride (m/z
36) gases. The configuration is shown in Figure 3. The
following procedure was followed for determining flow rates/
concentrations for each gas in and out of the scrubber:
Figure 1. Controlled flow (Matheson flow controller) of pure
methyl bromide and argon (reference gas) into an ethanolamine
scrubbing medium (500 mL). The inlet and outlet of the
scrubber bottle were monitored for methyl bromide and argon
gases using the process mass spectrometer (profile shown in
Figure 5). The flow of gas through the capillary is approximately
1 mL/min.
Figure 2. MS calibration of methyl bromide versus argon.
Methyl bromide and argon flows were controlled using a
Matheson flow controller. For example, a flow ratio of 2 refers
to 20 mL/min of methyl bromide because the argon flow was
held constant at 10 mL/min. The y-axis is the ion response ratio
of methyl bromide (m/z 94) relative to argon (m/z 40).
408
•
Vol. 6, No. 4, 2002 / Organic Process Research & Development

Figure 3. Pilot-plant equipment configuration for the methyl bromide scrubber system. Methyl bromide vapors evolved upon
heat-up of reactor (R-1) to 95 °C, were vented through condenser (C-1) into the venturi scrubber (SC-1). R-1 was under a slight
vacuum (8-20 in. H2O) during the reaction. The Dycor multiport MS was interfaced with the inlet and outlet scrubber vent lines
by connecting a 1/16 in. swagelock to a 1 in. valve fitting (Inset A).
(1) Nitrogen (m/z 28) flow rate was assumed constant and
thus was utilized as a reference gas. Note: this is a valid
assumption if the scrubber draft pressure is held constant.
(2) Sensitivity factors, calculated by dividing the MS
response by the gas concentration, were developed in the
laboratory for each gas (eqs 2a-d). Gas concentrations were
determined flowing a known amount of nitrogen (N2), methyl
bromide (MeBr), methyl chloride (MeCl), and hydrogen
chloride (HCl) using Matheson flow controllers and gas
standards. A HCl (1%) in nitrogen standard was used since
pure HCl is too corrosive for Matheson flow controllers.
Note: every gas introduced into the mass spectrometer will
exhibit a different affinity to produce ions (known as
ionization probability). When the ionization probabilities of
multiple gases are compared to each other, one can calculate
the “sensitivity” by dividing the measured amount of gas
by its concentration.
(4) The total corrected ion signal (It) was calculated with
respect to nitrogen using eq 4.
ion total (It) ) MeBrcorrected + MeClcorrected + HClcorrected
+
N2 (m/z 28) (4)
(5) The individual concentrations were calculated by
dividing the individual corrected ion signal by the total
corrected ion signal (It), as shown in eqs 5a-d.
MeBrcorrected
MeBr (%) )
MeCl (%) )
× 100%
× 100%
(5a)
(5b)
(
)
)
It
MeClcorrected
It
(
(
(
HClcorrected
HCl (%) )
N2 (%) )
× 100%
(5c)
(5d)
)
It
m/z 28
× 100%
)
m/z 94 A
MeBr
It
MeBr SENS )
MeCl SENS )
HCl SENS )
N2 SENS )
(%)
(%)
(%)
(%)
(2a)
(2b)
(2c)
(2d)
(6) The individual flow rates in and out of the scrubber
were calculated by multiplying the total flow by the
component concentration (eqs 6a-c). The total flow was
calculated by dividing the nitrogen flow by the nitrogen
concentration. Note: the nitrogen flow was unknown during
the pilot-plant campaign and was estimated by assuming 1
mol of methyl halide is released per mol of compound A
(Scheme 1), using an iterative trial and error approach.
m/z 50 A
MeCl
m/z 36 A
HCl
m/z 28 A
N2
(3) Sensitivity factors were applied to correct individual
MS response factors for methyl bromide, methyl chloride,
and hydrogen chloride relative to nitrogen (eqs 3a-c).
