566 Ind. Eng. Chem. Res.

1994, 33, 566-574

Chemical Processing in High-Pressure Aqueous Environments. 4.

Continuous-Flow Reactor Process Development Experiments for
Organics Destruction
Douglas C. Elliott,* M. R. Phelps, L. John Sealock, Jr., and Eddie G. Baker
Pacific Northwest Laboratory, P.O. Box 999, Richland, Washington 99352

A high-pressure (20 MPa) and high-temperature (350 °C) liquid water processing environment was
used to treat various wastewaters and model compounds. Organics were converted to methane and
carbon dioxide in the presence of a fixed bed of nickel or ruthenium catalyst. Nitrates were destroyed
by reaction with methanol in the presence of a nickel catalyst. Noncatalytic hydrolysis of carbon
tetrachloride and chloroform was also demonstrated. Three scales of continuous-flow reactors were
used in these tests. Extended tests to demonstrate catalyst lifetimes were also performed. Several
examples of test results with actual industrial waste streams showed that this process can be effectively
Downloaded via NANJING UNIV AERONAUTICS & ASTRNTCS on March 6, 2022 at 19:50:28 (UTC).

used with appropriate catalysts to clean wastewater and recover waste organics as useful fuel gas.
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Introduction term processing results (Baker et al., 1989a,b; Elliott et

A high-pressure (up to 22 MPa) and high-temperature
(250-350 °C) liquid water processing environment can be
used to treat contaminants in process waters by converting
the organics to gases. This system can be utilized for both
environmental cleanup and energy recovery. Under the
proper conditions, concentrated organics are converted to
useful fuel gas and dilute hazardous organics are destroyed.
Known as the Thermochemical Environmental Energy
System (TEES), this system is operated as a liquid-phase,
heterogeneously catalyzed process at nominally 350 °C
and 20 MPa to produce a methane/carbon dioxide product
gas from water solutions or slurries of organics. TEES is
a registered trademark of Onsite*Ofsite, Inc., of Duarte,
CA, who holds the commercial license.
The history of the development of the TEES concept
was recently reviewed (Sealock et al., 1993). This article
is the fourth in a series describing the most recent results
in the developmental effort. The earlier papers addressed
(1) the processing environment (Sealock et al., 1993), (2)
catalyst systems for this environment (Elliott et al., 1993a),
and (3) batch reactor tests with various organic chemical
components and waste streams (Elliott et al. 1994). Here
we report the results of continuous-flow reactor studies
with fixed beds of catalyst in a tubular reactor. These
tests demonstrate the scale-up of a useful reactor con-
figuration for industrial application of the TEES tech-
nology.
Background
The use of high-pressure liquid water at elevated
temperatures as a processing environment has recently
been reviewed (Sealock et al., 1993). One application of
this processing environment is for catalytic gasification of
al., 1991). This article provides additional results with
other feedstocks and with longer-term and larger-scale
operations.
Experimental Section
The equipment, procedures, and catalysts described
below were used for the continuous reactor testing of high-
pressure, aqueous-phase destruction of organics by gas-
ification.
Equipment. Gasification tests were carried out in three
scales of fixed-bed catalytic tubular reactor.
A. Bench-Scale Operation. The bench-scale unit,
shown schematically in Figure 1, was described in detail
in an earlier report (Elliott et al., 1991). The unit consisted
of a 1.83-m long X 25-mm i.d. (6-ft long X 1-in. i.d.) 304SS
tube which was fed from a cylindrical feed tank by a
reciprocating plunger pump. The reactor was heated by
an electrical resistance furnace and essentially served as
both the preheater and the reactor. Pressure was con-
trolled in the reactor by a dome-loaded, back-pressure
regulator. The pressure regulator was followed by a
condenser/ separator system in which liquid samples were
recovered. Uncondensed product gas was passed through
flowmeters and vented.
B. Microscale Operation. A microtubular version of
this reactor system was used for catalyst lifetime studies.
All the components shown in Figure 1 for the bench-scale
unit were the same for the microtubular reactor system
except for the pump and the reactor. A smaller recip-
rocating plunger pump was used to feed a reactor consisting
of 0.3 m of 13-mm (1/2-in.) 316SS tubing. In addition to
the reactor, there was a preheater section consisting of 2.1
m of 3.2-mm (1/8-in.) tubing. Both the reactor and the

organics. In this application, which we call TEES, catalysts
accelerate the reaction of organics with water and produce
methane and carbon dioxide as the product gases. TEES
has been reported both as a means of recovering useful
energy from organic-in-water streams and as a water
treatment system for dilute organic contaminants. Batch
reactor test results have demonstrated process applicability
to a wide range of organic components (Sealock et al.,
1988; Baker and Sealock, 1988; Elliott et al., 1994). De-
velopment of catalysts for this processing environment
has also been an important factor in making this processing
technology viable (Elliott et al., 1993a). Previous reports
of continuous reactor tests of TEES were primarily short-
0888-5885/94/2633-0566$04.50/0
preheater were contained in the electrical resistance reactor
furnace. This system was used for long-term tests because
it minimized chemical feed and waste handling and could
be run unattended.
C. Engineering-Scale Operation. The scaled-up
reactor system (SRS) was based on the bench-scale design.
It was a transportable system designed at a scale of 1/2
ton/day of wet feed for obtaining engineering data for
further scale-up of TEES. The SRS was mounted on a
single 2.44-m X 3.05-m (8-ft X 10-ft) skid platform that
could be transported on a single flat-bed truck. Equipped
with three fixed-bed tubular reactors and supporting
equipment to achieve conversion, the test system’s capacity
© 1994 American Chemical Society

