# I

i i

TECHNICAL REPORT September 1 through November 30,1994

Project Title: COMPOSITION MODIFICATION OF ZINC TITANATE SORBENTS . FOR HOT GAS DESULF'URUATION

DOE Cooperative Agreement Number: DE-FC22-92PC92521 (Year 3)

Principal Investigator: .

Co-Principal Investigator:

Project Manager:

ICCI Project Number: 94- 112.2A-2M James H. Swisher, Southern Illinois University at Carbondale Ranajit K. Datta, Virginia Poly technic Institute and State University Franklin I. Honea, Illinois Clean Coal Institute

ABSTRACT

For new coal gasification systems, zinc titanate sorbents are being developed to remove sufir fiom the hot product gas prior to its use in combined cycle turbines and high temperature fbel cells. Although most of the properties of these sorbents are very attractive, there are still concerns about durability over many sulfidation-regeneration cycles and zinc losses due to vaporization. Doping the zinc titanate with other metal ions could deviate both concerns, which are the objectives of this project. During the k s t quarter of effort, several sorbent formulations were prepared and testing begun. The dopants presently under study are Ni, Cr, Cu, and AI. Crush strength results obtained to date show that Ni and Cu dopants have a large positive effkct, while Cr gives a small improvement. Measurements were also made of sufir capacity and suKdation rates with a thermogravimetric analyzer. Of the three formulations, only the one containing Cr had a high sukr capacity. X-ray measurements will be relied upon heavily to obtain an understanding of solubility effects and sulfidation mechanisms. Screening experiments will continue on the doped sorbents mentioned above next quarter, and Mg will be studied also.

"U.S. DOE Patent Clearance is NOT required prior to the publication of this document."

EXECUTIVE SUMMARY

Nominal sorbent

Sulfbr sorbents which contain zinc as the primary desufirizing agent continue to show promise for coal gasification applications. Scale-up tests are now in progress for Integrated Gasiication Combined Cycle (IGCC) systems. There are still concerns, however with sorbent durability over many sulfidation-regeneration cycles and with vaporization losses of zinc. Therefore research and development directed at these potential problems is continuing. Zinc titanate sorbents are among the few that have reached the stage of large scale testing.

Bulk Porosity Porosity den& (assuming no so~ution (asslmring c~mplae

An approach that has received little attention to date for improving the properties of zinc titanate is to dope the material with other metal ions. In principle, it should be possible to both strengthen the material and reduce zinc vaporization losses in this way. In the portion of the literature review completed to date, not much usefbl information was uncovered on which dopants should be most effective. Therefore, principles of ceramic crystal chemistry and an experimental screening study are being used in the current project to select and evaluate the effects several dopants have on Zinc titanate properties. More specifically, nickel, chromium. copper, magnesium, aluminum and perhaps other dopants will be studied.

During the first quarter of effort, alternate approaches were taken in initial experiments . carried out at Southern Illinois University (STUC) and Virginia Tech (VPI). At SIUC, pellets

containing Ni Cr, and Cu were prepared and evaluated by measuring crush strength, porosity, and reactivity in a thermogravimetric analyzer (TGA). The compositions of the pellets were calculated to produce Zn,.QMo.4Ti04, where M is Ni Cr, or Cu, assuming al l of the dopant was incorporated into the titanate phase. A summary of the physical properties is given in Table I. The porosity values were all in the range from 45 to 56%, but there were much larger differences in crush strength. The formulations containing Ni and Cu had crush strengths more than twice that of a baseline material containing excess TiO, but no dopant.

Table 1. Porosity and Crush Strength Results for Doped ZincTitanate Sorbents.

Zn1.6Cr0.4Ti04

Zn1.6Cu0.4Ti04

2.25 56.05 53.22

2.79 49.04 45.83

of dopant) solution of dopant) I [g/cmiI I [%I I [%I I composition

Crush strength cN/mm3

55.8

113.7

65.3

137.5

DISCLAIMER

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, make any warranty, express or implied, or assumes any legal liability or responsibility 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. Reference herein to any specific commercial product, process, or service ,by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, 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.

DISCLAIMER

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

An example of the TGA results is given in Fig. 1 for the sorbent containing Ni. The initial weight loss is believed to be due to a reduction reaction which produced metallic Ni. The sample then sulfidized to a saturation level, followed by regeneration back to its original weight. The weight gain on &dation, a measure of sufir capacity, was much less than that expected for the conversion of Zn to ZnS and Ni to Ni3S2. Of the sorbents tested to date, only the one containing Cr had a high sulfur capacity.