N2 flow (L/min)
MeBr flow (L/min) )
MeCl flow (L/min) )
HCl flow (L/min) )
× MeBr (%)
× MeCl (%)
(
(
(
)
)
)
N2 (%)
(6a)
(6b)
N2 SENS
MeBrcorrected
MeClcorrected
HClcorrected
)
)
)
× m/z 94 A
× m/z 50 A
(3a)
(3b)
(3c)
(
(
(
)
)
N2 flow (L/min)
N2 (%)
MeBr SENS
N2 SENS
MeCl SENS
N2 SENS
N2 flow (L/min)
N2 (%)
× HCl (%) (6c)
× m/z 36 A
)
HCl SENS
Vol. 6, No. 4, 2002 / Organic Process Research & Development
•
409

Methyl bromide scrubber efficiencies were calculated by
comparing the area under the methyl bromide flow rate curve
in and out of the scrubber for the duration of the reaction.
Note that m/z 28, m/z 36, m/z 50, and m/z 94 are parent peaks
for nitrogen, hydrogen chloride, methyl chloride, and methyl
bromide, respectively. All are unique peaks with the excep-
tion of m/z 36, which includes both hydrogen chloride and
methyl chloride mass fragments. In this case, the HCl
contribution to m/z 36 was determined by subtracting the
contribution from methyl chloride (e.g., by applying a m/z
36 to m/z 50 ratio for methyl chloride, equal to approximately
0.013).
2.2. Ion Chromatography Analysis. Ion chromatography
was used to independently confirm and quantify the byprod-
ucts of the scrubber. Specifically, this off-line method was
incorporated to obtain quantitative results for residual 2-(me-
thylamino)ethanol, dimethylethanolamine, and associated
salts that are produced when methyl bromide comes into
contact with aqueous ethanolamine (20%) scrubber solutions.
Samples were assayed using a standard HPLC (HP1100)
connected to an external Waters conductivity detector (model
432). The mobile phase consisted of 0.005% HNO3 and 0.005
mM EDTA tetrasodium salt dihydrate in H2O. A Waters
cation column; 4.6 × 150 mm × 5 um, was used for the
separation. Samples were evaluated versus prepared standards
of 2-(methylamino)ethanol and dimethylethanolamine ob-
tained from Aldrich.
Figure 4. Correlation of product formation with methyl
bromide liberation during the demethylation reaction. The
reaction follows first-order reaction rate kinetics. Product
formation was measured by HPLC, and the methyl bromide
offgas profile was measured using a Hewlett-Packard 6890 GC/
MS. The reaction was performed in a 1-L RC-1 vessel, adding
83.2 g of compound A, 218 g of acetic acid, and 305 g of
hydrobromic acid (48%), with an approximate headspace of
500 mL.
Scheme 2. Reaction mechanism of methyl bromide with
ethanolamine, and the formation of monomethylated
(2-(methylamino)ethanol) and dimethylated
(dimethylethanolamine) byproducts
3. Results and Discussion
3.1. Demethylation Reaction. The demethylation reaction
monitored in the laboratory and pilot plant is shown in
Scheme 1. The reaction liberates primarily methyl bromide
during the aqueous hydrobromic acid addition, and a small
percentage of methyl chloride and hydrogen chloride are
evolved due to the chloride ion present as the HCl salt of
the substrate. Laboratory analysis of the demethylation
reaction indicates that product formation (measured by
HPLC) correlates with the methyl bromide offgas profile (i.e.,
when methyl bromide offgas profile reaches baseline, reac-
tion is complete). The reaction completion time during these
studies was shown to be 12-15 h and was shown to follow
first-order rate kinetics (Figure 4). The offgas monitoring
showed that methyl bromide was liberated to the extent of
0.9 mol/mol of substrate and that small amounts of methyl
chloride and HCl were also evolved. The headspace during
the demethylation reaction was 500 mL (based on a 1-L
vessel), which explains the slower (as measured) evolution
of methyl bromide gas before it reaches baseline (i.e., all
the methyl bromide must diffuse out of headspace before
no MS signal was detected by GC/MS). On-line mass
spectrometry could potentially be used as an indicator for
reaction endpoint, which could significantly reduce reaction
time upon scale-up. Originally in situ FTIR was employed
but the highly corrosive nature of the hot HBr/acetic acid
caused pitting of the hastelloy probes and thus was aban-
doned.
