Energy & Fuels 1994,8,896-906
896
Quantification of Nitrogen Forms in Argonne Premium Coals S. R. Kelemen,. M. L. Gorbaty, and P. J. Kwiatek Exxon Research and Engineering Co., Annandale, New Jersey 08801 Received September 9, 1993. Revised Manuscript Received March 23, 1994'
X-ray photoelectron spectroscopy (XPS) was used to identify and quantify the organically bound nitrogen forms present in fresh Argonne Premium coals. XPS spectra obtained on a wide variety of model compounds were used to establish a curve resolution methodology. Pyrrolic nitrogen is the most abundant form of organically bound nitrogen, followed by pyridinic types. Quaternary nitrogen species were also identified. An analysis of the sensitivity of XPS for primary amines indicates that these species are not detected at concentrations above 5 mol %. A trend of decreasing level of quaternary nitrogen forms along with increasing levels of pyridinic nitrogen was observed with increasing coal rank. The formation and reactivity of quaternary nitrogen forms in several coals were examined. Quaternary nitrogen forms decrease upon pyrolysis and hydropyrolysis. The level of quaternary nitrogen increases upon treating high-rank coal with strong acid.
I. Introduction
* Abstract published in Advance ACS Abstracts, May 1, 1994.
(1) Kelemen, S.R.;George, G. N.; Gorbaty, M. L. Fuel 1990,69,939. (2) Kelemen, S. R.;Gorbaty, M. L.; George, G. N.; Kwiatek, P. J.; Sansone, M. Fuel 1991, 70, 396. (3) Gorbaty, M. L.; George, G. N.; Kelemen, S. R. Fuel 1990,69,945. (4) George, G. N.;Gorbaty,M. L.; Kelemen, S.R.; Sansone, M. Energy Fueb 1991L5, 93. (6)Huffman, G. P.;Mitra, S.;Huggins, F. E.; Shah, N.; Vaidya, S.;Lu, F. Energy Fuels 1991,5, 574. (6) Taghiei, M. M.; Huggins, F. E.; Shah, N.; Huffman, G. P. Energy Fuels 1992,6, 293. (7) Brown, J. R.;Kasrai, M.; Bancroft, M. G.; Tan, K. H.; Chen, J. H. Fuel 1992, 71, 649. (8) Calkins, W. H.; Torres-Ordonez, R. J.; Jung, B.; Gorbaty, M. L.; George, G. N.;Kelemen, S.R. Energy Fuels 1992,6,411. (9) Kelemen. S.R.:Gorbatv. - . M. L.:. Vauahn. - . S.N.: Georae, - . G . Fuel 1993; 73, 645. (10) Jones. R. B.: McCourt. C. B.: Swift. P. Proc. Znt. Conf. Coal Sci.. Ddseldorf 1981,657. (11) Perry, D.L.;Grint, A. Fuel 1983, 62, 1029. (12) Bartle, K. D.;Perry, D. L.; Wallace, S. Fuel Process. Technol. 1987, 15, 351. (13) Wallace, S; Bartle, K. D.; Perry, D. L. Fuel 1989,68, 1450. (14) Burchill. P.: Welch. L. S. Fuel 1989. 68. 100. (15) Wilheha, A,; Patience,R. L.;Larter,S. R:;Jorgensen, S.Ceochim. Acta 1992,56, 3745.