In the research started at VPI, 30 specimens were prepared to determine the solubilities of Ni Cry and AI in Zn2Ti0,. X-ray measurements needed for the solubility determinations have been delayed because of an equipment failure. Progress was also made in setting up other equipment for the project.

I.

..

0 40 80 120 160 200 240 i ime [rnin]

Figure 1. Sulfidation of Zn,.,N&.,TiO, in 0.9% H,S, 92.4% H2 and 6.7% N2 at 650" C and Regeneration in 5% 0, and 95% N2 at 650" C (one cycle).

..

1

OBJECTIVES

The general objective of the project is to improve the engineering properties of zinc titanate sorbents for the removal of H,S from coal gasification product gases. These improvements will be made by systematicdy adding dopants, such as Ni, Cu, Cr, Mg, and AI, to zinc titanate formulations. The main tasks designed to meet the objectives are sorbent preparation, solubility measurements, and bench-scale testing of sorbent pellets. Expected results include both engineering property characterization and an understanding of the crystal chemistry of the doped sorbents.

INTRODUCTION AND BACKGROUND

All modem coal combustion and coal conversion processes must have provisions to meet strict specifications on emissions of H2S and SO,. For many systems, removal of dfk-bearing species fiom hot gases is also needed to prevent corrosive damage to materials and components. Both requirements apply to Integrated Gasification Combined Cycle (IGCC) and Molten Carbonate Fuel Cell (MCFC) systems. An important milestone in the development of IGCC systems was successful completion of the Cool Water Coal Gasification Program (Douglas, 1990). In this program, the fuel gas was cooled down for H2S removal, then reheated before injection into the turbine combustor. To avoid the associated efficiency loss with cooling of the fuel gas, most of the'active development projects now include hot gas desulfbrization with regenerable sorbents.

Many metals, oxides, and other compounds have been evaluated for their ability to remove K2S f?om hot, coalderived gases. In recent years, one of the classes of compounds that has been studied extensively and continues to show promise is zinc titanate (Swisher, O'Brien, and Gupta, 1994; Gupta, Gangwal and Jain, 1993; Gupta and Gangwal, 1993; Datta, 1994; Mei, et al., 1993; Ayala, et al., 1993; and Grindley, 1990). The status of the technology has matured to the point where large scale testing in process development units and pilot plants is in progress (Bevan, 1994; Konttinen, et al., 1994).

There are still a few concerns with zinc titanate sorbents which are being investigated in current projects. The most important of these is durability over many sulfidation cycles in fixed and fluidized bed reactors. There is a lesser concern with vaporization losses of zinc. In this regard, current sorbent formulations are adequate unless the sulfidation reaction is carried out above a temperature of approximately 600" C.

In the project described here, the research is directed toward both problem areas. By the incorporation of a third metal ion into the zinc titanate crystal structure, the thermodynamic activity and vapor pressure of zinc can be reduced. Also, the dopant may improve the strength and durability of the sorbents. Scanty information has been found in the literature on this subject, so the present work will be pioneering in nature.

2

During the past quarter and part of the second quarter, alternate but complementary approaches are being taken at SWC and VPI to screen several dopant possibilities for their potential. At SIUC, a standard dopant concentration was selected, and measurements are being made on crush strength, bulk density, and reactivity in a thermogravimetric analyzer (TGA). At VPI, the solubility of the dopants in the zinc titanate structure is being measured.

EXPERIMENTAL PROCEDURES

In the experiments done at SIUC, cylindrical pellets were prepared using the same procedures as in prior ICCI projects. Reagent-grade oxide powders were blended in a dilute solution of starch in water. Excess water was removed until the material had a consistency like toothpaste, then it was extruded through a 1 cm dia. metal tube and cut to lengths of approximately 1.5 cm. After drying krther, the green pellets were sintered for three hours at 900°C.

The weights of the oxide powders were calculated to be consistent with final compositions corresponding to Zn,.&.,TiO, (atomic basis), where M denotes the dopants, Ni, Cr, and Cu. It was not known in advance to what degree the dopants would be incorporated into the titanate. Thus the final pellets could contain unreacted NiO, Cr,O,, and CuO. The formula above also assumes Cr substitutes for Zn as a divalent ion, which may be incorrect. To obtain answers to these questions, a sample fiom each formulation was submitted to VPI for x-ray diffraction analysis.

Further characterization of the as-sintered pellets consisted of compressive strength measurements, and determinations of the bulk densities, from which the porosities were calculated. The remainder of the effort at SIUC consisted of experiments c d e d out in a thermogravimetric analyzer (TGA) and a controlled-atmosphere tube &mace to study the sulfidation-regeneration behavior and the composition of the &de phases formed. For these experiments, pellets were crushed to a particle size of 2' to 4 mm.