the mechanism of ethanolamine 1 with methyl bromide is
to form the 2-(methylamino)ethanol‚HBr salt 2-HBr (Scheme
2). Because of the excess ethanolamine in solution, however,
2-HBr is in equilibrium with the 2-(methylamino)ethanol
free base 2 and the ethanolamine‚HBr salt 1-HBr. Subse-
quent reaction of 2 with another equivalent of methyl
bromide then forms the dimethylethanolamine‚HBr salt
3-HBr, which is also expected to be in equilibrium with its
corresponding free base 3 and 1-HBr (not shown). As a
result, for each equivalent of 3-HBr that is formed, one
equivalent of 1-HBr must also be present.
Methyl chloride offgas was detected during MS monitor-
ing of the pilot-plant reaction. The reaction of ethanolamine
with methyl chloride follows Scheme 2, with the exception
that the HBr salt is replaced with the HCl salt.
3.2.b. Methyl Bromide Scrubber Analysis using On-Line
MS. Several ethanolamine scrubber systems were tested in
the laboratory using the dual port Dycor MS (Figure 1). The
scrubbing scenarios analyzed included: (1) aqueous etha-
3.2. Development of A Methyl Bromide Scrubber
System (Laboratory). 3.2.a. Reaction Mechanism of Aque-
ous Ethanolamine with Methyl Bromide. The first step in
410
•
Vol. 6, No. 4, 2002 / Organic Process Research & Development

Table 1. Results of Pure Methyl Bromide Scrubbing Experimentsa
ethanolamine
methyl bromide
removal
scrubbing medium
aqueous ethanolamine (20%)
equivb
flow rate (mL/min)
scrubber additions
efficiency (%)d
3.2
58.3
58.3
ethanolamine (84.2 g)
water (337 g)
ethanolamine (84.2 g)
methanol (253 g)
water (84.2 g)
99.1
99.5
ethanolamine 20%), methanol (60%), water (20%)c
3.2
ethanolamine
water
9.6
NA
87.5
58.3
ethanolamine (379 g)
water (400 g)
99.9
0.0
a Note: All experiments were run at ambient temperature (22 °C), MeBr input time ) 166 min. b Based on the amount of methyl bromide added. c Water and
methanol are miscible in ethanolamine. d Removal efficiency was calculated by comparing the area under the curve of the methyl bromide flow rates in and out of the
scrubber. Efficiency ) ((in - out)/in) × 100%.
Figure 6. x-axis is the wt % of methyl bromide in aqueous
ethanolamine (20%) as prepared gravimetrically. The y-axes
are the wt % of 2-(methylamino)ethanol and dimethylethanol-
amine, representing a calibration curve for ion chromatography
analysis.
Figure 5. Results of bubbling pure methyl bromide at a flow
rate of 58 mL/min into a solution containing 420 mL of aqueous
ethanolamine (20%), see Figure 1 for setup. Methyl bromide
flow rates (inlet and outlet) were calculated on the basis of a
MS calibration curve versus argon. Scrubbing efficiency )
((9209 mL - 81.5 mL)/9209 mL) × 100% ) 99.1%.
solutions were shown to be highly effective for removal of
methyl bromide vapors, with measured control efficiencies
of >99%. In addition, the gas-liquid contact time in the
laboratory is considered shorter than in the pilot plant, and
thus it was expected that pilot-plant efficiencies would exceed
our laboratory efficiencies.
nolamine (20% by wt); (2) ethanolamine (20%), methanol
(60%), water (20%); (3) ethanolamine (100%); and (4) water
(100%), and are listed in Table 1.
