r .

~ ~ ~~~~~

Quarterly Technical Report #11* December 1994

ADVANCED SEPARATION TECHNOLOGY FOR FLUE GAS CLEANUP

Abhoyjit S. Bhown Dean Alvarado Neeraj Pakala Susanna Ventura

SRI International 333 Ravenswood Avenue Menlo Park, CA 94025

SRI Project No. PYU-3501

Prepared for:

U.S. DEPARTMENT OF ENERGY Pittsburgh Energy Technology Center P.O. Box 10940

Pittsburgh, PA 15236-0940 MS 921-118

Attn:

DOE

Document Control Center

Contract No. DE-AG22-92PC91344

Kamalesh K. Sirkar Sudipto Majumdar Debabrata Bhaumick

New Jersey Institute of Technology University Heights Newark, NJ 07102

DlSTRlBUflON OF THIS DOCUMENT IS MASTER

CONTENTS

INTRODUCTION .................................................................................... 1 SUMMARY OF QUARTERLY PROGRESS .................................................. TASK 8: INTEGRATED NOx LIFE TESTS ................................................ 4

3

TASK 9: PERFORMANCE OF SCALABLE MODULES ................................. 8

This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsi- bility for the accuracy, completeness. or usefulness of any information, apparatus. product, or process disclosed, or represents that its use would not infringe privately owned rights. Refer- ence herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recom- mendation. or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

i

DlSCLAl MER

Portions of this document may be illegible in electronic image products. Images are produced from the best available original document.

I

FIGURES

1 Simultaneous NOx absorption and desorption tests using Co(I1)-phthalocyanine (30 hours of operation) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simultaneous NOx absorption and desorption tests using Co(I1)-phthalocyanine (40 hours of operation) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

5

2 6

3 Cross-flow pattern of gas and liquid flow in a rectangular hollow fiber cassette (module). ................................................. ........... 9

TABLES

1 Project tasks and schedule ................................................................ 2 2 Discussion with membrane manufacturers.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

ii

INTROD-UCTION

The objective of this work is to develop a novel system for regenerable So;2 and NO, scrubbing of flue gas that focuses on (a) a novel method for regeneration of spent So;! scrubbing liquor and (b) novel chemistry for reversible absorption of NOx. In addition, high efficiency hollow fiber contactors (HFC) are proposed as the devices for scrubbing the So;! and NO, from the flue gas. The system will be designed to remove more than 95% of the SOx and m r e than 75% of the NO, from flue gases typical of pulverized coal-fired power plants at a cost that is at least 20% less than combined wet limestone scrubbing of SO, and selective catalytic reduction of NOx. In addition, the process will make only marketable byproducts, if any (no waste streams).

The major cost item in existing technology is capital investment. Therefore, our approach is to reduce the capital cost by using high efficiency hollow fiber devices for absorbing and desorbing the So;! and NO,. We will also introduce new process chemistry to minimize traditionally well- known problems with S@ and NO, absorption and desorption. For example, we will extract the So;! from the aqueous scrubbing liquor into an oligomer of dimethylaniline to avoid the problem of organic liquid losses in the regeneration of the organic liquid. Our novel chemistry for scrubbing NO, will consist of water soluble phthalocyanine compounds invented by SRI and also of poly- meric forms of Fe* complexes similar to traditional NO, scrubbing media described in the open literature. Our past work with the phthalocyanine compounds, used as sensors for NO and NO;! in flue gases, shows that these compounds bind NO and NO2 reversibly and with no interference from 02, C02, SO2, or other components of flue gas.

The final novelty of our approach is the arrangement of the absorbers in cassette (stackable) form so that the NOx absorber can be on top of the SO, absorber. This arrangement is possible only because of the high efficiency of the hollow fiber scrubbing devices, as indicated by our preliminary laboratory data. This cassette (stacked) arrangement makes it possible for the So;! and NO, scrubbing chambers to be separate without incumng the large ducting and gas pressure drop costs necessary if a second conventional absorber vessel were used. Because we have separate scrubbers, we will have separate liquor loops and deconvolute the chemical complexity of simultaneous S02/NOx scrubbing.