In the work initiated at VPI, reagent-grade oxide powders were used to prepare zinc titanate specimens doped with Ni, Cr, and Al. The powders were mixed in appropriate proportions under acetone in a mortar and pestle and air dried. The process of mixing and drying was repeated 3 to 4 times. One gram from each of the 15 gram formulations was fired in an open alumina crucible. A total of 30 specimens were prepared with different dopant concentrations. The sintering time was 24 hours, and the sintering temperatures were 1000 and 1100°C.

RESULTS AND DISCUSSION

A summary of bulk density, porosity, and crush strength results obtained to date is given in Table I (see Executive Summary). For comparison purposes, data are included for a formulation studied in last y d s ICCI project, where excess TiO, was used instead of doping.

3

For the doped materials, two models were used in the calculation of porosity. In one, the theoretical density was calculated assuming no reaction of the NiO, Cr,O,, and CuO occurred. In the other, complete reaction to form Zn,,&,4Ti04 was assumed. It can be seen that the two models gave differences of only 3 to 4 percent. These calculations will be repeated after x-ray data are available to show the extent to which the dopants were incorporated into the titanate. The porosities for the doped materials can be seen to be a little higher, with the possible exception of the Cu-doped sorbent, than for the undoped formulation..

There is a much larger variation in the crush strength results. The Cr-doped material was a little stronger than the undoped material, but the formulations containing Ni and Cu were more than twice as strong. Thus the mechanical property results are very encouraging. There is also a possibility that the crush strengths would be still higher if the particle sizes of the initial oxide powders were larger and if a higher sintering temperature were used. These differences produced a high crush strength in the undoped zinc titanates evaluated last year.

The TGA tests for chemical reactivity produced rather complicated results. Fig. 1 (see Executive Summary) shows TGA results for a specimen doped with Ni. Sulfidation was carried out in a H2-H2S-N2 mixture at 650" C, and regeneration was carried out in an 02-N2 mixture at the same temperature. When the sulfidizing gas was admitted at 50 min. on the time scale, there was a rapid initial loss in weight, which was probably caused by a reduction reaction to produce metallic Ni. Between reaction times of 65 and 140 min., a weight gain attributed to sulfidation was observed. The regeneration reaction was started at a time of 140 min., producing a loss in weight to its original value. The theoretical weight gains for the conversion ofZnO to ZnS and NO to Ni3S2 are 10.7 and 0.9%, respectively. Thus complete sulfidation corresponds to 11.6% or 8.2 mg. The experimental weight gain from the original weight was only 2.4% or 1.7 mg.

Fig. 2 shows data for a two-cycle test on the same material. The only difference in test conditions was that the regeneration temperature was increased from 650 to 850" C to try to reduce the time for regeneration. For the first cycle, the'weight gain on sulfidation was 4.5, nearly twice that of the test described in Fig. 1. In the second cycle in Fig. 2, not much su1fi.u was absorbed, so the 850" C regeneration temperature probable caused structural damage to the material. X-ray difhction measurements are being done to determine the extent to which Zn vs Ni are sulfidizing in these tests.

One other experiment was done on the Ni-containing material. It was heated in simulated coal gas without sulfur to measure the tendency for metallic Ni to form. These results are shown in Fig. 3. There was a pronounced weight loss just after the coal gas was introduced at a time of 45 min., and the amount is nearly the same as one would expect for reduction of NiO to Ni. The slow weight loss at longer times is believed to be due to vaporization of Zn.

Au things considered, the sulfidation behavior of the Ni-doped material is not very promising. Whether or not additional experiments will be done on this material depends on the outcome of the x-ray studies.

4

The Cr-doped formulation gave quite different TGA results. Fig. 4 is an analog to the left portion of Fig. 1 for a specimen containing Cr. There is no pronounced initial weight loss, indicating there was no reduction to metallic Cr. The sulfidation rate was high for a period of 50 min., but then the rate became slow over a period of 4 hrs., and saturation was not attained. The.total weight gain was 14.5%, compared to an expected gain of 10.6% if only the Zn sulfidized. If all of the Cr were present as Zn,.,Cr,,,TiO,, conversion of the Cr to CrS would produce a 2.7% gain. Total sulfidation to ZnS and CrS would therefore produce a 13.3 wt. gain. Since the experimental wt. gain was higher, there must have been an incorrect assumption made in thestoichiometric calculation. Again, x-ray data should clan@ the nature of the sulfidation reaction. It is clear, however, that both Cr and Zn in the material are forming sulfides.