The water/methanol/ethanolamine (99.5%) showed a
slightly higher removal efficiency as compared to water/
ethanolamine (99.1%) due to the greater solubility of methyl
bromide in methanol. Methyl bromide is freely soluble in
alcohol, whereas its solubility in water is only 1.75 g/100 g
of solution.10 Ethanolamine (100%) resulted in the highest
removal efficiency (99.9%) utilizing 9.6 equiv as compared
to 3.2 equiv for scrubber solutions (1) and (2). In the case
of pure water as a scrubber it was shown that essentially all
of the methyl bromide passed through the scrubber. The
aqueous ethanolamine (20%) scrubbing medium is preferred
and recommended for pilot-plant usage since the reaction
byproducts (2-(methylamino)ethanol, dimethylethanolamine,
and hydrogen bromide salts) are soluble in water and the
solution is more suitable for wastewater treatment disposal
(than methanol-containing waste streams). Thus 20% etha-
nolamine provides a good balance between high removal
efficiency, ease of waste treatment, and lower VOC loss. In
Figure 5, a typical methyl bromide profile is obtained for
inlet flow and outlet flow during a laboratory experiment.
On the basis of the measured flows the efficiency of the 20%
aqueous ethanolamine was 99.1%. In summary, ethanolamine
3.2.c. Ion Chromatography Analysis. In addition to
analyzing the aqueous ethanolamine scrubber inlet and outlet
using MS, ion chromatography (IC) was utilized to verify
that methyl bromide was being consumed. Standards were
prepared gravimetrically by bubbling varying amounts of
methyl bromide (wt %) into the aqueous ethanolamine (20%)
scrubber solution. IC analysis identified two byproducts from
the aqueous ethanolamine (20%) scrubbing of methyl
bromide: 2-(methylamino)ethanol, and dimethylethanola-
mine (Table 2). IC results have shown methyl bromide molar
recoveries of >95% in the laboratory, on the basis of analysis
of the two byproducts above. Table 2 represents the IC
standards that were made and plotted in Figure 6. In a
separate experiment, the scrubber solution was weighed
before and after the introduction of methyl bromide, and this
value was consistent with the calculated amount of methyl
bromide delivered by the flow control meter. This experiment
verified the flow meters were accurate and that methyl
bromide was being consumed by the scrubber solution.
3.2.d. SilVer Nitrate Titrations to Determine Bromide
Content. An alternative method of calculating recovered mass
from the methyl bromide/aqueous ethanolamine scrubbing
(10) Braker, W.; Mossman, A. L. Matheson Gas Data Book, 6th ed.; Matheson
Gas Products: East Rutherford, NJ, 1980; p 456.
Vol. 6, No. 4, 2002 / Organic Process Research & Development
•
411

Table 2. Ion chromatography results: methyl bromide scrubber standards prepared gravimetrically by bubbling MeBr into
20% ethanolamine solution
MeBr input
scrubber solution
2-(methylamino) ethanol
mol MeBr
dimethyl ethanolamine
mol MeBr
total mol
MeBr
consumed
MeBr
initial
final
weight (g)
MeBr
a
%
mol
weightb (g)
wt (%)c
consumedd
wt (%)c
consumede
recovery (%)f
0.185
0.537
0.674
1.400
1.897
2.509
0.0021
0.0061
0.0076
0.0162
0.0216
0.0284
107.9
108.10
107.94
106.90
109.99
108.07
107.60
0.165
0.39
0.44
0.89
1.15
1.47
0.0024
0.0056
0.0063
0.0130
0.0165
0.0211
0
0
0.0024
0.0060
0.0070
0.0155
0.0209
0.0271
112.7
98.1
91.8
95.6
96.9
95.3
107.36
106.18
108.45
106.02
104.9
0.016
0.029
0.1
0.18
0.25
0.00039
0.00070
0.00247
0.00437
0.00604
a MeBr % ) (MeBr input (g)/scrubber solution final weight (g)) × 100%. b Weighed from aqueous ethanolamine (20 wt %) stock solution. c Ion chromatography
results, wt % of scrubber solution. d 2-(Methylamino)ethanol (MW ) 75.1): mol MeBr consumed ) 1 × mol 2-(methylamino)ethanol. e Dimethylethanolamine
(MW ) 89.1): mol MeBr consumed ) 2 × mol dimethylethanolamine. f % Recovery ) (total mol MeBr consumed/mol MeBr input) × 100%.