Liquid/Gas Cooler
Feed Pump
Figure 1. Schematic of bench-scale reactor system.
Figure 2. Schematic of scaled-up reactor system.
was a1/2 ton/day liquid feed with a design flow rate of 22
L/h. Design working conditions for the reactors were 350
°C at 24 MPa.
As diagrammed in Figure 2, the aqueous organic
feedstock was either loaded into or prepared for processing
in feed tanks at the front end of the process. Each of the
300-L tanks was constructed of mild steel, lined with a
chemical-resistant epoxy, and had a centrifugal pump to
circulate and agitate the tank’s contents. These pumps
were also used to transfer feedstock from one tank to
another or from the feed tank to the high-pressure feed
pump. A hinged lid on the mixing tank provided access
for either loading or preparing feedstock, and the feed
tank was completely sealed. Feedstock could be loaded
into the feed tank from the mixing tank, or it could be
directly pumped in from an outside source.
Once prepared, the feed stream was pumped through
the tube side of the preheater by a positive displacement
pump. This pump was a standard Kerr KD 1250B fitted
with a 16-mm (5/8-in.) plunger and a half-stroke crank-
shaft. Constructed of mild steel, the pump was capable
of delivering 38 L/h at 27.6 MPa. The preheater was a
Ind. Eng. Chem. Res., Vol. 33, No. 3, 1994 567
Gas Analysis
and Metering
o i-
o o
\73 v-
=
g.
Back Pressure
Li (/)
Regulator
Cooling
Water
Condensate
Collector
Vent Line
Regulator
double-tube type heat exchanger that was constructed of
316SS tubing. The tube side was 9.5-mm (3/8-in.) o.d. X
1.2-mm (0.049-in.) wall. The shell was 19-mm (3/4-in.)
o.d. X 2.1-mm (0.083-in.) wall. Pressure in the heat
exchanger was limited by the shell side, which had a
maximum allowable working pressure (MAWP) of 23.4
MPA at 350 °C. With a total length of 15 m, the preheater
could bring the feedstock to within 30 °C of the final
operating temperature.
The final heating of the feed was accomplished in the
preheater vessel which was enclosed in a 1.45-m X 0.4-m
X 0.4-m cabinet fitted with ceramic furnaces. The
preheater vessel (H-l) was 1.8 m (6 ft) long and had a
76-mm (3-in.) o.d. with a 51-mm (2-in.) i.d. High Pressure
Equipment Company manufacturered the vessel out of
304SS and rated it to 23.40 MPa at 350 eC. The furnace
assembly contained three sets of tube-type ceramic heaters.
Each set operated independently of the other two, thereby
creating a three-zoned furnace. The furnace was rated at
538 °C with power requirements of 7500 W at 230 V.
After heating, the feed was routed to the two reactor
vessels (R-l and R-2). Flow could be directed through the
568 Ind. Eng. Chem. Res., Vol. 33, No. 3, 1994
reactors in either a series or parallel configuration or
through only R-2. The system has been run only in the
series configuration because of inconsistent flow charac-
teristics in parallel reactor operation. Each reactor was
fitted with two sets of ceramic-tube-type external furnaces
to provide makeup heating of the feed stream. The first
set of heaters, located on the bottom third of the reactors,
was 46 cm (18 in.) long and rated at a total of 2200 W at
230 V. The reactors were of a tubular fixed-bed design
with volumes of 3.7 L each. The preheater vessel was of
the same design and size as each of the reactors and, if
filled with catalyst, could serve as a third reactor. After
leaving the reactor(s), the product stream was routed
through the outer tube of the double-tube heat exchanger
to provide heat for preliminary heating of the feed stream.
After passing through the exchanger, the product stream
entered a liquid/gas cooler that fed a liquid/gas separator
tank where process water was reclaimed and combustible
gases were vented.
Pressure in the system was regulated to approximately
20.8 MPa by a back-pressure regulating valve capable of
yielding a 41.5-MPa pressure drop across its restriction
piston. One rupture disk assembly and two pressure relief
valves were incorporated in the system to ensure system
pressure never exceeded 23.5 MPa. The rupture disk
assembly, located at the front end of the system between
the feed pump and the heat exchanger, was equipped with
a disk rated at 21.8 MPa. The first pressure relief valve,
located at the top of the preheater vessel, was set to relieve
at 23.5 MPa. The second pressure relief valve, located at
the top of the preheater vessel, was also set to relieve at
23.5 MPa. The rupture disk vented feedstock from the
feed pump into the feed tank. A check valve between the
rupture disk assembly and the heat exchanger ensured
that the remainder of the system did not depressurize
through the rupture disk assembly. Both pressure relief
valves went through a catch pot and into a vented catch
barrel located outside the west wall of the operations
facility.
Procedures. For these experiments, the test was
preceded by a period of heating of the catalyst bed in
hydrogen to ensure reduction of the catalyst. Actual
startup of the bench-scale equipment usually required an
hour or more to bring operating conditions to the desired
levels. Startup of the SRS required in excess of 4 h.
Operating data were recorded, and data windows were
defined based on steady-state (or near-steady-state)
operating conditions.
A. Elemental Analysis. A Perkin-Elmer (P-E) 240B
analyzer was used for analysis of carbon deposits on the
catalysts. The P-E 240B determines carbon, hydrogen,
and nitrogen by detecting and measuring their combustion
products (CO2, H2O, and N2). Combustion occurs in pure
oxygen under static conditions, and the products are
analyzed by thermal conductivity. Helium is used to carry
the combustion products from the combustion furnace
through the reduction tube (which reduced any NO) and
through a series of detectors and traps. The instrumental
readout was in millivolts, which was used to calculate the
composition based on reference compounds.
B. Gas Analysis. Gas samples were withdrawn from
the smaller two systems by an automated sampler and
manually. Automated samples were withdrawn typically
every 5 min and manual samples every 20-30 min.
However, in the longer-term tests, the automatic procedure
was limited to once every 60 min, and manual samples
were withdrawn twice a day. The gaseous stream was
composed principally of CO2, CH4, H2, and C2+ hydro-
carbons, as well as water vapor.
The online gas analysis equipment was able to measure
concentrations of carbon oxides, hydrogen, methane, and
C2 hydrocarbons with reasonable accuracy and precision
on a near-continuous (real-time) basis with a very short
(<5 min) turnaround. A thermal-conductivity detector
was used.
Manual gas analysis was performed by gas chromatog-
raphy (GC) using a Carle AGC. The AGC used a set of
three columns (80% Porapak N + 20% Porapak Q, 50/80
mesh; Molecular Sieve 13X, 80/100 mesh; and 8% -
on Chromosorb W AWDMCS, 80/100 mesh) with
automatic switching controls to separate hydrogen, carbon
dioxide, carbon monoxide, methane, ethane, ethylene,
oxygen, and nitrogen. Propane and butane were detected
in the backflush from the columns. A thermal conductivity
detector was used in the AGC. Hydrogen was determined
through the use of the Carle hydrogen transfer system, a
metallic membrane for hydrogen separation by diffusion.
The GC signal was processed in a Spectra-Physics 4290
integrator, which plotted the signal, integrated peaks, and
determined retention times; it then printed a table of data
including peak identifications and gas composition based
on previously analyzed standard gas mixtures. A 0.7-cm
gas sample was used for analysis.
C. Calculation of Carbon Conversion to Gas.
Following the analysis of the gas samples from the
experiments, calculations were made to determine the
composition of the actual gas product and to determine
the conversion of the organic feedstock to gases. For the
manual GC analyses, an air-free gas composition correction
was used. Carbon conversion to gas was then calculated
on a mass basis for the carbon in the product gases (from
either the direct online analysis or the manual analysis
corrected to air-free basis) as a percent of the carbon in
the feedstock. In those cases in which the mass balance
was deficient, the gas yield was normalized to a 100%
recovery.
D. Aqueous Byproduct Analysis. The aqueous
byproducts were analyzed for ammonia, chemical oxygen
demand (COD), and pH. The ammonia analytical pro-
cedure involved the use of an ammonia-specific gas-sensing
electrode (ORION 95-12). The electrode was calibrated
in 10 M NaOH. The COD method used was the dichromate
closed reflux colorimetric method, No. 5220D (American
Public Health Association, 1992). The pH was measured
by an electrode calibrated against buffered pH 4.0,7.0, or
10.0 solutions.
Catalysts. The catalysts used in these experiments
were acquired from commercial catalyst manufacturers.
The available information on their compositions and
properties is given in Table 1.
Results and Discussion
The experiments yielded information on catalyst effects
and lifetimes and use of the TEES process with various
industrial-related feedstocks.
Evaluation of Catalysts. Several metal catalysts have
been evaluated in the TEES continuous reactor systems.
On the basis of catalyst screening in batch reactor tests
(Elliott et al., 1993), the focus of these evaluations was on
the use of nickel and ruthenium as the catalytic metal.
Catalyst support effects are also an important consider-
ation, and several different support materials and un-
supported Raney nickel catalyst were evaluated.
Short-term test results (<8 h), given in Table 2, show
the variation in catalyst activity attained in TEES with
nickel catalysts. The Ni-0750 and BASF catalytic results
are nearly equivalent. The Raney nickel test is slightly
different because only a limited amount was available for
the test. With the smaller catalyst bed, the lower feed
rate achieves an equivalent space velocity for comparison
 Ind. Eng. Chem. Res., Vol. 33, No. 3, 1994 569
Table 1. Conventional Metal Catalysts Tested in TEES
catalyst manufacturer composition SA, m2/g PV, cm3/g
Ni-0750 Engelhard 48% Ni on 7-alumina 145 0.50
BASF BASF Corp. NiO on proprietary NA NA
Raney* 5886 Grave/Davison pellets 25 NA
C150 United Catalyst 62% Ni on SÍO2/AI2O3 200 0.4-0.5
5% Ru/5-A1203 Engelhard 1/8-in. star extrudate 147 0.61
5% Ru/ZrOa-D Engelhard 1/8-in. extrudate NA6 NA
3% RU/7-AI2O3 Engelhard 1/8-in. star extrudate NA NA
0
Raney is a registered trademark of the W. R. Grace Co., Davison Specialty Chemical Division.6 NA, not available.
Table 2. Short-Term Experimental Results with Nickel
Catalysts*
catalyst Ni-0750 BASF Raney C150
feedstock p-cresol p-cresol p-cresol p-cresol Ni-0750