in the auantification of organic nitrogen forms. Develop-
binding energy.13 The first XANES studies of coal-and petroleum asphaltenes support the position that pyrrolic and pyridinic groups are the most abundant nitrogen forms in these materials.la20 An error estimate, based on the unnormalized XANES nitrogen fraction, is considered to be about 10% The present work describes a self-consistent curve resolution methodology based on XPS data from model compounds. The repeatability of obtaining raw XPS spectra and the reproducibility of the curve resolution analysisare also examined. A detailed analysis of the XPS nitrogen (le) line shape provides quantitative data on both
'
(16) Patience, R.L.; Baxby, M.; Bartle, K. D.; Perry, D. L.; Reee, A. G. W.; Rowland, S.J. Org. Geochem. 1991,18,161. (17) Buckley, A. N.h e 1 Process. Technol., in press. (18) Mitra-Kirtley, 5. M.;Mullins, 0. C.; van Elp, J.; Cramer, S. P. Fuel 1993, 72, 133. (19) Mitra-Kirtley, S.;Mullins, 0.C.; van Elp, J.; Cramer, S.P. J. Am. Chem. SOC.1993,115,252. (20) Mullins, 0.C.; Mitra-Kirtley, S;Van Elp, J.; Cramer, S. P. Appl. Spectrosc. 1993,8, 1268.
0 1994 American Chemical Society
Nitrogen Forms in Argonne Premium Coals the major and minor forms of nitrogen present in fresh Argonne Premium coals. Changes in the distribution of nitrogen functionalities due to pyrolysis and hydropyrolysis are examined for Illinois No. 6 bituminous and Wyodak subbituminous coals. The relevance of these results toward understanding the chemistry of coal metamorphism is discussed. 11. Experimental Section
The XPS spectra were obtained with a Vacuum Generators (VG) ESCA Lab system using Mg Ka nonmonochromatic radiation and either a single- or a fivechannel detection arrangement. The five-channel signal system provided a 5 times greater XPS signal for the same X-ray exposure time. Data acquisition and analysis were done using the VG supplied VGS 5000 software package. The model compounds used in this study were obtained from Aldrich Chemical Co. and were used without any further purification. The coals were obtained from the Argonne Premium Coal Sample Program.21 The model compound samples were made into fine powders and mounted to a metallic sample block by means of Scotch double-sided nonconducting tape. The -100-mesh coals samples as received were mounted in the same way. In all cases care was taken to produce a uniform coating of sample on the nonconducting tape substrate. This was found to generate the most uniform level of sample charging and minimize distortions to the XPS signal. An energy correction was made to account for sample charging based on the carbon (1s) peak a t 284.8 eV. All spectra were obtained at an analyzer pass energy of 20 eV and a constant analyzer transmission mode. The coal samples were also made into finer powders by grinding in anitrogen drybox using a wig-l-bug. These samples were mounted in the same way. The identical XPS spectrum was obtained for the as received and the more finely ground coal sample for all coals except IllinoisNo. 6 and Pittsburgh No. 8 coal. With the latter coals a slightly narrower (0.10.2 eV) full width half-maximum (fwhm) peak was found for the XPS carbon (1s) spectrum for the more finely ground samples. The fwhm for these fiiely ground samples was similar to the other Argonne Premium coals of similar rank. The relative levels of heteroatoms were the same before and after grinding. Since fine grinding was necessary in the case of Illinois No. 6 and Pittsburgh No. 8 coal to obtain a powdered sample that gave the same uniformity of sample charging as observed in all of the other Argonne Premium coals, only results from ground samples of Illinois No. 6 and Pittsburgh No. 8 are reported in this study. (21) Vorres, K. S.,Ed. The Users Handbook for the Argonne Remium Coal Sample Program; Argonne National Laboratory: Argonne, 1989; Vol. 11; ANL-PCSP-89-1;Energy Fuels 1990, 4, 420. (22) Clark, D. T.; Wilson, R. Fuel 1983, 62, 1034. (23) Weitzsacker, C. L.; Gardella,J. A,, Jr. Anal. Chem. 1992,64,1068. (24) Solomon, P. R.; Serio, M. A.; Carangelo, R. M.; Bassilakis, R.; Gravel, D.; Baillargeon, M.; Baudais, F.; Vail, G. Energy Fuels 1990,4, 320.