Scheme 3. Reaction mechanism of ammonium bromide
salts with silver nitrate, forming silver bromide as a
precipitate
for methyl bromide and nitrogen during batch 4, whereby
the methyl bromide response on the outlet of the scrubber is
3 orders of magnitude lower than the inlet response. Methyl
bromide scrubber removal efficiencies for each data point
were calculated by comparing the inlet response ratio (m/z
94/28) to the outlet response ratio, resulting in >99.5%
removal. Figure 7b shows the methyl bromide offgas profile
for batch 4 (obtained from Figure 7a and using eq 6a), which
starts during reactor heat-up to 95 °C (t ) 1 h), then increases
to peak flow rate (t ) 6 h), followed by an exponential
decrease with an estimated decay constant of 0.08/h. The
decay constant (0.08/h) was calculated as the slope of the
natural log of methyl bromide flow rate versus time (h) line
during exponential decrease (t ) 6-24 h). The decay
constant was applied to estimate the methyl bromide scrubber
efficiency for batches 1-3 as well. The methyl bromide gas
flow was estimated assuming 1 mol of the methyl halide
(CH3) is released per mol of substrate (compound A) during
batch 4. The methyl bromide flow rate does not reach
baseline (t ) 25 h, reaction completion) due to the lack of
a nitrogen “sweep” (e.g., scrubber draft pressure) and large
headspace present during the reaction. Note: the headspace
volume in the 500-gal R-1 reactor was 450 gal. The scrubber
draft pressure applied to R-1 was 8-20 in. H2O. As shown,
the scrubber draft pressure and temperature remained con-
stant during the reaction period, resulting in a smooth methyl
bromide response profile. The difference between the pilot-
plant run (Figure 7b) and the laboratory run (Figure 4) is
primarily the headspace-to-reaction volume ratio: laboratory
) 1:1, versus pilot plant ) 10:1. In addition we use a slight
argon sweep in the lab where the pilot plant provided no
sweep. The scrubber analysis resulted in >99.8% removal
efficiency based on measured inlet and outlet methyl bromide
concentrations for all 4 batches (Table 3).
Table 3. Pilot-plant scrubber conditions and results
scrubber additions
MeBr
removal
efficiency (%)
ethanolamine
gal equivb (gal) (kg)
SM inputa
(kg)
water mass % ethanolamine
batch
(wt %)
1
2
3
4
23.9
26.5
25.1
28.7
32
32
27
27
39
35
31
27
100
100
100
100
499
499
480
480
24.6
24.6
21.5
21.5
av
99.9
99.9
99.8
99.9
99.9
a Compound A (MW ) 463.75). b Ethanolamine (MW ) 61.08, density )
1.012 kg/L) equivalents, based on SM.
solution is to quantify the amount of bromide in solution, as
opposed to calculating the byproduct amines 2 and 3 (Scheme
2). One way to do this is by the addition of aqueous silver
nitrate (Scheme 3), which reacts with the ammonium bromide
salts and precipitates silver bromide, leaving the resulting
ammonium nitrates in solution.
Immediately upon addition of an aqueous solution con-
taining 4.80 g (1.2 equiv based on methyl bromide) of silver
nitrate (MW ) 169.88) to the aqueous ethanolamine (20%)
scrubber solution, silver bromide (MW ) 187.78) precipi-
tated as a white solid. It should be noted that silver bromide
is a light-sensitive solid, such that all manipulations were
done with the lights off whenever possible. The solid was
collected on a Buchner funnel and dried in a vacuum oven
at 40 °C with a slow nitrogen sweep. Following filtration
and drying, 3.90 g of silver bromide was recovered,
corresponding to approximately 95% yield based on methyl
bromide input.