gasification of carbon, % 99.4 97 99.9 68.6 D BASF

reduction of COD, % 99.994 99.9 99.8 97.3 *
Raney Nickel
feed, L/h 1.6 1.6 1.1 1.1
LHSV, L L"1 h-i 2.7 2.8 2.8 1.8 ® Ru/AI203
gas yield, L/h 41.6 40 30.9 21.1
gas yield, L/L feed 26.0 25.0 28.1 19.2
effluent COD, ppm 3 76 105 1300
effluent pH 4.3 4.8 3.5 4.0
gas composition, vol %
methane 62 55 57 61 Time, hr on stream
carbon dioxide 36 44 30 33
hydrogen 1.5 1.2 12 4.0 Figure 3. Catalyst metal crystalline growth over time in TEES.
ethane 0.2 0.2 0.2 0.4
backflush 0.4 0.4 0.7 1.7 than the Ni-0750. Note that Table 3 is a combination of
Btu/SCF6 644 577 640 688 bench-scale results for the Ni-0750 catalyst and microscale
0
Catalyst bed conditions: 350-360 °C and 20-22 MPa. Organic results for the BASF catalyst, so the gas yield results cannot
loading of 1.8 wt % in water as feedstock. 6 SCF, standard cubic feet. be compared directly from one catalyst to the other.
Changes in the catalyst resulting from exposure to the
Table 3. Extended-Time Experimental Results with Nickel TEES environment is the primary explanation for the
Catalysts* loss in activity. As illustrated in earlier tests (Elliott et
catalyst Ni-0750 Ni-0750 BASF BASF al., 1993a), the composition of the catalyst support can be
feedstocks p-cresol p-cresol p-cresol& p-cresol& chemically changed and the metal can become agglom-
phenol phenol
time on stream, h 38 72 240 620 erated, limiting its active surface area. Long-term results
presented here further demonstrate the effect of the change
gasification of carbon, % 99.1 23.4 101.4 89.2
(and lack of change) in catalysts upon exposure to the
reduction of COD, % 99.94 68.2 99.8 97.6
feed, L/h 1.6 1.6 0.036 0.017 high-temperature liquid water.
LHSV, L L'1 h-1 2.7 2.7 1.2 0.8 Metal agglomeration is a primary concern for catalysts
gas yield, L/h 41.5 8.9 1.1 0.5 used in TEES. Catalytic metal crystallite size has been
gas yield, L/L feed 25.9 5.6 30.6 29.4 correlated with the loss of active surface area and, hence,
effluent COD, ppm 30 14900 95 1710 catalytic activity (Pearce and Patterson, 1981). Figure 3
effluent pH 3.9 4.7 NA NA shows that different catalysts exhibit different rates and
gas composition, vol %
amounts of agglomeration. For example, the Ni-0750
methane 61 45 59 59
carbon dioxide 36 41 39 37 catalyst shows a nearly straight line growth in crystal size,
hydrogen 1.6 6.2 0.4 0.7 while the BASF catalyst shows an initial growth of crystal
ethane 0.4 1.4 0.5 0.7 size followed by a stable period which has been demon-
backflush 0.6 5.7 1.0 1.7 strated for up to 4 weeks. The change in the Raney nickel
Btu/SCF 646 669 634 666 is more nearly like the Ni-0750 than the BASF catalyst.
“
Catalyst bed conditions: 350-369 °C and 21-22 MPa. Organic The data in Figure 3 are derived from tests in all three
loading of 1.8 wt % in water as feedstock. scales of continuous-feed reactor system. There is no
significant difference in catalytic metal agglomeration that
with the Ni-0750 and BASF catalysts. The lower feed can be correlated to any of the scales of operation.
rate results in a lower gas yield, albeit a similar yield in Changes in the catalyst support contributed to the
liters of gas per liter of feed. The higher hydrogen content activity changes. The BASF catalyst support was con-
in the product gas is attributable to the higher temperature verted to a hydroxide form and crystal growth paralleled
in the preheat zone (284 vs 227 °C for BASF), also because the metal crystal growth, as shown in Figure 4. As reported
of the lower feed rate. The fourth catalyst, C150, is in our earlier publication (Elliott et al., 1993a), high-
noticeably less active even at a lower space velocity. surface-area 7-alumina, as used in the Ni-0750 catalyst,
More important is the long-term activity of the catalyst. was readily converted to bóhmite (AlOOH), which has a
As shown by the results in Table 3, the long-term operation much reduced surface area. The change to bóhmite may
of Ni-0750 shows a definite loss of activity over time. The have allowed the water to interact more strongly with the
long-term loss of activity is apparent even when stable nickel and caused the nickel agglomeration that was seen
short-term activity has been demonstrated as shown in in the Ni-0750 catalyst.
column 2 of Table 2 (see also Elliott et al., 1993b). The loss A different situation was found with ruthenium. Ru-
of activity is evident in the reduced level of COD thenium showed no evidence of crystal growth. As seen
destruction, lower gas yields, and shift in gas composition in Figure 3, even a slight reduction in crystal size was
away from methane to hydrogen and higher hydrocarbons. apparent (although the change is just at the limit of the
The BASF catalyst shows much better long-term stability experimental error of the analysis). There was little
570 Ind. Eng. Chem. Res., Vol. 33, No. 3, 1994