(25) Kelemen, S. R.; Gorbaty, M. L.; Kwiatek, P. J. R o c . Int. Conf. Coal Sci., Banff 1993,1, 270. (26) Bumham, A. K.; Oh, M. S.; Crawford, R. W. Energy Fuels 1989, 3, 42. (27) Bassilakis, R.; Serio, M. A.; Solomon, P. R.; R e p r . Pap.-Am. Chem. SOC.,Diu. Fuel Chem. 1992, 37, 1712. (28) Fiedler, R.; Bendler, D. Fuel 1992, 71, 381. (29) Batich, C. D. Appl. Surf. Sci. 1988, 32, 57. (30) Kelemen, S. R.; Freund, H. Energy Fuels 1988,2, 111. (31) Kelly, M. D.; Buckley, A. N.; Nelson, P. F. Znt. Conf. Coal Sci., Newcastle 1991, 356.
Energy & Fuels, Vol. 8, No. 4, 1994 897 Elemental data for the coal samples were obtained from the Users Handbook for the Argonne Premium Sample Program.21 Other elemental analytical data were obtained from Galbraith Analytical Laboratories, Knoxville, TN. Elemental analysis data were also derived from measurements of the areas of the XPS peaks after correction for atomic sensitivity. The sensitivity factors were obtained from VG sensitivity tables and experimentally measured standards. The elemental concentrations are presented relative to carbon. Pyrolysisexperiments were done in a quartz lined reactor in helium a t 1atm. The reactor temperature was raised to 400 OC at approximately 0.5 OC/s and held a t 400 "C for 5 min. Under these conditions, little of the ultimate amount of hydrocarbons expected from the volatile matter determination is released while much of the oxygen as C02, H20, and CO evolves.2 Hydropyrolysis was accomplished in a closed reactor pressurized a t room temperature to 70 atm with a 95% hydrogen4 % helium mixture. The reactor temperature was raised to 427 "C at a heating rate of 0.5 "C/s and held for 30 min. These conditions favored the retention of coal hydrocarbon components and hydrocarbon products in the hydropyrolysis residues. The following procedure was used for the "titration" of basic nitrogen functionalities in coal. One gram of coal was added to 1.0 g of acid titrant dissolved in 75 mL of THF. After 24 h of stirring under nitrogen at room temperature the mixture was decanted and the remaining coal washed several times with fresh THF. The treated coal was dried at 50 "C overnight under vacuum. 111. Results
Under XPS data acquisition conditions described in the Experimental Section the nitrogen (1s)spectrum from model compounds containing a single nitrogen species could be curve resolved with a single peak. The full width at half-maximum of 1.7 (*0.05) eV and a mixed 30% Gaussian-Lorentzian line shape provided the best fit to the model compound data. The energy position found for the nitrogen (1s) peak in different model compounds is shown in Table 1. These values are in good agreement with available data determined for model compounds having similar chemical types of nitrogen.12J3 In model compounds where two different nitrogen types exist the nitrogen (1s) spectrum was curve resolved using two peaks of equal intensity and the same line shape as used for the other model compound data. In these cases the energy position of the peak used in the curve resolution procedure is reported in Table 1. The amount of nitrogen measured by XPS for the fresh coal samples was compared to that obtained on the whole sample by elemental analysis. The results expressed as N/Catom ratios are shown in Table 2 and there is excellent agreement for all coals except the Illinois No. 6 and Lewiston samples. The generallygood agreement indicates that there is no systematic enhancement or depletion of nitrogen concentrations a t the coal surface relative to the bulk sample. XPS analysis of Illinois No. 6 and Lewiston coals gave lower nitrogen to carbon ratio values than found in the bulk elemental analyses. The same values for nitrogen for these coals were obtained after grinding the fresh coals to expose new surfaces. The XPS nitrogen values for Illinois No. 6 coal were consistent throughout thermal and reductive treatment. The origin for the
Kelemen et al.