3.3.b. Material Balance Analysis Using MS. By observing
the demethylation reaction for the pilot-plant campaign
(Scheme 1), there is 1 mol equiv of methyl halide (CH3)
and 1 mol equiv of hydrogen chloride (HCl) per mol of
substrate (compound A) available to form reaction byprod-
ucts. Reaction offgas byproducts monitored during the pilot-
plant reaction included methyl bromide (CH3Br, m/z 94),
methyl chloride (CH3Cl, m/z 50), and hydrogen chloride
(HCl, m/z 36). Figure 8 shows the offgas flow rate profiles
3.3. Scale-Up Results (Pilot Plant). 3.3.a. On-Line
Monitoring of Methyl Bromide Using MS. The scrubber
solution makeup and starting material inputs for each batch
used in the pilot plant are shown in Table 3, whereby the
ethanolamine equivalents (relative to methyl bromide) ranged
from 27 to 39 and the wt % aqueous ethanolamine used was
21.5-24.6%. Figure 7a shows the MS raw data response
412
•
Vol. 6, No. 4, 2002 / Organic Process Research & Development

Figure 7. (a) MS raw data response for methyl bromide (m/z 94) and nitrogen (m/z 28) in and out of the scrubber (left axis) during
batch 4. The right-hand axis shows the methyl bromide removal efficiency by comparing the inlet response ratio to the outlet
response ratio for each data point (t), whereby removal efficiency ) (((94/28)In - (94/28)Out)/(94/28)In) × 100%. (b) Methyl bromide
flow rate ( L/min), temperature profile, and scrubber draft pressure profile during the demethylation reaction for batch 4. The
methyl bromide flow rate reaches peak (1.6 L/min) at t ) 6 h, then follows an exponential decrease with a decay constant of 0.08/h.
Scrubber efficiency was calculated by comparing the area under the methyl bromide flow rate curve in and out of the scrubber.
Scrubbing efficiency ) ((1194 L - 0.68 L)/1194 L) × 100% ) 99.9%.
Table 4. Methyl halide (CH3) and chloride (Cl) material
balance: batch 4
methyl halide
(equiv)
chloride
(equiv)
constituent
methyl bromde offgas
methyl chloride offgas
HCl offgas
HCl in solution
totals
0.86a
0.14a
-
-
0.14a
0.06a
0.80b
1.0
-
1.0
a Amount based on area under inlet flow profile curve, assuming 1 equiv of
compound A is available for methyl halide generation. b Material balance closure
based on 1 mol of HCl per mol of substrate (compound A).
Figure 8. Scrubber inlet offgas response ratio and offgas flow
rate profile for methyl bromide, methyl chloride, and HCl. The
flow rate profiles were calculated assuming one mol of methyl
(CH3) is released per mol of substrate (compound A). The total
mass for each offgas (shown in brackets) is calculated as the
area under the inlet flow rate curve.
3.3.c. Ion Chromatography Analysis (Pilot Plant). In
addition to MS, an independent method (ion chromatography)
was developed to verify the amount of methyl bromide that
was being consumed by the aqueous ethanolamine scrubber
solution during the pilot-plant reactions. Ion chromatography
analysis has verified that methyl bromide (in addition to
methyl chloride vapors) was consumed, whereby the overall
methyl halide recovery over all 4 batches was 94%. The
theoretical methyl halide recovery is 100%, since the
demethylation reaction was shown to go to completion (i.e.,
compound A remaining ) 0%). The overall methyl halide
recovery was estimated by comparing the total moles of
methyl halide available (based on compound A input) to the
amount of methyl halide consumed by the scrubber based
on the 2-(methylamino)ethanol and dimethylethanolamine IC
measurements. Methyl halide recovery results for IC analysis
for batches 1 through 4 are shown in last column of Table
5 and ranged from 62 to 127%.
At this point a discussion of the MS results (Table 3)
versus the IC results (Table 5) is warranted. As discussed
above, the MS provides relative concentrations on the inlet
and outlet sides of the scrubber, which in the pilot plant
indicated concentrations up to 3 orders of magnitude higher
for methyl bromide on the inlet side than on the outlet side.
One reason we believe that methyl bromide vapor as
measured by m/z 94 does not rapidly return to baseline at
the end of the reaction, as seen in Figure 7a, is because there
for all three gases, whereby methyl chloride and HCl flow
rates are much smaller compared to the methyl bromide flow
rate profile. The offgas response ratio represents a relative
concentration. The total value of gas flow for each byproduct
was estimated by assuming that 1 mol of the methyl halide
(CH3) is released per mol of substrate (compound A) during
batch 4 (Note: reaction completion was verified during pilot-
plant runs). Scrubber removal efficiencies for methyl chloride
and HCl were calculated as 99.9 and 92.2%, respectively,
using the same methodology as methyl bromide removal
efficiency calculations.