40 conversion for comparison of processing rate. There was

35 a noticeable improvement in processing rate achievable
with the 5% Ru catalyst compared with the 3% Ru (both
on high-surface-area alumina supports). Operation of the
C. 25
<u 3% Ru catalyst at higher flow rate resulted in less than
I 20 99% conversion. The 5% catalyst showed a significant
| 15 loss of activity over a 9-month period between tests
resulting from unknown effects in handling and storage.
The 5 % Ru on zirconia was tested only in the microreactor
5
system. It showed good activity but was tested at a lower
0
space velocity.
0 120 240 360 480 600 Table 5 provides results of long-term tests for compar-
Time, hr on stream ison of two ruthenium catalysts. These tests showed good
maintenance of conversion over time for both catalysts.
Figure 4. Relation of crystallite growth in nickel metal and catalyst
support material in the BASF catalyst.
COD destruction remained high with good gas yield. The
alumina-supported catalyst showed significant loss of
Table 4. Short-Term Experimental Results with activity after 10 days of operation as evidenced by the
Ruthenium Catalysts* lower space velocity required to maintain conversion levels.
catalyst 3% on 5% on reused 5% on A shift away from methanation was noticeable toward the
AI2O3 AI2O3 5% ZrC>2 end of a 6-week test with the zirconia-supported ruthe-
feedstock p-cresol p-cresol p-cresol phenol nium, while the level of conversion remained high.
gasification of carbon, % 100.4 99.9 96.3 101.2 Evaluation of Feedstocks. A variety of feedstocks
reduction of COD, % 99.8 99.95 99.7 99.4 have been tested in the continuous reactor systems. This
feed, L/h 1.4 1.95 1.2 0.03 discussion focuses on the testing in the following order:
LHSV, L L-1 h-1 2.4 3.6 2.0 1.0
bench-scale reactor tests with chemical model compounds;
gas yield, L/h 36.1 47.3 30.0 0.8
gas yield, L/L feed 25.8 24.3 25.0 26.7 processing tests with actual waste streams in the bench-
effluent COD, ppm 80 20 118 300 scale system; and test results at the engineering devel-
effluent pH 3.1 2.8 3.1 NA opment scale in the SRS.
gas composition, vol % A. Bench-Scale Reactor Tests with Chemical
methane 61 60 65 64 Models. Three groups of chemical model compounds were
carbon dioxide 36 38 33 35
tested. The first group consisted of oxygenates, phenol
hydrogen 1.6 1.8 1.1 0.5
ethane 0.6 0.1 0.1 0.1
and methyl isobutyl ketone (MIBK). The results were
backflush 0.8 0.0 0.3 0.0 used to compare short-term operation with Ni catalyst
Btu-SCF 659 616 676 654 with extended operating time on the Ru catalyst. As shown
in Table 6, these tests further confirmed the long-term
0
Catalyst bed conditions: 350°Cand20-21MPa. Organic loading
of 1.8-2.0 wt % in water as feedstock. activity of Ru catalysts for the TEES processing system.
Two differences among the tests were the concentration
evidence of an effect of the support on crystal growth. The
of the phenol and the catalysts used. The activity of phenol
ruthenium did not agglomerate when supported on the conversion later in the test with ruthenium is comparable
5-alumina (which is also readily converted to bóhmite), as to that from the short-term nickel test, even with the nickel
is shown in Figure 3, or zirconia. Analyses of the ruthenium processing at a lower feed rate (lower liquid hourly space
on zirconia showed no sign of growth as the crystallite size velocity, LHSV). The MIBK also is readily processed in
of much less than 10 nm for several samples recovered TEES. The short-term results with Ni catalyst are
during 6 weeks of operation. Zirconia was relatively stable comparable to the long-term results with Ru catalyst.
in the TEES environment (Elliott et al., 1993a). X-ray The second group of chemicals was the chlorinated
diffraction (XRD) analysis of the zirconia support showed hydrocarbons. Tests were performed with carbon tetra-
no significant growth of the crystallite size after the first
chloride in water, both with and without a nickel catalyst.
2 days of operation.
A preliminary test with TCE (trichloroethylene) is also
Table 4 provides results of TEES processing with several reported here. As shown in Table 7, there was ready
Ru catalysts. The results are for test conditions with high destruction of the chlorinated hydrocarbons at moderate
processing rates at 300 and 350 °C, as suggested by earlier
Table 5. Extended-Time Experimental Results with 5% Ruthenium Catalysts*
catalyst RU/AI2O3 RU/AI2O3 RU/AI2O3 Ru/Zr02 Ru/ZK>2
feedstock p-cresol p-cresol p-cresol phenol phenol
time on stream, h 80 140 230 432 864
gasification of carbon, % 95.4 90.6 97.6 100.2 83
reduction of COD, % 98.5 96.6 99.4 99.4 95.6
feed, L/h 2.0 1.3 1.2 0.038 0.036
LHSV, L L-1 h-1 3.7 2.1 1.9 1.5 1.5
gas yield, L/h 46.3 27.7 29.1 1.1 0.85
gas yield, L/L feed 23.2 21.3 24.2 28.9 23.6
effluent COD, ppm 718 1500 279 365 2300
effluent pH 3.8 3.6 3.9 NA NA
gas composition, vol %
methane 60 58 58 58 47
carbon dioxide 36 38 38 38 45
hydrogen 1.6 1.6 1.2 1.6 2.7
ethane 1.2 1.5 1.2 1.2 1.7
backflush 1.6 1.9 1.3 1.5 3.4
Btu/SCF 681 671 656 654 615
0
Catalyst bed conditions: 350-360 °C and 21 MPa. Organic loading of 1.8-2.0 wt % in water as feedstock.
 Ind. Eng. Chem. Res., Vol. 33, No. 3, 1994 571
Table 6. Extended-Time Experimental Results with Oxygenates*
oxygenate 2% phenol 5% phenol 1.5% MIBK 1.5% MIBK
catalyst 5% RU/AI2O3 Ni-0750 5% RU/AI2O3 Ni-0750
time on stream, h 115 10 195 9