We will conduct our work in a 60-month period (5/92 to 4/97), encompassing 16 tasks (Table l), beginning with studies of the fundamental chemistry and of the mass transfer character- istics of small HFC modules in the laboratory. We will then examine the most favorable method of So;! liquor regeneration, determine the ability of the HFC devices to withstand particulate matter,

1

and examine the behavior of scalable modules. In the final 15 months of the program, we will determine the fundamental mass transfer behavior of a subscale prototype system. Based on these data, a computational design model will be devised to guide further scaleup efforts that may follow. ,.

Table 1

PROJECT TASKS AND SCHEDULE

Task Number

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

Project Definition

Capacity, Reversibility and lifetime

Chemical Synthesis

SO2 Scrubbing with HFCs

NOx Scrubbing with HFCs

SO2 Liquor Regeneration

Particle Deposition

Integrated NOx Life Tests

Scalable Modules

Computational Model

Construction of Subscale Prototype

Operation of Subscale Prototype

Refinement of Computational Model

Economic Evaluation

Reporting

Chemical Synthesis for Process Scale-up

Duration

5/92 - 8/92 7/92 - 6/94 7/92 - 6/94 7/92 - 9/93

2/93 - 3/94 7/93 - 9/94

8/93 - 9/94 8/94 - 4/95

8/94 - 495 1 1/94 - 7/95 8/95 - 1/96 2/96 - 4/97 9/96 - 497

Various

5/92 - 4/97 5/94-1196

2

SUMMARY OF QUARTERLY PROGRESS

During the fourth quarter of 1994, we continued work on Task 8. We also began working on Task 9.

In Task 8, we completed the integrated experimental setup for testing the efficacy of C o o - phthalocyanine solution for NOx absorptiorddesorption. Over 40 hours of life testing operation, the Co@)-phthalocyanine solution exhibited stable and good performance.

In task 9, we carried out preliminary design calculations to determine the performance of a rectangular module for Sa absorption. The calculations indicate that a 0.8' x 1' x 1" size module with 240 pm fibers will effectively function as an S @ absorber. Also, OUT efforts to locate a rectangular membrane module manufacturer for this phase of the project are presented. Based on these consultations with module manufacturers, we placed an order for two modules with Sectec, Inc., Livermore, CA.

3

TASK 8: INTEGRATED NO, LIFE TESTS

Liquid scrubbing systems for NO, have traditionally been plagued by degradation of performance over time with subsequent need for convoluted liquor regeneration schemes. Therefore, it is essential to determine whether or to what extent there is any loss with time in the performance of the NOx absorptioddesorption system. In Task 2, we have showed C o o phthalocyanine as a promising candidate for NO, absorption in the presence of e. In Task 5, we also demonstrated the superior mass transfer characteristics of Co(II)-phthalocyanine scrubbing system in a 300 fiber HFC. Consequently, the objective here is to devise a continuously operating system for determining if the NO, absorption/desorption chemistry has the potential to be commercially robust.

In the last quarter, we reported that the problems related to the desorption step were fixed. During this reporting period, we continued collecting data using combined absorption-desorption apparatus. 100 mM Co(1 1)-phthalocyanine solution was recirculated between 300 fiber HFC and the desorption flask (250 ml) at 3 d m i n flow rate. The flue gas flow rate through the lumens of HFC was 100 sccm. This feed gas stream had a composition of 500 ppm NO, 4.5% 02 and balance N2. Pure N2 (sweep gas) after humidification was sparged into the liquid, contained in the desorption flask. The sweep gas stream flow rate is maintained at 300 sccm. This high flow rate compared to feed is used in order to create higher driving force for NO between liquid phase and sweep gas phase.