898 Energy & Fuels, Vol. 8, No. 4,1994 Table 1
~~
sample phenazine poly(2-vinylpyridine) 3-hydroxypyridine norharman PIB amine l-aminopyrene norharman 2-hydroxycarabazole dibenzocarbazole l-hydroxyquinoline l-methyl-4-pentadecyl2(1H)-quinolone 6-(2,2-diphenyl-2-hydroxylethyl)-2( ZH)-pyridone 1-ethyl-4methoxypyridinumiodide pyridinium 3-nitrobenzenesulfonate(pyridinium) 3-hydroxypyridineN-oxide 9-hydroxy-3-nitroflourene pyridinium 3-nitrobenzenesulfonate(nitro) l-nitropyrene
Table 2. Nitrogen/Carbon Atom Ratio (XlOO) coal total XPS total bulk Beulah Zap 1.5 1.5 1.3 1.4 Wyodak 1.2 1.7 Illinois No.6 1.6 1.7 Blind Canyon Pittsburgh No. 8 1.2 1.3 1.0 1.7 hwiston 1.5 1.6 Upper Freeport Pocahontas 1.2 1.3
difference between XPS and bulk values for these two fresh coals is yet to be resolved. Previous XPS investigations of other coal samples gave reasonable agreement between XPS and bulk values for nitrogen.1°-14*2 2 ~ x j A tendency toward lower XPS values was n0tedl1J3J4and a possible uncertainty in the XPS sensitivity factor for nitrogen was suggested as one possible cause for the discrepancy. This is an inadequate explanation for our results on Argonne Premium Coals. It is also unlikely that poor experimental precision is the cause of the disparities. For example, based on 12 measurements of fresh Wyodak coal the value for the nitrogen to carbon atom ratio was 1.3 (*0.09) at the 95% confidence level based on a Student's t distribution analysis method. A similar level of precision is expected with XPS for other coals. Good precision is an essential prerequisite for quantification of different nitrogen functionalitiesinitially present in coal and identification of subsequent changes due to chemical reaction. The XPS nitrogen (1s) signals for the fresh Argonne Premium coal samples are shown in Figure 1. Included in Figure 1are the results of the curve resolution analysis for nitrogen forms. Peaks at 398.8,400.2, and 401.4 (fO.l) eV were used in the curve resolution process. These peaks correspond to the energy positions found for pyridinic, pyrrolic, and quaternary type nitrogen functionalities, respectively. In each case the nitrogen (la) signal was curve resolved using a 70 % Gaussian 30 % Lorentzian line shape and a peak fwhm of 1.7 eV for each coal. The peak shape and peak energy positions are fixed in the curve resolution process adopted in the present study and only the amplitudes of these peaks were varied to obtain the best fitto the experimental XPS data. It should be pointed out that the binding energy of primary amine species in model compounds appear close to pyridinic forms. The special quantificationproblems that this situation presents are discussed in the Appendix where the sensitivity of the
nitrogen (Is) binding energy (eV) 398.7 398.8 398.8 398.9 399.1 399.3 400.1 400.2 400.2 400.3 400.4 400.5 401.3 401.4 403.0 405.3 405.8 405.9
~
nitrogen type pyridinic pyridinic pyridinic pyridinic amine amine Pyrrole pyrrole pyrrole pyridone pyridone pyridone quaternary quaternary N-oxide nitro nitro nitro
analysis procedure to changes in the fwhm and energy position of the peaks used in the curve resolution analysis are examined. This analysis of the sensitivity of XPS for primary amines indicates that they may be present in most coals, however, their concentration does not exceed 5 mol 7%. The presence of primary amines at these low levels cannot be unequivocally distinguished from pyridinic forms. For this reason the potential relative concentration of primary amines does not explicitly appear in the tabulated values for nitrogen forms (see Appendix). A self-consistent curve resolution methodology based on peaks at the energy position characteristic of ppidinic, pyrrolic, and quaternary nitrogen provided optimum fits to the experimental data. The line shape and fwhm for the peaks chosen to fit the XPS nitrogen (Is) spectra of coal were the same as those used to fit individual nitrogen functionalities in model compounds. A further experimental check on the use of 1.7 eV for the fwhm to represent the experimentalresponse for a single functional group comes from the carbon (1s) spectrum of Pocahontas coal. For Pocahontas coal the carbon (la) signal is dominated by hydrocarbons because of the very low heteroatom content and a fwhm of 1.7 eV is observed. For other coals a slightly broader fwhm is observed which is related to the higher organic oxygen content of the lower rank coals.