gasification of carbon, % mi 99.5 100.6 98.7

reduction of COD, % 99.98 99.95 99.96 99.90
feed, L/h 2.0 1.6 1.4 1.8
LHSV, L L"1 h"1 3.8 2.7 2.4 3.2
gas yield, L/h 49.6 103.3 31.0 35.1
effluent COD, ppm 7 62 20 40
effluent pH 3.7 3.6 3.6 4.5
gas composition, vol %
methane 54 54 67 66
carbon dioxide 42 40 29 29
hydrogen 1.2 2.9 1.4 3.3
ethane 1.5 1.0 0.6 0.3
backflush 1.9 2.2 1.7 2.4
Btu/SCF 629 620 743 751
*
Catalyst bed conditions: 350-360 °C and 20-21 MPa.
Table 7. Experimental Results with Chlorinated Hydrocarbons
hydrocarbon 500 ppm TCE 200 ppm CCL 200 ppm CCI4 200 ppm CCI4 200 ppm CCI4
catalyst Ni-0750 Ni-0750 Ni-0750 none none
temp, °C 360 350 300 350 300
destruction of chlorinated, % >99 est. >99.99 >99.99 >99.99 >99.99
feed, L/h 1.5 1.5 2.2 1.6 2.3
LHSV, L L-1 h-1 2.5 2.5 4.4 2.4 4.6
effluent level of chlorinated <1 ppm est. <1 ppb <1 ppb 15 ppb 10 ppb
effluent pH 7 5.5 3.8 3.6 2.5
gas composition, vol %
methane 23-51 7-14 5-6 2 1
carbon dioxide 14-1 5-12 1 15 13
hydrogen 25-57 50-65 79- 88 82 85
ethane 0.4 0.2 0.2 0 0
backflush 0.1 0.2 0.5 0.2 0.1

Table 8. Results with TEES Processing of Nitrates in

batch reactor tests (Elliott et al., 1994). The number listed Water at 21 MPa
in Table 7 for destruction of chlorinated species applies
to the main feedstock component; however, it was also HNOs/MeOH NaNOg/MeOH
determined that a chloroform impurity in the carbon temp, °C 350 360

tetrachloride was destroyed to < 1 ppb by TEES processing. COD destruction, % 97.65 99.40
nitrate destruction, % 99.87 99.96
Analysis of the product effluent for chlorinated hydro- feed, L/h 2.5 1.5
carbons was performed by GC with an electron capture LHSV, L L-1 h-1 5.0 2.5
detector by the direct injection method. Little gas was effluent COD, ppm 30 61
produced because of the low amount of organic in the effluent nitrate, ppm 0.2 0.8
feedstock. Because of the limited amount of gas product, effluent pH 7.1 9.7
its recovery and analysis were inconsistent. Hydrogen gas composition, vol %
methane 50 81
production likely results from reaction of product HC1 carbon dioxide 13 1.0
with the metal reactor walls. Acidification of the effluent hydrogen 32 14
was noticeable. A significant degradation (loss of physical ethane 0.0 0.2
integrity) of the alumina-supported nickel catalyst was backflush 0.0 0.2
also observed. In addition, deposits of bóhmite were found nitrogen 5.5 2.8
in the product lines, evidence of the breakup of the catalyst
support material. The noncatalytic results are nearly as For the nitrate experiments, two tests were made in the
good as those with a catalyst. Since no catalyst is needed bench-scale reactor system using the Ni-0750 catalyst.
for carbon tetrachloride and chloroform destruction, Nitrate and methanol were fed as dilute contaminants in
further development of this process application should water. In a nitric acid test, 0.02 M MeOH was in 2:1 molar
not be affected by catalyst stability limitations. ratio to nitrate; in a sodium nitrate test, 0.2 M MeOH was
Nitrates comprised the third group of chemicals. Earlier in 2:1.4 molar ratio to nitrate. Processing conditions and
results are presented in Table 8.
unreported results at PNL suggested the reaction of nitrate
with organics (MeOH and formate) at TEES operating Effective reactions of both the organic and the nitrate
conditions. Those tests were followed by a more definitive were achieved in the presence of the nickel catalyst. No

study of the effect of pH on the reaction of nitrate with
other organics (formate, urea, and glucose) (Cox et al.,
1992). Although the early tests suggested a catalytic effect
in the process, the later studies were done without a
catalyst. More recent results have shown that the catalyst
is unnecessary for the reaction of nitrate with p-cresol up
to the extinction of the nitrate (Elliott et al., 1994). The
continuous reactor tests conducted here provide additional
information on the reaction of nitrates with organics
(methanol) in the presence of a nickel catalyst.
NaO or NO product gases were detected in either test. In
the acid test, nitrogen recovery accounted for 31% of the
nitrate fed with 0.2% nitrate converted to nitrite. In the
base test, only 1.3 % of the nitrate was recovered as nitrogen
gas and no nitrite was detected. The balance of the
nitrogen in the feedstock was unaccounted for in the
products in the two tests.
The catalyst bed was recovered intact in both of these
tests. A loss of crush strength of the catalyst pellets was
evident, as has been typical with the Ni-0750 catalyst when
572 Ind. Eng. Chem. Res., Vol. 33, No. 3, 1994