Regeneration of the same specimen at 650" C is described in Fig. 5. The wt. loss was 10.8% over a period of 2 hrs., compared to an expected loss, matching the sulfidation gain, of 14.5%. In Fig. 6, a second cycle of sulfidation resulted in a wt. gain of 12.3%.

A new specimen was sulfidized at 725" C to try to obtain a faster approach to equilibrium. These results are shown in Fig. 7. The wt. gain was 12.2%, and the rate of sulfidation was considerably faster at 725 than at 650" C.

Although these results are not hlly understood at present, the Cr-doped material appeared to have promise, so an additional TGA experiment was carried out in simulated coal gas. The H2S concentration was 0.3% and the temperature was 650" C. The results are shown in Fig. 8. Again, there was no evidence that metallic Cr was produced. The rate was rather slow, and the wt. gain was only 4.6% after nearly 6 hr. of reaction time. A major factor was the lower H2S concentration in the gas (0.3 vs 0.9%), which has little to do with the properties of the sorbent.

Only one TGA experiment was carried out on the Cu-doped sorbent. As shown in Fig. 9, there was an initial wt. loss when the sulfidizing gas was introduced, and then an increase back to approximately the starting weight. Reduction of CuO to Cu in the material initially produces a theoretical wt. loss of 2.6%, and the measured loss was 2.9%. Also conversion of CuO to Cu2S produces a theoretical wt. gain of zero, and no overall wt. gain was observed. A tentative explanation for the data is that only Cu in the material is reacting, and for an unknown reason the Zn is not converted to ZnS. Another possible explanation is that both ZnS and Cu,S form near the surface of the particle, and the molar volume changes seal off the pores in the.material;

To summarize the results of the TGA experiments, only the Cr-doped material gave a high d f b r capacity and was stable against initial oxide reduction by the sulfidizing gas mixtures. While the Ni- and Cu-doped materials do not at present look promising, a final judgment should not be made until x-ray difEaction data are available and perhaps some additional tests are conducted.

5

At VPI, results of solubility measurements will not be available until the next quarter, due in part to a breakdown of x-ray dfiaction equipment.

CONCLUSIONS AND RECOMMENDATIONS

All experiments planned for this quarter were initiated. From the limited data obtained to date, it can be concluded that doping of zinc titanate sorbents with third metal ions can result in sigdicant increases in crush strength. Data fiom thermogravimetric analysis experiments showed that chromium doping results in a high sulfur capacity and moderately high sulfidation and regeneration rates. Sorbents doped with nickel and copper do not seem to be as attractive with respect to reactivity. It is recommended that the screening experiments be continued during the next quarter to obtain a more complete set of data for sorbents doped with aluminum and magnesium, in addition to the dopants mentioned above.

DISCLAIMER STATEMENT

This report was prepared by James H. Swisher of Southern Illinois University at Carbondale . with support, impart by grants made possible by the U.S. Department of Energy Cooperative

Agreement Number DE-FC22-92PC92521 and the Illinois Department of Energy through the Illinois Coal Development Board and the Illinois Clean Institute. Neither James Swisher of Southern Illinois University at Carbondale nor any of its subcontractors nor the U.S. Department of Energy, Illinois Department of Energy and Natural Resources, Illinois Coal Development Board, Illinois Clean Coal Institute, nor any person acting on behalf of either:

(A) Makes any warranty of representation, express or implied, with respect to the accuracy, completeness, or usemess of the information contained in this report, or that the use of any information, apparatus, method or process disclosed in this report may not infiinge privately-owned rights; or

(€3) Assumes any liabilities with respect to the use of, or for damages resulting fiom the use o$ any information, appartus, method or process disclosed in this report.

Reference herein to any specific commercial product, process, or service by trade name, trademark, mandacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the U.S. Department of Energy. The views and opinions of authors expressed herein do not necessarily state or reflect those of the U.S. Department of Energy.

Notice to Journalists and Publishers: If you borrow information fiom any part of this report, you must include a statement about the DOE and Illinois cost-sharing support of the project.

6

REFERENCES

1. Ayala, R.E., E. Gal and R.P. Gupta, 1993, "Testing Zinc Titanate Desulfurization Sorbents for Moving Bed Applications," Proc. Coal-Fired Power Svstems 93-- Advances in IGCC and PFBC Review Mtg, D.L. Bonk, ed., Report No.

DOE/METC-?3/613 1, US DOE Morgantown Energy Technology Center, Morgantown, WV, pp. 136-145.