The resulting material balance for methyl halide (CH3)
and chloride (Cl) is shown in Table 4. On the basis of a
batch 4 compound A input of 28.7 kg (0.062 kmol), there
are 0.062 kmol of methyl halide and 0.062 kmol of chloride
available. The amount of the methyl halide which forms
methyl bromide is 0.86 equiv, compared to 0.14 equiv for
methyl chloride. The amount of chloride released as offgas
is 0.20 equiv (0.14 equiv as methyl chloride, 0.06 equiv as
HCl), with an estimated 0.80 equiv remaining in solution.
In summary, these data obtained from pilot-plant runs are
consistent with the demethylation laboratory results discussed
above and shown in Figure 4.
Vol. 6, No. 4, 2002 / Organic Process Research & Development
•
413

Table 5. Ion chromatography results: pilot plant
monoa
dib
scrubber
weight (kg)
kmol Me
consumed
kmol Me
consumed
total kmol Me
consumed
SM Inputd
(kg)
kmol Me
availablee
methyl
batch
wt %c
wt %c
recovery (%)f
1
2
3
4
499
499
480
480
0.565
0.84
0.49
1.1
0.0378
0.0563
0.0315
0.0711
0.01
0.00113
0.00401
0.00206
0.00746
totals
0.0389
0.0603
0.0335
0.0786
0.211
23.9
26.5
25.1
28.7
0.0515
0.0571
0.0541
0.0619
0.225
75
106
62
127
94%
0.0355
0.019
0.0685
a 2-(Methylamino)ethanol (MW ) 75.1): kmol methyl (Me) consumed ) 1 × kmol 2-(methylamino)ethanol. b Dimethylethanolamine (MW ) 89.1): kmol methyl
(Me) consumed ) 2 × kmol dimethylethanolamine. c Ion chromatography results, wt % of scrubber solution. d Compound A input (MW ) 463.75). e kmol Me available
) SM input (kg) × [Me (MW ) 15)/SM (MW ) 463.75)]. f % methyl recovery ) (total kmol methyl consumed/kmol methyl available) × 100%.
is residual vapor present in the headspace and vent-lines.
The vapors are not swept with nitrogen flow but rather only
a slight vacuum (20 in. H2O) is created downstream of the
scrubber to pull any excess vapor. This would explain why
batch 1 resulted in only 75% of the theoretical amount of
methyl bromide coming in contact with the scrubber solution.
Approximately 25% remained in the reactor headspace and
vent lines (remaining amount of methyl bromide gas was
consistent, using ideal gas law calculation). In the present
campaign we had a significant amount of headspace relative
to reaction volumesonly 50 gal of reaction mixture in a 500-
gal tank. In addition, the entire system upstream of the
scrubber is sealed between batches. The scrubber solution
is also replaced with fresh ethanolamine solution between
batches. Thus, when batch 2 began, we believe the residual
vapors that remained at the end of batch 1 were forced into
the scrubber solution. The final analysis of batch 2 appears
to be consistent with this theory, provided some residual
vapor was left behind at the end of batch 2. Between the
batches there may be varying amounts of residual vapors
left behind, depending on draft pressures employed during
the runs. At the end of the campaign (batch 4) nitrogen was
used to purge the reactor, filter, and all vent lines through
the scrubber. Thus, residual vapors left over from batches 3
and 4 would have been forced into the scrubber (Note:
concentration of methyl bromide (m/z 94) was 2 orders of
magnitude higher between batches than at the start of the
campaign, indicating residual methyl bromide). This explains
the 127% theoretical recovery for batch 4. In addition, the
magnitude of the MS scrubber inlet concentration response
(area under m/z 94/m/z 28 profile) correlated with IC results
for batches 1 through 4. In summary, an excellent methyl
halide recovery balance is achieved for the direct analysis
of the scrubber byproducts of methyl bromide with ethanol-
amine versus starting material input over all 4 scrubber
batches as is appropriate due to the process procedures used
during the campaign.
that flow is necessarily occurring through the scrubber. More
important, the MS provided real-time measurements that
showed that methyl bromide concentrations at the outlet of
the scrubber were negligible. In short, a reference gas such
as nitrogen or argon with a constant flow rate will ensure
that gases are swept through the scrubber and that the MS
method can be used to calculate an actual flow rate in and
out of the scrubber.