Table 9. Experimental Results with Industrial Waste Streams1

nylon olive cheese whey, cheese whey,
wastewater wash water delactosed delactosed
catalyst Ni-0750 Ni-0750 Ni-0750 5% Ru on AI2O3
gasification of carbon, % 90.6 41.9 85.0 76.5
reduction of COD, % 99.3 99.9 97.5 93.4
feed, L/h 2.7 1.5 1.3 1.8
LHSV, L L"1 h-1 8.2 2.5 2.3 3.4
gas yield, L/h 5.1 7.3 80.9 96.1
effluent COD, ppm 18 30 3700 8500
effluent NH4"1", ppm 600 350 NA NA
effluent pH 7.2 6.9 7.0 NA
gas composition, vol %
methane 72 62 48 48
carbon dioxide 18 35 48 47
hydrogen 9.1 2.1 1.9 3.8
ethane 0.1 0.5 1.3 1.0
backflush 0.0 0.0 1.3 0.0
Btu/SCF 764 649 555 515
0
Catalyst bed conditions: 350 °C and 21 MPa.

Table 10. Experimental Results with Brewer’s Spent Grain1

tubular only tubular/Carberry
filtered, decanted, liquor, liquor,
28 500 ppm COD 41 000 ppm COD 61 500 ppm COD 65 000 ppm COD
time on stream, h 4 11 4 11

gasification of carbon, % 100.1 67.4 84.7 70.4

reduction of COD, % 97.7 71.2 96.2 82.2
feed, L/h 1.4 1.2 1.2 1.5
LHSV, L L-1 h-1 2.3 2.0 1.3 1.7
gas yield, L/h 26.1 21.5 49.9 54.0
effluent COD, ppm 670 12290 2450 12100
effluent NH4+, ppm 220 285 3000 3000
effluent pH 6.3 NA 7.0 NA
gas composition, vol %
methane 40 35 50 44
carbon dioxide 58 58 45 42
hydrogen 1.0 2.5 4.5 11
ethane 0.6 1.4 0.6 0.9
backflush 0.5 2.6 0.8 1.6
Btu/SCF 434 466 550 548
a
Catalyst bed conditions: BASF catalyst at 350-360 eC and 21 MPa.

used in TEES. Deposits were found in the product line,
indicating movement of the support material. As in the
chlorinated solvent tests, the material was determined to
be bóhmite, based on XRD. Much more deposited ma-
terial was found in the sodium nitrate test, suggesting
that the higher solubility of the bóhmite in the high-pH
water was a factor in the movement.
B. Bench-Scale Reactor Tests with Industrial
Wastes. Bench-scale reactor tests of TEES were also
performed for actual waste materials from commercial
processes. These materials included nylon manufacturing
wastewater, olive wash water, delactosed cheese whey, and
brewer’s spent grain and spent grain liquor. As shown in
Table 9, the two tests with nylon manufacturing waste-
water and the olive wash water had good results in terms
of COD destruction. In both cases, there was a significant
ammonia product in the effluent water from nitrogen in
the organic feed. Further treatment for ammonia removal
may be required or incorporated into the process. One
potential process modification for ammonia removal
compatible with TEES is the reaction with stoichiometric
amounts of nitric acid (Elliott et al., 1994). The low gas
recovery in these tests was probably as a result of carbonate
formation with the ammonia. Alkali in the olive (and
cheese whey) feedstock would also form carbonates at these
conditions and further sequester carbon dioxide product
gas.
In both of the cheese whey tests the result presented
was only maintained for a limited time. Loss of catalyst
activity occurred over a period of hours, apparently as a
result of precipitation of alkali salts in the catalyst bed,
coating the pellets and plugging the catalyst pores.
Whitlockite (CaieMgaHalPOilu) was identified by XRD
in both spent catalysts. The delactosed cheese whey is a
much higher concentration feedstock than tested in TEES
in the past (Elliott et al., 1991, 1993b). The delactosed
feed was actually diluted 1:1 with water for these tests but
was still about twice as concentrated as in earlier tests. In
addition, the delactosed feed would likely be less readily
processed than whole cheese whey, as the more easily
pyrolyzed components, the sugars, have been removed to
a large degree, leaving the fats, proteins, and minerals.
Long-chain hydrocarbons, as would be found in fats, have
been shown to react more slowly in TEES (Elliott et al.,
1994). Therefore, the results reported in Table 9 are not
quite as good as those reported earlier for conventional
cheese whey.
Results from two tests with brewer’s spent grain using
the BASF catalyst are provided in Table 10. The first
test was done in the standard tubular reactor configuration.
In this test, initial conversion was quite good, but a steady
loss of activity was noted. After the test the catalyst bed
was removed for analysis. The lower portion of the bed
contained catalyst pellets coated with a brown material,
which was analyzed by XRD and found to contain, as the
principal crystalline phases, hydroxylapatite [CasiPCU^-
OH] and NÍ3S2. However, the nickel catalyst itself showed
no unusual changes. Reuse of this catalyst bed (the 11-h
data set) exhibited even poorer activity after handling.
Regeneration of the bed was not attempted.
 Ind. Eng. Chem. Res., Vol. 33, No. 3,1994 573

Table 11. Experimental Results in the Scaled-Up Reactor System*

propylene glycol,
sucrose,
5% in water
nylon
5% in water manufacturing wastewater
reactor bed temp, °C H-l 275-339 303-350 309-345 283-333
R-l 326-339 329-345 328-343 330-340
R-2 313-342 336-348 336-347 332-340
gasification of carbon, % 85.3 95.0 93.6 97.8
reduction of COD, % 99.96 99.99 99.99 99.77
feed, L/h 15.4 16.2 22.8 33.0
LHSV, L L-i h-1 1.7 2.2 3.2 4.6
gas yield, L/h 488 669 917 67.9
effluent COD, ppm 20 6 11 8
effluent NH4"1", ppm 290
effluent pH NA NA NA NA
gas composition, vol %
methane 45 64 64 68
carbon dioxide 53 33 34 18
hydrogen 1.2 3.0 2.2 1.2
ethane 1.0 0.3 0.5 0.5
backflush NA 0.0 0.0 NA
Btu/SCF 480 659 662 685
“
Catalyst bed conditions: 20-21 MPa.

