Energy & Fuels 1994,8,886-889
886
XPS Studies of Residues from Liquefaction of Blind Canyon Coal Mixed with an Iron-Based Catalyst J. Y. Kim, P. J. Reucroft,' and M. Taghiei Department of Material Science and Engineering and CFFLS, University of Kentucky, Lexington, Kentucky 40506
V. R. Pradhan and I. Wender Chemical and Petroleum Engineering Department, University of Pittsburgh, Pittsburgh, Pennsylvania 15261 Received July 12, 1993. Revised Manuscript Received April 26, 1994'
X-ray photoelectron spectroscopy (XPS) has been used to investigate the surface and bulk characteristics of residue from liquefaction of Blind Canyon coal mixed with an iron-based catalyst. The liquefaction of Blind Canyon coal was investigated at three different processing times (17,30, and 60 min); residues were studied by XPS. It is shown that the concentration of elements at the outermost surface layer of samples, measured by XPS, is different from the bulk. Ar+ ion sputtering followed by XPS was carried out to delineate the differences in the distribution of elements and chemical changes that occurred as the exposed surface varied from initial surface to the bulk. The elemental composition of the catalyst mixed with coal residue surface changes with time of Ar+ ion sputtering, i.e., with the depth from the initial surface of the coal residue particles. The S/Fe (catalyst element) ratio, which was initially greater than 1,decreases to less than 1as time of Ar+ ion sputtering is increased.
Introduction Iron-based catalysts with a specific surface area and fine particulate size can be utilized at small concentrations to achieve better performance in terms of overall coal conversion and selectivityto lighter products (oils)in direct coal liquefaction.l-3 The catalyst dispersion and its intimate contact with the coal also can play an important role in coal liquefaction. Since the catalyst surface is the primary site of catalytic activity,efforts have been direded to characterizing catalyst surfaces in coal liquefaction environments. X-ray photoelectron spectroscopy (XPS) has been used to study iron-based catalyst mixed with liquefied Blind Canyon coal sample residues that have been processed at three different liquefaction processing times (17,30,and 60 min). It was shown previously that the concentration of elements at the outermost surface layer of catalyst and liquefied coal residue samples measured by XPS was different from the bulk composition in some important re~pects:~*s less iron concentration and enriched oxygen concentration was observed in the surface regions of the samples compared to the bulk composition. In order to more carefully delineate the differences in the distribution of the elements and chemical changes that might occur between the surface and the bulk, the samples have been further analyzed by combining Ar+ ion sput-
* Abetract published in Advance ACS Abstracts, June 1, 1994.
(1) Derbyehire, F. J. Cataly8is in Coal liquefaction: New Directions for Research; IEA Coal Research London, 1988; IEA CR-08. (2) Cugini, A. V.; Lett, R. G.; Wender, I. Energy Fuels 1989,3,120-
126. (3) Weller, S.; Pelipetz, M. G. Znd. Eng. Chem. 19S1,43,1243-1246. (4) Kim, J. Y.;Reucroft, P. J.; Taghiei, M.; Pradhan, V. R.;Wender, I. R e p r . Pap.-Am. Chem. SOC.,Diu. Fuel Chem. 1992,37,756-763; Fuel ROCe88. Technol. 1993, 34,207-215. (5) Kim, J. Y.;Reucroft, P. J.; Pradhan, V. R.; Wender, I. R e p r . Pap.-Am. Chem. SOC.,Diu. Fuel Chem. 1993,38,220-226.
0887-0624/94/250&0886$04.50/0
tering with XPS measurements. Special emphasis has been focused on the surface to bulk distributions of the catalytic elements, iron and sulfur.