% In Figure 1 the individual peaks used in the curve resolution process and the total simulated spectra are shown with the actual spectra. The anticipated pyrrolic (400.2 eV) and pyridinic (398.8 eV) nitrogen forms were the most abundant species identified. The two forms could almost completely describe the nitrogen (1s) spectrum of a high-rank coal such as Pocahontas; however, in all other cases it was necessary to include a peak at 401.4eV. This peak is interpreted to arise from quaternary nitrogen species. Notice in Figure 1the very good signal to noise (S/N) characteristics (in general, greater than 501) of each spectrum and how well the s u m of theoretical peaks fit the actual data. There are limited degrees of freedom in the curve resolution methodology, described in the Experimental Section. Data withgood S/N provide the basis for assessing the choice of the number and kinds of peaks in the curve resolution methodology that are required to provide optimum fits to the experimental data. Experimentswere performed to determinethe precision of the reproducibility of the nitrogen (Is) spectrum for
Energy & Fuels, Vol. 8, No. 4,1994 899
Nitrogen Forms in Argonne Premium Coals I
P’Y RRO‘l
A
I
c
I
I
I
Beulah Zap
I
I
I
I
A /; ’\\
I 399
I
397
I
401
I 403
I
I
I
I
407
387
389
401
I
405
I
I
Lewisto nStockton
I 403
I
405
I 407
Figure 1. XPS nitrogen (1s) spectra and curve resolution of fresh Argonne Premium coals. Table 3. XPS Mole Percent (h3 at 95% Confidence Level) Pyrrolic Quaternary Wyodak Coal Pyridinic 15 26 59 repeatibility reproducibility
25
60
15
coal and to access the repeatability of the curve resolution process. The repeatability of the curve resolution process, defined as the closeness of agreement between successive independent attempts at curve resolution of the same XPS spectrum, was evaluated for Wyodak coal. The results for the repeatability of the curve analysis of Wyodak coal are shown in Table 3 based on 10 individual curve resolution attempts. The standard deviation is &3mol 7% at a 95 7% confidence level based on a Student’s t distribu-
tion analysis. The reproducibility of the nitrogen XPS (1s) spectrum on subsequent curve resolution, defined as the closeness of agreement between individual results on different coal samples, is also included in Table 3 for Wyodak coal. As examples,four individual XPS nitrogen (1s) spectra for different Wyodak coal samples are shown in Figure 2 along with the results of curve resolution. The results in Table 3 for reproducibility are based on eight individual XPS nitrogen (1s) spectra of Wyodak coal. The standard deviation was f 3 mol 7% at a confidence level of 95% based on a Student’s t distribution analysis. The standard deviation at the 95 7% confidence level is similar for both repeatability and reproducibility of the curve resolution results for the nitrogen (1s)spectrum of Wyodak
Kelemen et al.
900 Energy & Fuels, Vol. 8, No. 4,1994
Fresh Wyodak
I
I
I
1
I
I
396
398
400
I
I 402
1
1
I
I
I
I
1
396
398
400
402
Binding Energy (eV) Binding Energy (eV) Figure 2. XPS nitrogen (le) spectra and curve resolution of different samples of fresh Wyodak coal. Table 4. XPS Mole Percent coal Beulah Zap Wyodak Illinois No. 6 Blind Canyon Pittsburgh No.8 Lewiston Upper Freeport Pocahontas
Pyridinic 25 25 26 31 32 31 28 33
Pyrrolic 59 60 62 55 61 60 65 64
I
I
I
1
ARGONNE PREMIUM COAL SAMPLES
Quaternary 16 15 12 14 7 9 7 3
coal. The standard deviation is approximately *3 mol % . It is expected that a similar result for the standard deviation would be obtained with different Argonne Premium coals. It is important to emphasize that the anticipated standard deviation for quantification of nitrogen forms based on the XPS curve resolution method is sensitive to the signal to noise characteristics of the XPS spectrum. In the present study S/N greater than 501 is required to achieve the stated standard deviation. It has been our experience that there is an unacceptably large uncertainty associated with XPS spectra with S/N below 101, and this uncertainty precludes a detailed interpretation of such XPS data. Attempts at doing so yield highly questionable results. The average numerical results of the curve resolution analysis for nitrogen species in fresh Argonne Premium coal are shown in Table 4. In all cases the peaks that are associated with pyrrolic and pyridinic species are the dominant features of each spectrum. The observation is in agreement with previous findings for other ~0als.lO-l~ There is a significant contribution of a peak in the energy position associated with quaternary nitrogen for lower rank Argonne Premium Coals. The results of curve resolution
65
75
85
95
WEIGHT PERCENT CARBON (Dmmf)
Figure 3. Nitrogen functional groups by XPS curve resolution analysis.
