Energy Fuels 2010, 24, 557–562 Published on Web 10/28/2009
: DOI:10.1021/ef900908p
Sulfur Speciation of Crude Oils by Partial Least Squares Regression Modeling of Their Infrared Spectra Peter de Peinder,*,†,‡ Tom Visser,*,‡ Rudy Wagemans,§ Jan Blomberg,§ Hassan Chaabani,§ Fouad Soulimani,‡ and Bert M. Weckhuysen‡ †
VibSpec, Tiel, The Netherlands, ‡Inorganic Chemistry and Catalysis Group, Department of Chemistry, Utrecht University, The Netherlands, and §Shell Global Solutions International B.V., Amsterdam, The Netherlands Received August 20, 2009. Revised Manuscript Received October 9, 2009
Research has been carried out to determine the feasibility of partial least-squares regression (PLS) modeling of infrared (IR) spectra of crude oils as a tool for fast sulfur speciation. The study is a continuation of a previously developed method to predict long and short residue properties of crude oils from IR and near-infrared (NIR) spectra. Retention data of two-dimensional gas chromatography (GC GC) of 47 crude oil samples have been used as input for modeling the corresponding IR spectra. A total of 10 different PLS prediction models have been built: 1 for the total sulfur content and 9 for the sulfur compound classes (1) sulfides, thiols, disulfides, and thiophenes, (2) aryl-sulfides, (3) benzothiophenes, (4) naphthenic-benzothiophenes, (5) dibenzothiophenes, (6) naphthenic-dibenzothiophenes, (7) benzonaphthothiophenes, (8) naphthenic-benzo-naphthothiophenes, and (9) dinaphthothiophenes. Research was carried out on a set of 47 IR spectra of which 28 were selected for calibration by means of a principal component analysis. The remaining 19 spectra were used as a test set to validate the PLS regression models. The results confirm the conclusion from previous studies that PLS modeling of IR spectra to predict the total sulfur concentration of a crude oil is a valuable alternative for the commonly applied physicochemical ASTM method D2622. Besides, the concentration of dibenzothiophenes and three different benzothiophene classes can be predicted with reasonable accuracy. The corresponding models offer a valuable tool for quick on-site screening on these compounds, which are potentially harmful for production plants. The models for the remaining sulfur compound classes are insufficiently accurate to be used as a method for detailed sulfur speciation of crude oils.
developed and as a consequence the sulfur content of fuels is dramatically reduced. Nowadays, the maximum sulfur concentration in Europe is 10 ppm (w/w S-total) for gasoline and diesel4 and 1000 ppm for marine diesel.5 Desulfurization is therefore a big topic in oil industries. The current method of choice in refineries is HDS by means of, e.g., a cobaltmolybdenum based catalyst. This method is expected to stay the dominant technology for the coming years, even though it is still not possible to eliminate the sulfur completely.6 On the other hand, HDS is an expensive treatment for deep desulfurization, while the removal of heterocyclic aromatic sulfur compounds is not very effective. This is particularly relevant since the exploration of the tar sand fields in Canada has brought large amounts of crude oils with high sulfur concentrations onto the world market. For that reason, research for alternative methods and ways to enhance the efficiency of the HDS process is ongoing.4 Obviously, also in this process, detailed knowledge of the qualitative and quantitative composition of sulfur compounds in crude oils is essential. Besides, the type and molecular structure of the sulfur compounds are found to affect the crackability and detachability,7 while the
Introduction Crude oils are highly complex mixtures of organic compounds with a large variety in elemental composition and chemical structures. All crude oils contain sulfur in concentrations between 0.1 wt % in light samples up to 10% in, for example, bitumen and tar sands.1 The majority of the sulfur is present as organic molecules in more than 10 000 different structures, ranging from aliphatic sulfides, disulfides, and alkyl-substituted thiophenes to a variety of large polycyclic benzothiophenes.2 The presence of sulfur species in crude oils has a severe impact on oil production and refinery processes. Next to direct corrosive effects on the plant infrastructure and equipment, macromolecular sulfur compounds form a substantial part of the solid asphaltenes and may cause clogging of pipelines.3 Therefore, an important task at production platforms and refineries is to quickly identify the compounds that are harmful for the production plant. Another, well-known drawback of sulfur in crude oils is the release of sulfur oxides (SOx) upon combustion of crude oil based fuels. This environmental effect has led to more and more severe directives on SOx emission. As a result, novel or improved hydrodesulfurization (HDS) catalysts have been
(4) Ali, M. F.; Al-Malki, A.; El-Ali, B.; Martinie, G.; Siddiqui, M. N. Fuel 2006, 85, 1354. (5) Directive on fuel quality, 98/70/EC as amended by 2003/17/EC. (6) Ring, Z.; Chen, J.; Yang, H.; Du, H.; Briker, Y. Proceedings of the AIChE Spring National Meeting, New Orleans, LA, April 25-29, 2004; pp 1386-1406. (7) Xialolan, Z.; Jun, J.; Jianhua, L.; Yongtan, Y.; Chin, J. Anal. Chem. 2006, 34, 1546.