Experimental Section Materials. The liquefaction of Blind Canyon coal was investigated at three different proceesing times (17,30, and 60 min); residues were studied by XPS. After adding FezO$S042by physically mixing, so that the proportion of Fe was 1.75 atom % ,liquefactionconversion was carried out at 400 "C withtetralin as solvent (1:1.5 coal/solvent ratio). The surface Fe and S concentration in Blind Canyon coal residue before adding catalyst was -0.5 atom % . Dimethyl disulfidewas also added in amounta stoichiometricallyequivalent to Fe to promote sulfidation of the catalyst. The liquefied Blind Canyon coal residues plus catalyst were then removed in liquidlike form after 17, 30, and 60 min and kept in a vacuum oven at 110 "Cto remove volatile materials and moisture in the residue without exposingtoan air atmosphere. A solid powder residue sample was obtained from this treatment that could then be analyzed by XPS. Methods. The samples were examined by XPS employing a Kratos XSAM 800 spectrometer using Mg Ka (1253.6 eV) radiation. The samples were mounted on the spectrometer probe tip by means of double-sided adhesive insulating tape. After heating under vacuum (1W Torr) at 65 O C for 10 h in a pretreatment chamber to remove volatiles, the samples were inserted into an ultrahigh vacuum chamber for surface analysis. The spectrometer was run in fixed analyzer transmission (FAT) mode at a pass energy of 11 kV and 13 mA. Under these conditions,the full width half-maximum (fwhm)of the Ag (3612) peak is 21.0 eV. The system pressure was normally maintained below 3 X 108 Torr using a 300 L/s ion pump and Ti sublimators to minimize contributions from vacuum contaminants. To excludethe possibility of recordingcontaminanta associatedwith the tape, the tape surface was also analyzedseparately. However, it was found that the constituents of the tape were not detected by XPS when residue samples were present and that the 0 1994 American Chemical Society
Liquefaction of Blind Canyon Coal
Energy & Fuels, Vol. 8, No. 4, 1994 887
Table 1. Analysis (atom %) of Liquefied Catalyet/Coal Residue Samples sample C 0 N S Si Fe 17min 30min 60min
85.55 86.10 87.90
11.02 11.09 10.00
1.10 0.80 0.92
0.38 0.25 0.62
1.95 1.76 0.56
-
photoionization signals were characteristic of the sample alone. All binding energies were referred to carbon 1s at 285 eV to compensate for sample charging. Atomic concentrations were estimated from the area of peaks by applying the atomic sensitivity factom6 I n situ Ar+ ion sputtering of the residue samples was carried out using a differentially pumped and computer controlled 3M minibeam ion gun. The incident ion gun was operated at 3.5 keV and sample currents were kept around 3 MAacross a sample area of about 5 X 5 "2. The pressure in the main chamber was kept below 4 X 10-6Torr during ion sputtering. After ion sputtering, a consistent increase in the fwhm for all elementa was found with increasing ion dose. No compensation was made for charging of the samples after Ar+ ion sputtering .
175
I70
165
160
175
I70
165
I60
I75
I70
I65
160
Rssults and Discussion The surface XPS spectra showed a distinct peak for each of the major elements, carbon, oxygen, aluminum, and silicon, as well as the minor components, sulfur, and nitrogen in each sample. The initial surface XPS elemental concentrations of coal residue samples are given in Table 1. The mineral element surface concentration showed progressive reduction as the liquefaction time increased due to a process which encapsulates the mineral matter with carbonaceous materials. This is indicated by the increase in carbon surface concentration. The binding energies of these elements showed the expected values for the most stable oxidation states. A relatively high oxygen concentration was also determined initially at the surface, which can be ascribed to air oxidation. Two peak components were observed in the case of the initial surface sulfur XPS signal. The peak at 168 eV, which was relatively broad, could be ascribed to oxidized sulfur,718while the peak at 163eV corresponded to inorganic sulfide plus the usual organic sulfur forms present in coal (thiophene, sulfides, ~nercaptans).~ However, Fe was not observed initially in the surface regions of the liquefied coal residue particles. From the above results, it was apparent that catalytic elements were encapsulated by carbonaceous organic materials. In order to delineate the differences in the distribution of elements and chemical changes between the surface and bulk, the samples were analyzed by combining Ar+ ion sputtering with XPS measurements. Due to the edge effect associated with the unetched sample surface, the depth profiles for materials in powder form can be difficult to obtain. Selective areal sputtering was carried out to reduce this effect. Although different features such as preferential sputtering, atomic mixing, particle size effects, etc., can effect the etching profiles, sputtering was performed in the present studies for comparative purposes on samples where such effects were expected to be very similar. Ar+ ion sputtering was carried out for times up to 60 min.
Binding Energy
Figure 1. XPSspectraof sulfur in liquefiedcatalpt/coalresidues at three Ar+ion sputtering times: (a) 17 min processing time, (b) 30 min processing time, (c) 60 min processing time.
Figures 1and 2 show XPS spectra of the liquefied coal residues, at three different Ar+ ion sputtering times, for the S 2p and Fe 2p regions, respectively. Two peak components were initially observed in the case of the surface sulfur XPS signal. As sputtering increased, the sulfate peak (168 eV) decreased and finally disappeared, while the sulfide peak (164 eV) showed an increase in intensity with further etching. This observation is in agreement with the decrease in the oxygen concentration which was observed as the sputtering increased. Yoshimura et al. also reported similar XPS results of catalysts where surface sulfide was covered with oxidelsulfate layer^.^ The Fe peaks were not observed initially for most samples unless ion sputtering was carried out for at least 30 min. Subsequent to this, an increase in Fe concentration was observed with further etching. Weak Fe signals were still observed, however. As the liquefaction reaction time increased from 17 to 60 min, sputtering time had to be further increased in order to observe the Fe peaks. The (6) Wagner,C.D.;Rigp,W.M.;Davis,L.E.;Moulder,J.F.;Muilenberg, surface chemical state of Fe 2p showed the expected 2~312
G. E. Handbook of X-ray Photoelectron Spectroscopy; Perkin Elmer Corp. Physical Electronics Div.: Norwalk, CT, 1979. (7) Perry, D. L.; Grint, A. Fuel 1983,62, 1024-1033. (8)Frost, D. C.; Leeder, W. R.; Tapping, R. L. Fuel 1974,53,206-211.