analysis for organic nitrogen forms for each individual XPS experiment were plotted as a function of the dmmf weight percent carbon in coal. Figure 3 showsthese results. Because some of the data overlap, not all of the data obtained can be represented as individual points. These results show a distinct trend of decreasing level of the peak associated with quaternary nitrogens with increasing coal rank for the Argonne Premium coals. The relative
Nitrogen Forms in Argonne Premium Coals
Energy & Fuels, Vol. 8, No. 4,1994 901
Table 5. XPS Atom Ratio ( M O O ) organic organic coal oxygen/C nitrogenic Illinois No. 6 (initial) 9.0 1.2 Illinois No.6 (pyrolysis) 6.7 1.3 Illinois No.6 (hydropyrolysis) 4.8 1.5 Wyodak (initial) 16.9 1.3 Wyodak (pyrolysis) 8.8 1.3 Wyodak (hydropyrolysis) 5.6 1.4
level of the pyrrolic nitrogen peak is similar for most coals. There appears to be a tendency for the relative level of the peaks associated with pyridinic nitrogen to increase corresponding to decreasing contribution of the peaks associated with quaternary nitrogen as a function of increasing coal rank. The changes in the total level of organic nitrogen and oxygen following pyrolysis and hydropyrolysis of Illinois No. 6 and Wyodak coal were examined by XPS. The results of total nitrogen in the residual pyrolysis and hydropyrolysis chars, expressed as atomic N/C, are shown in Table 5 along with the results for total organic oxygen based on XPS analysis. For Illinois No. 6 coal the level of organic oxygen drops nearly in half during hydropyrolysis. Less organicoxygen is lost during pyrolysis. There is only a slight increase in the relative level of total nitrogen to carbon after pyrolysis and hydropyrolysis. For Wyodak coal the organic oxygen level drops in half after pyrolysis and to one-third of its initial value after hydropyrolysis. In contrast, the relative level of atomic nitrogen to carbon remains close to the initial value. The loss of organic oxygen after mild pyrolysis and hydropyrolysis conditions employed here is likely a result of loss of organic oxygen functionalities as C02, H20, and C0.26128 The loss of nitrogen as small gaseous molecules (Le., NH3, HCN, etc.) does not seem to occur at low temperature (2' < 450 OC)28 prior to the devolatilization of hydrocarbons. Since some carbon is lost with C02 and CO as well as through formation of small quantities of methane and similar light hydrocarbon gases, a slight increase in the relative level of nitrogen to carbon ratio would be expected if nearly all of the initial nitrogen is retained in the pyrolysis and hydropyrolysis chars. The changes in the XPS nitrogen (1s) line shape following mild pyrolysis and hydropyrolysis of Illinois No. 6 and Wyodak coal have been studied. Figure 4 shows the XPS nitrogen (1s) signalsfollowing hydropyrolysis of these coals. Included in Figure 4 are the results of the curve resolution analysis for nitrogen forms. In each case the individual peaks and the total simulated spectrum are shown with the actual spectrum. The anticipated presence of peaks associated with pyrrolic (400.2eV) and pyridinic (398.8 eV) nitrogen forms was the dominant feature in both hydropyrolysis chars. The two forms could describe the nitrogen (1s) spectrum following hydropyrolysis of Illinois No. 6 coal;however, it was still necessary to include a small peak at the position expected for the quaternary nitrogen species in the XPS spectrum of Wyodak hydropyrolysis char. The numerical results of the curve resolution analysis for nitrogen forms after pyrolysis and hydropyrolysis are shown in Table 6. There is a decline in the relative level of quaternary nitrogen for both coals after reaction. The decline is greater following hydropyrolysis. There is a significant increase in the relative level of the pyridinic nitrogen peak for both coals after hydropyrolysis.