2. 3evan, S., D. Najewicz, E. Gal, AH. Furman, RE. Ayala and A. Feitelberg, 1994, "Integrated Operation of a Pressurized Fixed Bed Gasifier, Hot Gas Desulfurization and Turbine Simulator System," Proc. Coal-Fired Power Systems 94-Advances in IGCC and PFBC Review Mtn., H.M. McDaniel, RK. Staubly and V.K. Venkataraman, eds., Report No. DOE/METC-94/1008, US DOE Morgantown Energy Technology Center, Morgantown, WV, pp. 222-235.

3. Datta, RIC, 1994, Tomposition Modification and Characterization of ZnO-TiO, Based (ZTB) DesulfUrization Sorbents," Final Report under U.S. Dept. of Energy Contract No. DE-Ap21-93MC53415, Vignia Polytechnic Institute and State Univ., Blacksburg, VA 24060. . .

4. Douglas, J., Dec. 1990, "Beyond Steam: Breaking Through Performance Limits," EPRI -9 Journal Electric Power Research Inst., Palo Alto, C q pp. 4-1 1.

5. Grindley, T., 1990, "Effect of Chlorine on Hot Gas DesulfUrization Sorbents," Proc. Tenth Annual Gasification and Gas Stream Cleanup Svstems Contractors' Review Meeting, Vol. 1, Report No. DOE/METC-90/6115, US DOE Morgantown Energy Technology Center, Morgantown, WV, pp. 215-226.

6. Gupta, R.P. and S.K. Gangwal, 1993, "High-Temperature, High-pressure Testing of Zinc Titanate in a Bench-Scale Fluidized Bed Reactor for 100 Cycles," Topical Report under U.S. Dept. of Energy Contract No. DE-AC21-88MC25006, Research Triangle Inst., Research Triangle Park, NC 27709-2194.

7. Gupta, R.P., S.K. Gangwal and S.C. Jain, 1993, "Fluid-izable Zinc Titanate Materials with High Chemical Reactivity and Attrition Resistance," U.S. Patent No. 5,254,516, Awarded to Research Triangle Inst., Research Triangle Park, NC 27709-2194.

8. Konttinen, J., R Ghazanfari A Lehtovaara and W. Mojtahedi, 1994, "Enviropower Hot Gas Desulfiukation PDU," Proc. Coal-Fired Power Systems 94--Advances in IGCC and PFBC Review Mtg., H.M. McDaniel, RK. Staubly and V.K. Venkataraman, eds., Report No. DOE/METC-94/1008, US DOE Morgantown Energy Technology Center, Morgantown, WV, pp. 236-245.

9. Mei, J.S., L. Gasper-Gale C.E. Everett and S. Katta,1993, "Fixed-Bed Testing of a Molybdenum-Promoted Zinc Titanate Sorbent for Hot Gas Desulfurization," Proc.

. .

7

Coal-Fired Power Systems 93-Advances in IGCC and PFBC Review Mtg., D.L. Bonk, ed., Report No. DOE/METC-93/6131, US DOE Morgantown Energy Technology Center, Morgantown, WV, pp. 315-325.

10. Swisher, J.H., W.S. O'Brien, and RP. Gupta, 1994, "An Attrition-Resistant Zinc Titanate for Sulfur," Find Technical Report on Project No. 93-113.2A-3M. Available from the Illinois Clean Coal Insititute, Carterville, IL 62918-0008.

56

55

54

53

52

51

Figure 2. Sulfidation of Zn,,,Ni,,TiO, in 0.9% H,S, 92.4% H, and 6.7% N, at 650" C and Regeneration'in 5% 0, and 95% N, at 850" C (two cycles).

9

0 so (v

0 ' C D T

n c o n

.- E

i=

0 'co

0 ' e

.o

[Bw] SSEW

68

67

66

65

62

61

60

59 I I I I I I

Time [min] 0 50 100 150 200 250 300 350

c-, 0

Figure 4. First Sulfidation of Zn,,Cr,,TiO, in 0.9% H,S, 92.4% H, and 6.7% N, at 650°C.

. .. . - .

11

0 'co 7

0 'a 7

,o 03

0 d-

0

. . .

[Bw] SSEpJ

0 'cu m

0 'co cu

0 ' -4 - cu

0 '0 cu

0 ' C D T-

O (u r

0 co

0 -4-

0

13

. .

[fju] SSEW

0 '0 d-

0 '0 m

- c .- E

E i=

u

Q,

0 0 cu

0 0 7

0

14

0 '0 d-

0 '0 rn

E i= 0

'0 CY

0 '0

T-

(D

[~LIJ] SSER

0 (D

.-

15

[Bu] ssen