(9)Yoshimura,Y.;Mataubayashi,N.; Yokokawa, H.; Sato,T.;Shimada, H.; Nishijima, A. Znd. Eng. Chem. Res. 1991,30,1092-1099.
Kim et al.
888 Energy & Fuels, Vol. 8,No.4, 1994 100
n
v
"
"
1
8
w
I
si1
I
\\ L "
--
'
1
'
I
'
I
'
0
~
"
I
I
I
I
IW
730
720
710
700
690
730
720
710
700
690
700
690
C)
n
730
720
A
710 Binding Energy
Figure 2. XPS spectra of iron in liquefied catalyst/coal residue at three Ar+ ion sputteringtimes: (a) 17 min processing time, (b) 30 min processing time, (c) 60 min processing time. binding energy value of 710.3 eV in most samples with 13.6 f 0.05 eV splitting between 2~112and 2 ~ 2 1 3peaks? Figures 3,4, and 5 show depth elemental concentrations, as a function of etching time, of the liquefied coal residue samples at liquefaction reaction times of 17, 30,and 60 min, respectively. A relatively high oxygen concentration was determined initially at the surface before sputtering, but a sharp drop in oxygen concentration was subsequently observed. This surface oxidized layer thickness was estimated to be approximately 50-60 nm. Carbon shows no significant change in concentration before and after Ar+ ion sputtering. These concentration changes with sputtering time, for carbon and oxygen, are similar to those observed in previous etching studies on a catalyst (FeOOH/SOP) impregnated Illinois No. 6 coal sample? In this case the sample wa8 not liquefaction processed and catalytic elements, such as iron and sulfur, showed a slight trend towards lower concentrations after Ar+ ion sputtering, in contrast to the present studies. From the present studies, it can be concluded that the iron catalyst particles are well encapsulated in either oxidized layers or within other organic materials from the coal residues, as a result of the liquefaction process.
0
10
20 30 40 Etching Time (min.)
50
60
Figure 4. Depth profile of element concentrations for 30 min liquefied catalyst/coal residue. The S/Fe (catalyst element) ratio, which was initially greater than 1, decreased consistently to less than 1 as time of Ar+ ion sputtering was increased (Tables 2,3, and 4). This can also be seen in Figures 3-5 for liquefaction times of 17, 30, and 60 min, respectively. EXAF'S and MiidJsbauer studies have shown that under the liquefaction conditions that were employed i.e., after adding dimethyl disulfide to the catalyst/coal residue mixtures, the ion catalyst is converted to iron sulfide (pyrrhotite) within a few minutes of starting the reaction.1° Conclusions
It is apparent from this study that XPS can be used to obtain surface chemistry information and elemental identificationson the exposed solid surfaces of liquefaction processed coal residue samples. A relatively high oxygen concentrationwas determined initiallyat the surface which (10)Taghiei, M. Private communication.
Energy & Fuels, Vol. 8, No.4, 1994 889
Liquefaction of Blind Canyon Coal 100
Table 2. Depth Profile of Catalyst Elements for 17 min Liquefied Coal Residue (Element Concentrations (atom % ) as a Function of Etching Time) Omin 10min 20min 30min 40min 50min 60min S 0.38 0.48 0.76 0.50 0.58 0.73 0.89 0.40 0.73 0.81 1.24 Fe S/Fe 1.25 0.79 0.82 0.72
4
0
I
10
20 30 40 Etching Time (An.)
50
60
Figure 5. Depth profile of element concentrations for 60 min liquefied catalyst/coal residue.
can be ascribed to air oxidation. Fe was not observed initially in the surface regions of most samples since it was apparently encapsulated by other organic materials. Ar+ ion etching reveals increasing concentrations of Fe in the bulk of the coal residue particles. Surface analysis combined with depth profiling shows that the sulfur/iron stoichiometryvaries in iron catalyst/liquefied coal residue particles that have been extracted from coal liquefaction environments. The S/Fe atomic ratio approaches 0.7-0.9
Table 3. Depth Profile of Catalyst Elements for 30 min Liquefied Coal Residue (Element Concentrations (atom % ) as a Function of Etching Time) Omin 10min 20min 30min 40min 50min 60min S 0.25 0.23 0.34 0.43 0.51 0.72 0.97 Fe 0.20 0.45 0.69 1.12 2.15 1.13 1.04 0.87 S/Fe Table 4. Depth Profile of Catalyst Elements for 60 min Liquefied Coal Residue (Element Concentrations (atom % ) as a Function of Etching Time) Omin 10min 20min 30min 40min 50min 60min S 0.62 0.53 0.42 0.38 0.37 0.62 0.63 0.17 0.23 0.52 0.74 Fe 2.24 1.61 1.19 0.85 S/Fe -
inside coal residue particles. These results have provided valuable and useful information which allows a better understanding of catalyst element distributions during coal liquefaction processing.
Acknowledgment. We gratefully acknowledge financial support from the Consortium for Fossil Fuel Liquefaction Science, University of Kentucky, under Department of Energy Contract No. DE-FC22-92 PC 90029. We also thank Dr. G. P. Huffman for his interest in this work.