The second test used a reactor configuration incorpo-
rating a Carberry, stirred-tank catalytic reactor as a
preheater in combination with the tubular, fixed-bed
catalytic reactor. Similar results were obtained in this
reactor configuration as with the tubular reactor alone.
The use of the preheater to bring the feed to temperature
more quickly and to increase the amount of the tubular
catalyst bed at temperature did not appear to improve the
conversion in the combined reactor system. Catalyst
overload resulting from mass-transfer limitations using
the higher concentration feedstock in the combined reactor
system may explain the poorer results at 4 h time on stream.
Comparing the two data sets, there is an apparent feed
rate effect of lower conversion at faster throughput,
although the effect also correlates with time on stream.
The higher levels of hydrogen in the product gas are a
function of the higher temperature in the catalyst bed
(360 vs 350 s in the tubular only test). There did not appear
to be as quick a loss of activity in the catalyst bed as in
the tubular reactor alone test.
Again, in the combined reactor system, the catalyst bed
had a coating, this time after the 12-h test. In addition,
there was a deposit recovered from the bottom of the
Carberry reactor. The catalyst coating was similar to that
found in the test with the tubular reactor only. The
Carberry deposit consisted of the hydroxylapatite and the
nickel subsulfide, but other phases were also identified by
XRD, including iron phosphate, ammonium iron sulfate,
potassium aluminosilicate, calcium carbonate, calcium
magnesium sulfate, and anorthoclase (an alkali silicoalu-
minate). This deposit in the Carberry reactor suggests a
possible pretreatment step for removal of inorganic
components in the TEES process.
These tests show an improved operability compared with
the earlier test results from the tubular reactor (Elliott et
al., 1991). In the earlier tests with a less active catalyst,
true steady-state conditions were not achieved. The
carbon gasification was far lower than the COD conversion,
suggesting that carbon was building up in the reactor
system as unconverted material. The tests reported here
have a good match between carbon gasification and COD
conversion with the small difference probably resulting
from soluble carbonates formed from the alkali and
ammonia byproduct. The higher concentration feedstocks
reported in Table 10 have COD loadings more represen-
tative of the industrial situation.
C. Scaled-Up Reactor Tests with Several Feed-
stocks. The scaled-up reactor system (SRS) has been
used to show the effects of scale-up on the reactor design
and to demonstrate the complete processing system. The
system has been run with two chemical model compounds,
sucrose and propylene glycol, and the nylon manufacturing
wastewater stream. Sucrose and propylene glycol were
used as models because they are nonhazardous and have
been shown to readily react in earlier TEES tests (Elliott
et al. 1989,1994). The Ni-0750 catalyst was used in all the
SRS tests, as it was reduced and stabilized and did not
require a prereduction step in the reactor.
Sucrose was used as a model carbohydrate. It exhibited
ready pyrolysis and decomposition in the tube-in-tube heat
exchanger with formation of a char material found in the
heat exchanger after the tests. During an initial run, the
preheater vessel was packed with alumina balls and after
the test was found to have a significant amount of
carbonaceous particulate residue mixed with the balls.
Similar material also was collected from the bottom of the
R-1 catalytic reactor. In the subsequent test (whose results
are presented in Table 11) the preheater, as well as the
reactors, was packed with nickel catalyst. As a result, the
preheater contained essentially no carbon deposits except
for particulate formed in the heat exchanger which
collected at the inlet to the preheater. No particulate
collected at the inlet of the R-l reactor.
The particulates produced in both tests were similar—a
pyrolytic, condensed tar phase from the carbohydrate (62-
69% carbon and 4.2-4.4% hydrogen). Carbon buildup
on the catalysts in the R-l and R-2 vessels was almost
negligible compared with the initial carbon content of the
catalyst as manufactured. There was a noticeable amount
of carbon on the catalyst in the preheater in the second
test, but the amount lessened over the length of the vessel.
Hydrogen content in the used catalysts was higher as a
result of the increased hydroxyl content in the support by
the reaction of the alumina to the hydroxide.
Overall, sucrose was an easily treatable feedstock in
TEES. High conversion of COD was achieved at moderate
feed rate. The gas composition was as expected with little
high molecular weight hydrocarbon byproduct gases. The
mismatch of COD conversion with carbon gasification
probably resulted from the carbon deposition in the reactor
system. The processing results compare very well with
the earlier reported test result (Elliott et al., 1991), which
was achieved with a less useful catalyst in the TEES
processing environment.
Propylene glycol was a very reactive feedstock at these
conditions, as had been seen earlier in batch reactor tests
574 Ind. Eng. Chem. Res., Vol. 33, No. 3, 1994
Table 12. Heat Exchanger Data from the SRS
flow effluent__feed feed AT/ effective
approach
rate, hot, cold, cold, hot, effluent heat cold side,
L/h °C °C °C °C , % recovery, %
16.2 324 53 31 288 94.8 87.7
22.8 336 63 35 309 100.4 91.0
30.0 330 68 33 283 95.4 84.2
33.6 325 75 36 277 96.4 83.4
(Elliott et al., 1994). The results of the SRS tests are given
in Table 11. Very clean process effluent with good quality
gas byproduct was produced at two different processing
rates. The total time on stream with this feed was 13 h
with one minor stoppage because of a filter plug.
The nylon manufacturing wastewater was chosen for
demonstration of TEES on an industrial waste stream. As
shown in Table 11, this feed could be processed at high
feed rates without difficulty. The reactor system was on