3.3.d. Demethylation Reaction Completion Analysis (Pilot
Plant). The reaction completion time during laboratory
studies was shown to be 12-15 h. By observing Figure 7b
(batch 4 reaction) the majority of the methyl bromide has
evolved after 12-15 h, and the slow decrease towards
baseline is primarily due to the lack of a sweep present and
the large headspace effect during the reaction. On-line mass
spectrometry could potentially be used as an indicator for
reaction endpoint, which could significantly reduce the batch
time (Note: reaction time during the campaign was set at
24 h). A nitrogen sweep would help move the methyl
bromide through the reactor headspace into the scrubber more
rapidly, and MS end point detection could be enhanced.
3.3.e. Recommendations for MS Monitoring On-Scale. (1)
Reduce reactor headspace during reaction, this will increase
the correlation between reaction endpoint and methyl bro-
mide concentration detected by MS. (2) Provide a small
nitrogen sweep preferably with a known flow rate through
the scrubber system. (3) Maintain steady scrubber draft inlet
pressure during on-line monitoring for stabilization of the
MS signal. (4) Purge after the reaction to displace all residual
methyl bromide vapors to the scrubber before proceeding to
the next batch. (5) Utilize MS as a quality-control tool to
ensure methyl bromide vapor has been purged from system
before the vessel is opened.
4. Conclusions
Methyl Bromide is a VOC produced during demethylation
reactions, that poses adverse safety and environmental issues.
Development of the methyl bromide scrubber system using
ethanolamine resulted in a successful transfer of the MS
technology from the laboratory to the pilot plant. The MS
on-line monitoring showed >99% removal of methyl bro-
mide at both scales. The ability to monitor the offgas profile
on-line for both the inlet and outlet of the scrubber has been
demonstrated, thereby ensuring a safe and environmentally
benign process during production. In addition, the ability to
The IC results and vapor phase MS results obtained from
the pilot plant highlights an important point. In the lab testing,
where we have a controlled flow rate of reference gas (argon)
that is blended with the methyl bromide, there is an
established flow rate of gases in and out of the scrubber. On
the other hand if the flow of gas stops, the MS continues to
detect levels of gases on the inlet and outlet side in proportion
to their concentration. These measurements do not ensure
414
•
Vol. 6, No. 4, 2002 / Organic Process Research & Development

monitor multiple offgases (including methyl bromide, methyl
chloride, and HCl) and determine a material balance estimate
by MS was also shown. Ion chromatography (IC) and silver
nitrate were used to confirm scrubber composition and
verified that methyl bromide and methyl chloride were being
trapped. The aqueous ethanolamine (20%) scrubber medium
effectively converted the methyl bromide offgas to the
2-(methylamino)ethanol and dimethylethanolamine byprod-
ucts (and associated HBr salts) for aqueous waste disposal.
Another application for MS monitoring during demethylation
reactions is to utilize the methyl bromide offgas profile as
an indicator for reaction completion, whereby significant
time- and cost savings may be realized.
Acknowledgment
The following Pfizer Pilot Plant Personnel are acknowl-
edged for their assistance, guidance, and support: Charles
Santa Maria, Nick Andreopoulos, and Eileen Callaghan;
Steve Colgan and Mike Preigh for ion chromatography and
LC/MS support as well as their helpful discussion; William
Sands and Tony Slapikas (Ametek Process Instruments, 150
Freeport Road, Pittsburgh, PA 15238) for their assistance
and guidance using MS technology.
Received for review February 14, 2002.
OP025520O
Vol. 6, No. 4, 2002 / Organic Process Research & Development
•
415