The outlet gas streams from both the 300 HFC and the desorption the flask are monitored for NO concentration and the results are plotted in Figure 1. The evaporation losses that occur due to repeated exposure of liquid to the gas stream in the HFC as well as in the desorption flask are monitored by watching the liquid level in the flask. The losses were noticed to be small. However, the fluctuation in data points could be attributed somewhat to this concentration changes due to evaporation of water.

During the middle of the quarter, we continued further testing on absorptioxddesorption characteristics of Co(1 1)-phthalocyanine solution. Figure 2 shows an additional 10 hours worth of data on simultaneous absorption/desorption of NO. Despite some fluctuations, both absorption and desorption rates are approximately equal to one another. Figure 2 also indicates the steady operation owing to small deviations from flatness of the data points.

4 I I 1 1 I 1

a0 r 0

w 7- v

m 0

LT 8

1 2 1

I 1 I 1 1

0 5 10 15 20 25 30 35

TIME (h) CAM-3501-96

Figure 1. Simultaneous NO, absorption and desorption tests using Co(ll) - phthalocyanine (30 hours of operation).

5

w

4

0 10 20 TIME (h)

30 40

CAM-3501-97

Figure 2. Simultaneous NO, absorption and desorption tests using Co(ll) - phthalocyanine (40 hours of operation).

6

However, fdrther continuation of experiments were hampered due to the NO, analyzer. The analyzer drifts by about 15% per an hour leading to false readings. Therefore, we discontinued NO, life tests. Our attempts to assess the problems with regard to calibration of the equipment were not successful. Due to continued fluctuations in the reading, we have sent the equipment to the regional manufacturer in Los Angeles. We hope to continue to do further experiments on the absorptioddesorption characteristics of Co(l1)-phthalocyanine solution once we receive the equipment.

TASK 9: PERFORMANCE OF SCALABLE MODULES

Because of the need for billions of (approximately 30-cm long) fibers to treat the flue gas from a 500 MW(e) plant, it is critical to establish ;he mass transfer characteristics of a module that can be scaled up to a prototypical size. To appreciate this point, one must recognize that approximately 250,000 modules of 2" diameter would be required to provide one billion fibers. Such an arrangement would provide a ducting, plumbing, and maintenance nightmare in a full- scale plant and clearly would not be economical or workable. A new design concept, such as rectangular modules, is needed for a full-scale plant. Therefore, the objective of Task 9 is to develop the mass transfer fundamentals of rectangular modules.

The term "scalable module" is an important concept because it influences our ability to think about an eventual application of HFC devices to a 500 MW(e) power plant. We use the term to mean a module that exhibits the important phenomena that would appear in modules of the 500 MW(e) plant. For example, mass transfer in the rectangular (square) module will be different than that in the cylindrical modules because the flow pattern in a large-scale rectangular module absorber or liquid-liquid extractor will be crossflow, not countercurrent (Figure 3). Probably more importantly, flow distribution and liquid pressure drop behavior will be much different in a large "scalable" module than in our laboratory modules. Therefore, Task 9 is designed to obtain the fundamental mass transfer behavior, including issues of flow distribution and pressure drop, on modules that reflect these basic features of large-scale modules with gas and liquid flow rates in the range of 5-10 cfm and 1-5 Vmin respectively.

We began carrying out preliminary evaluation on the efficacy of the rectangular modules. The assumptions involved in the calculations and the fiber dimensions are given below.

(1) Rectangular module dimensions: 0.8' x 1' x 1".

(2) Packing density of fibers = 40%.

(3) Fiber dimensions = ID - 240 pm; OD - 300 pm.

(4) Maximum pressure drop along the length of the fiber = 10" water.

(5) Gas flow through the fiber is laminar.

Number of fibers that can be placed in rectangular

(Cross sectional area of shell) Packing density Cross sectional area of fiber based on OD geometry =

= 43,000

8

c

Vertical Bundles of Hollow Fibers

Inlet Scrubber

Liquor

Inlet G a s Flow

Potting Compound

Hollow Fiber

CM-360583-52

Figure 3. Cross-flow pattern of g a s and liquid flow in a rectangular hollow fiber cassette (module).