stream for 14 h straight without shutdown, with two barrels
of waste processed during that period. The principal
organic components in this waste stream were hexameth-
ylenediamine and cyclopentanone. Clean process effluent
was achieved along with a good quality gas product. There
was an ammonia residual in the water, which may require
further treatment in actual applications.
In these tests, the tube-in-tube heat exchanger gave good
performances. The heat exchanger was the major process
difference, besides scale, between the bench-scale unit and
the SRS. Although there were significant losses of heat
in the transfer lines between the reactors and the heat
exchanger (10-20 °C), inside the exchanger there was an
effective exchange of heat from the reactor effluent to the
feedstock. Data for several flow rates are given in Table
12. The effluent from the exchanger was adequately cooled
to eliminate the need to operate the air-blown cooler. The
exchanger recovered >90% of the heat in the effluent
stream with a temperature approach to within <30 °C.
Conclusions
The TEES process has now been demonstrated in
continuous-feed, fixed-bed catalytic reactor systems on
three scales of operation ranging from 0.03 to 33 L/h. The
systems have been operated with consistency at conditions
of 350 °C and 21 MPa. The demonstrated effective heat
recovery in a tube-in-tube heat exchanger should be
beneficial for economical operation of TEES. Aqueous
effluents with low residual COD (<100 ppm) and a product
gas of medium-Btu quality have been produced. Catalysts
have been demonstrated for up to 6 weeks of operation
with reasonable stability. Ruthenium appears to be a more
stable catalyst than nickel in this respect. A wide range
of feedstocks have been processed effectively, including
oxygenates, chlorinated hydrocarbons, nitrates, food pro-
cessing wastes, and organic chemical manufacturing
wastewaters. Reduced nitrogen in these feedstocks is often
converted into ammonia, which is stable at TEES con-
ditions. Inorganic components in these feedstocks have
sometimes been deposited in the catalyst bed when
compounds were produced at concentrations above their
solubility limits.
Acknowledgment
The authors acknowledge the support for this research
provided by the U.S. Department of Energy/Energy
Efficiency and Renewable Energy through its Office of
Industrial Technologies and the program manager, Mr.
Stuart Natof. Some of the work reported was performed
under a separate DOE program in the Environmental
Restoration and Waste Management/Office of Technology
Development. Pacific Northwest Laboratory is operated
for the U.S. Department of Energy by Battelle Memorial
°C Institute under Contract DE-AC06-76RLO 1830. We
22
gratefully acknowledge the catalyst samples received from
28 BASF, United Catalysts, Grace-Davison, and Engelhard
35 Corporation and the industrial participation of DuPont-
39 Fibers, the Anheuser Busch Companies, Sheffield Prod-
ucts, and Oberti Olive Company, who provided samples
of process wastes for testing. We have also had assistance
from numerous technicians and engineers of PNL who
operated the high-pressure reactor systems, but partic-
ularly Gary Neuenschwander and Todd Hart. Finally,
we acknowledge the support of our licensee, Onsite*Ofsite,
Inc., in interacting with the industrial participants.
Literature Cited
American Public Health Association. Standard Methods for the
Examination of Water and Wastewater, 18th ed.; APHA
Publishing: Washington, DC, 1992.
Baker, E. G.; Sealock, L. J., Jr. Catalytic Destruction of Hazardous
Organics in Aqueous Solutions, PNL-6491-2; Pacific Northwest
Laboratory: Richland, WA, 1988.
Baker, E. G.; Butner, R. S.; Sealock, L. J., Jr.; Elliott, D. C.;
Neuenschwander, G. G. Thermocatalytic Conversion of Food
Processing Wastes: Topical Report FY 1988, PNL-6784; Pacific
Northwest Laboratory: Richland, WA, 1989a.
Baker, E. G.; Sealock, L. J., Jr.; Butner, R. S.; Elliott, D. C.;
Neuenschwander, G. G.; Banns, N. G. Catalytic Destruction of
Hazardous Organics in Aqueous Wastes: Continuous Reactor
System Experiments. Hazard. Waste Hazard. Mater. 1989b, 6
(1), 87-94.
Cox, J. L.; Lilga, . A.; Hallen, R. T. Thermochemical Nitrate
Reduction, PNL-8226; Pacific Northwest Laboratory: Richland,
WA, 1992.
Elliott, D. C.; Sealock, L. J., Jr.; Butner, R. S.; Baker, E. G.;
Neuenschwander, G. G. Low-Temperature Conversion of High-
Moisture Biomass: Continuous Reactor System Results, PNL-
7126; Pacific Northwest Laboratory: Richland, WA, 1989.
Elliott, D. C.; Neuenschwander, G. G.; Baker, E. G.; Sealock, L. J.,
Jr.; Butner, R. S. Low-Temperature Catalytic Gasification of
Wet Industrial Wastes, FY1989-1990InterimReport,PNL-7671·,
Pacific Northwest Laboratory: Richland, WA, 1991.
Elliott, D. C.; Sealock, L. J., Jr.; Baker, E. G. Chemical Processing
in High-Pressure Aqueous Environments. 2. Development of
Catalysts for Gasification. Ind. Eng. Chem. Res. 1993a, 32,1542-
1548.
Elliott, D. C.; Butner, R. S.; Baker, E. G.; Sealock, L. J., Jr. Bench-
Scale Reactor Tests of Low-Temperature, Catalytic Gasification
of Wet Industrial Wastes. J. Solar Energy Eng. 1993b, 115 (1),
52-56.
Elliott, D. C.; Sealock, L. J., Jr.; Baker, E. G. Chemical Processing
in High-Pressure Aqueous Environments. 3. Batch Reactor Process
Development Experiments for Organic Destruction. Ind. Eng.
Chem. Res. 1994, preceding paper in this issue.
Pearce, R.; Patterson, W. R. Catalysis and Chemical Processes, John
Wiley & Sons: New York, 1981, pp 14-15.
Sealock, L. J., Jr.; Elliott, D. C.; Butner, R. S.; Neuenschwander, G.
G. Low-Temperature Conversion of High-Moisture Biomass:
Topical Report January 1984-January 1988, PNL-6726; Pacific
Northwest Laboratory: Richland, WA, 1988.
Sealock, L. J., Jr.; Elliott, D. C.; Baker, E. G.; Butner, R. S. Chemical
Processing in High-Pressure Aqueous Environments. 1. Historical
Perspective and Continuing Developments. Ind. Eng. Chem. Res.
1993, 32, 1535-1541.
Received for review May 10, 1993
Revised manuscript received October 5,1993
Accepted November 4, 1993®
®
Abstract published in Advance ACS Abstracts, January 15,
1994.