I
I
I
I
I
PYRROLIC
A
PYRlDlNlC
I
Illinois #6
// \
Hyd ropy rolysi s 4 2 7 OC,
I
I
I
I
397
399
401
403
I
I
407
405
Binding Engergy (eV) I
I
I
1
I
Hydro py rol ysis
I
I
I
I
I
397
389
401
403
I 407
I
405
Binding Energy (eV)
Figure 4. XPS nitrogen (IS)spectraand curveresolutionanalysis of hydropyrolyzed coals.
Table 6. XPS Mole Percent coal Illinois No. 6 (initial) Illinois No.6 (pyrolysis) Illinois No.6 (hydropyrolysis) Wyodak (initial) Wyodak (pyrolysis) Wyodak (hydropyrolysis)
Pyridinic 26 30 35 25 29 35
P F rolic 62 64 65 60 62 65
~
quaternarv 12 6 0 15 9 3
The relative level of peaks associated with quaternary nitrogen forms in coal are small and a confident distinction from nitrogen oxides and assignment has not been made in the past. Examples of XPS nitrogen (1s) curve resolution results from other closely related systems provides support for the choice of the energy positions and interpretation of different peaks. Figure 5A shows the XPS nitrogen (1s) spectrum of a Wyodak coalwhich was treated with a solution of triethylamine and hydrocarbon solvent for 24 h at 75 O C . The coal was dried under vacuum. The total level of nitrogen retained corresponded to the addition of roughly 40% more nitrogen to the coal. An intense peak appeared in the nitrogen (1s) spectrum in addition to those that could be associatedwith pyridinic and pyrrolic nitrogen forms initially present in the coal. The peak appeared a t roughly the same energy position (401.6 eV) assigned to quaternary nitrogen. The intense peak is interpreted to arise from the reaction of triethylamine with acidic sites of the coal to produce a quaternary nitrogen species. This chemistry is not surprising. The derivatization of acidic surface sites with amines followed
Kelemen et al.
902 Energy & Fuels, Vol. 8,No. 4, 1994
$ 4
I
0 C
I
I
I I
I
I
I
I
A ’ j \A
I
1
Pocahontas
A
After Strong Acid Treatment
D
iij
F
Wyodak Coal
v
c Q)
D
e
U
f
v)
n
x
397
399
401
403
405
I
I
I
397
399
401
cp
c
cn
I
I
I
I
407
407
Binding Energy (eV)
Binding Energy (eV)
3 U
I
405
403
I
I
I
I
I
I
Upper F r e e p o r t
PYRROLIC
Nitric Acid Ox idized
I
After Strong Acid Treatment
iij c
5 W
c Q)
cn
e
U
f
v)
n
x
I
I
397
I 399
I
I
401
403
I
405
I 407
Binding Energy (eV) Figure 5. (A) XPS nitrogen (1s) spectrum of Wyodak coal following exposure to triethyl amine. (B) XPS nitrogen (Is) spectrum of nitric acid oxidized glassy carbon. by XPS analysis has be noted as a possible way to “tag” and quantify acidic groups in polymer systems.29 The oxidation of glassy carbon surfaces by nitric acid is an example that illustrates the spectroscopic signature of nitrogen oxides.30 The initial XPS nitrogen to carbon atom ratio for glassy carbon is usually low (