*To whom correspondence should be addressed. E-mail: info@ vibspec.com (P.d.P.),
[email protected] (T.V.). (1) Hua, R.; Li, Y.; Liu, W.; Zheng, J.; Wei, H.; Wang, J.; Lu, X.; Kong, H.; Xu, G. J. Chromatogr., A 2003, 1019, 101. (2) Beens, J.; Thijssen, R. J. High Resolut. Chromatogr. 1997, 20, 131. (3) Marshal, A. G.; Rodgers, R. P. Acc. Chem. Res. 2004, 37, 53. r 2009 American Chemical Society
557
pubs.acs.org/EF
Energy Fuels 2010, 24, 557–562
: DOI:10.1021/ef900908p
de Peinder et al.
desulfurization efficiency for an individual sulfur compound differs with the type of crude.8 Evidently, sulfur speciation of crude oils, either into detail or indicative and fast, is an important task in oil industries. Many analytical techniques have been explored for this purpose, ranging from square wave voltametry9 and liquid chromatography10 to conventional gas chromatography (GC) and two-dimensional GC (GC GC).11-13 A variety of sulfur selective detectors have been used in combination with GC, such as atomic emission detection (GC-AED),14 sulfur chemiluminescence detection (GC-SCD),1,7,15-18 and mass spectrometric detection (MSD).19-24 Other detection techniques have been based on X-ray spectroscopy including X-ray fluorescence (XRF)25-28 and X- ray absorption nearedge structure (XANES) spectroscopy.29-32 Furthermore, potential of temperature programmed reduction and oxidation methods has been studied27,33 as well as the new but powerful technique of Fourier transform ion cyclotron
Table 1. Sulfur Compound Classes As Applied in This Study 1 2 3 4 5 6 7 8 9 10
STD Ar-S BT NBT DBT NDBT BNaT NBNaT DNaT S total
sulfides, thiols, disulfides, thiophenes aryl-sulfides benzothiophenes naphthenic-benzothiophenes di-benzothiophenes naphthenic-di-benzothiophenes benzo-naphthothiophenes naphthenic-benzo-naphthothiophenes dinaphthothiophenes total sulfur amount (including elemental S)
Table 2. Schematic Representation of the 18 Preprocessing Methods Used 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
(8) Yang, Y. T.; Yang, H. Y.; Lu, W. Z. Chin. J. Chromatogr. 2002, 20, 493. (9) Serafim, D. M.; Stradiotto, N. R. Fuel 2008, 87, 1007. (10) Sinkkonen, S. J. Chromatogr. 1989, 475, 421. (11) Beens, J.; Blomberg, J.; Schoenmakers, P. J. J. High Resolut. Chromatogr. 2000, 23, 182. (12) Blomberg, J.; Schoenmakers, P. J.; Brinkman, U. A. Th. J. Chromatogr., A 2002, 972, 137. (13) Blomberg, J.; Riemersma, T.; Van Zuijlen, M.; Chaabani, H. J. Chromatogr., A 2004, 1050, 77. (14) Hegazi, A. H.; Andersson, J. T.; Abu-Elgheit, M. A.; El-Gayar, M. Sh. Polycyclic Aromat. Compd. 