9

where

x A P 4 128 pg L as flow per fiber =

AP = Pressure drop along the length of fi&r = 24866 dynes/cm2 di = ID of fiber = 240 pm pg = Viscosity of gas = 1.7 x 10-4 gm/cmsec L = Length of fiber = 24 cm.

upon substitution of above parameter values into Eq (I), we get

Qg/fiber = 3.0 cc/min

Gas flow through the module, Qg = Qg/fiber (total number of fibers)

= 125 L/min

As this gas flow rate is within the range mentioned above, it is of further interest to see the gas scrubbing capacities of the module. From experimental data obtained so far (see Task 14, Quarterly Technical Report #9), the mass transfer correlation is given as:

KO, = 0.079257 Reo33211

where Re is the gas side Reynolds number and is given as

Here pg indicates gas density.

From Eqs. (2) and (3), we obtain &g to be 0.23 cm/sec.

Equilibrium calculation for N e So3/S@ system indicate that the partial pressure of So;! @*) in equilibrium with the liquid is very small compared to the actual partial pressure @) and therefore mass balance yields:

In = KOg xd, LN/Qg PL

Here Po and PL indicate the partial pressures of SO2 in the entering and exiting gas streams, respectively . From Eq. (4), we estimate that the Sa removal rate is more than 99%. In summary, these approximate estimations indicate that a rectangular module with dimensions of 0.8' x 1' x 1" containing 240 pm ID fibers will effectively function as an SO2 scrubber.

10

(4)

Thus far, we have approached ten membrane companies to solicit their interest in this project. Summary of events are given in Table 2. Based on these consultations with different membrane module manufacturers, we identified that Separation Equipment Technologies, Inc. (Setec), Livermore, CA, as a suitable supplier for the rectangular modules. Therefore, we placed an order for two modules with Setec. The characteristics of the hollow fibers and the anticipated performance of the resulting modules will be reported in the coming months.

Manufacturer Hoechst-Celanese Charlotte, NC

TNO, Netherlands

Compact Membrane Systems, Inc. (CMS) Wilmington, Delaware

LSR TecQnologies Massachusetts

Innovative Membrane Systems, Inc. (IMS) Massachusetts W.L. Gore & Associates, Inc. - Elkton, Maryland

Mitsubishi, Japan

Toyobo, Japan Nitto Denko, Japan

Setec, Inc. Livermore, CA

Table 2.

Date(s) of C o r r e w o n d e n c e

1990 - Sept, 1994

Sept, 1994 - Jan, 1995

Nov - Dec, 1994

Nov - Dec, 1994

Nov - Dec, 1994

July - Sept, 1994

Nov, 1994 Nov, 1994 Nov, 1994

Nov, 1994 - present

Discussions With Membrane Manufacturers

Comments Hoechst-Celanese was initially the membrane company that showed interest in commercializing these modules. For business reasons, they decided not to pursue it further. Already makes rectangular modules. Ideal candidates. Prototype module would cost approx. $30,000 for lab-scale. Already trying to commercialize. Rectangular lab-scale module is priced at $36,000.

Rectangular lab-scale module is priced at $9 , 000.

No business interest in fabrication of microporous fiber modules.

Interested to use their PTFE (Teflon) fibers; Gore had also received the cost estimates under non-disclosure agreement. Manufactures microporous hollow fibers. Same as above. No good fibers for the project.

Rectangular lab-scale module is priced at $8,000.

End R e s u l t Not a viable option.

Too expensive, but may be viable later.

Too expensive, but may be viable later.

cu Does not have reliable source of fiber supplier, no prior expertise with microporous fibers and also modules are expensive.

c

No Response. Probably because the process becomes too expensive with PTFE fibers.

No response. Same as above.

Prior expertise with microporous polypropylene fibers. Placed an order for 2 modules.

Co(I1)-phthalocyanine (30 hours of operation)Co(I1)-phthalocyanine (40 hours of operation)

hollow fiber cassette module).1 Project tasks and scheduleDiscussion with membrane manufacturers