2004, 24, 123. (15) Andari, M. K.; Behbehani, H. S. J.; Stanislaus, A. Fuel Sci. Technol. Int. 1996, 14, 939. (16) Behbehani, H. S. J. 219th National Meeting of the American Chemical Society, San Francisco, CA, March 26-30, 2000; American Chemical Society: Washington, DC, 2000; PETR-057. (17) Hua, R.; Wang, J.; Kong, H.; Liu, J.; Lu, X.; Xu, G. J. Sep. Sci. 2004, 27, 691. (18) Lee, I. C.; Ubanyionwu, H. C. Fuel 2008, 87, 312. (19) Glinzer, O.; Severin, D.; Beduerftig, C.; Czogalla, C. D.; Puttins, U. Fresenius’ Z. Anal. Chem. 1983, 315, 208. (20) Dzidic, I.; Balicki, M. D.; Rhodes, I. A. L.; Haskell, I. A. L. J. Chromatogr. Sci. 1988, 26, 236. (21) Payzant, J. D.; Montgomery, D. S.; Strausz, O. P. AOSTRA J. Res. 1988, 4, 117. (22) Nishioka, M.; Tomich, R. S. Fuel 1993, 72, 1007. (23) Sinninghe Damste, J.; Rijpstra, W. I. C.; de Leeuw, J. W.; Lijmbach, G. W. M. J. High Res. Chromatogr. 1994, 17, 489. (24) Ma, X.; Sakanishi, K.; Isoda, T.; Mochida, I. Fuel 1997, 76, 329. (25) Waldo, G. S.; Mullins, O. S.; Penner-Hahn, J. E.; Cramer, S. P. Fuel 1992, 71, 53. (26) Snape, C. E.; Ismail, K.; Mitchel, S. C.; Bartle, K. Speciation of Organic Sulfur Forms in Solid Fuels and Heavy Oils. In Composition, Geochemistry and Conversion of Oil Shales; Snape, C. E., Ed.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 1995; pp 125-142. (27) Snape, C. E.; Yperman, J.; Franca, D.; Bartle, K. Eur. Comm., [Rep.] EUR 1998, EUR 17947, 1–100. (28) Barker, L. R.; Kelly, W. R.; Guthrie, W. F. Energy Fuels 2008, 22, 2488. (29) Waldo, G. S.; Carlson, R. M. K.; Moldowan, J. M.; Peters, K. E.; Penner-Hahn, J. E. Geochim. Cosmochim. Acta 1991, 55, 801. (30) Kasrai, M.; Bancroft, G. M.; Brunner, R. W.; Jonasson, R. G.; Brown, J. R.; Tan, K. H.; Feng, X. Geochim. Cosmochim. Acta 1994, 58, 2865. (31) Sarret, G.; Connan, J.; Kasrai, M.; Eybert-Berard, L.; Bancroft, G. M. J. Synchrotr. Rad. 1999, 6, 670. (32) Mijovilovich, A.; Pettersson, L. G. M.; Mangold, S.; Janousch, M.; Susini, J.; Salome, M.; de Groot, F. M. F.; Weckhuysen, B. M. J. Phys. Chem. A 2009, 113, 2750. (33) Snape, C. E.; Mitchel, S. C.; Ismail, K.; Garcia, R. Rev. Anal. Chem.-Euroanal. VIII 1994, 103. (34) Guan, S.; Marshall, A. G.; Scheppele, S. E. Anal. Chem. 1996, 68, 46. (35) Hughey, C. A.; Rodgers, P. R.; Marshall, A. G.; Qian, K.; Robbins, W. K. Org. Geochem. 2002, 33, 743. (36) Klein, G. C.; Rodgers, R. P.; Marshall, A. G. Fuel 2006, 85, 2071. (37) Hughey, C. A.; Galasso, S. A.; Zumberge, J. E. Fuel 2007, 86, 758. (38) Panda, S. K.; Schrader, W.; Al-Hajji, A.; Anderson, J. T. Energy Fuels 2007, 21, 1072.
MC, 1800-650 MC, 3500-650 MSC, MC, 3500-650 SNV, MC, 3500-650 SNV, Detrend (2), MC, 3500-650 SNV, Detrend (3), MC, 3500-650 SG (25 2 0), MC, 3500-650 SG (25 2 0), MC, 1800-650 SG (25 2 1), MC, 3500-650 SG (25 2 1), MC, 1800-650 SG (25 2 1), MSC, MC, 3500-650 SG (25 2 1), MSC, MC, 1800-650 SG (25 2 1), SNV, Detrend (2), MC, 3500-650 SG (25 2 1), SNV, Detrend (2), MC, 1800-650 SG (35 2 2), MC, 3500-650 SG (25 2 2), MSC, MC, 3500-650 SG (49 2 2), MSC, MC, 3500-650 SG (49 2 2), MSC, MC, 1800-650
resonance (FT-ICR)-MS.3,34-38 Occasionally, infrared (IR) spectroscopy has been used, either including an oxidation pretreatment39 or without it.40 The advantage of IR is that it can be easily performed on location without any preparation of the sample. In previous papers,41-43 we have demonstrated the viability of chemometric modeling IR spectra of crude oils to predict long and short residues properties of crude oils straightforward from their spectra. This method, based on partial least squares (PLS) regression models, has been patented.44 It is currently tested on-site as a fast alternative for the much more elaborate physicochemical American Society for Testing and Materials (ASTM) and Institute of Petroleum (IP) methods used so far. Also, the method turned out to be able to predict the sulfur content with high accuracy. For that reason, a study to the potentials of PLS modeling of IR spectra as a tool for sulfur speciation is a logical next step. This article describes the results of that study using the speciation data obtained from standard GC GC analysis as reference values. Methods and Materials A set of 47 crude oil samples, representing a wide range of geographical oil wells and hence a large variety of different sulfur (39) Saetre, R.; Somogyvari, A. Prepr.-Am. Chem. Soc., Div. Pet. Chem. 1989, 34, 268. (40) Samedova, F. I.; Martynova, G. S.; Yusifov, Y. G.; Guseinova, B. A.; Ismailov, E. G. Azarb. Neft Tasarrufati 2008, 4, 39. (41) De Peinder, P.; Petrauskas, D. D.; Singelenberg, F.; Salvatori, F.; Visser, T.; Soulimani, F.; Weckhuysen, B. M. Appl. Spectrosc. 2008, 62, 414. (42) De Peinder, P.; Petrauskas, D. D.; Singelenberg, F.; Salvatori, F.; Visser, T.; Soulimani, F.; Weckhuysen, B. M. Energy Fuels 2009, 23, 2164. (43) De Peinder, P.; Petrauskas, D. D.; Singelenberg, F.; Salvatori, F.; Visser, T.; Soulimani, F.; Weckhuysen, B. M. Vib. Spectrosc. 2009, 51, 205. (44) De Peinder, P.; Petrauskas, D. D.; Singelenberg, F.; Salvatori, F.; Visser, T.; Weckhuysen, B. M. PCT Patent Application WO 2008/ 135411, 2008.
558
Energy Fuels 2010, 24, 557–562
: DOI:10.1021/ef900908p
de Peinder et al.
Table 3. Crude Oil Samples for Calibration (C1-C28) and Validation (V1-V19) Used for Modeling of 10 Different Sulfur Classesa concentration (ppm) sample
STD
Ar-S
BT
NBT
DBT
NDBT
BNaT
NBNaT
DNaT
S total
C1 C2 C3 C4 C5 C6 C7 C8 C9 C10 C11 C12 C13 C14 C15 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25 C26 C27 C28
14 4 62 53 11 26 586 1803 138 62 385 170 2488 130 269 1848 1039 1622 623 1179 1228 71 1022 333 1448 57 871 757
8 1 254 59 67 15 272 607 91 70 318 830 557 612 1412 421 262 640 289 293 451 239 501 182 520 35 332 247
359 20 1083 461 598 98 811 6791 111 366 2458 2129 6340 2096 2704 4907 950 3400 5689 1052 2821 1606 3070 5114 5452 160 4506 4264
57 3 759 191 332 24 308 929 58 167 600 2149 1193 2020 2306 974 200 867 1075 209 929 492 875 928 1292 58 717 756
742 59 1521 490 1038 85 412 5335 124 218 2097 3312 4357 3193 4248 2810 1371 2033 4872 1475 1694 1790 2448 4264 2652 208 4290 3323
114 4 683 195 442 18 116 1068 37 64 638 1425 1301 1270 1345 757 250 756 1166 277 608 533 844 1124 599 51 1086 681
187 13 519 171 436 24 63 2678 50 31 728 1529 1943 1415 1597 773 322 664 1816 331 540 705 1053 1577 566 57 1987 1101
26 2 197 61 164 9 18 804 16 13 309 632 726 571 589 270 74 289 438 72 268 297 471 471 112 29 714 201
15 1 62 38 111 6 9 870 11 6 205 527 589 522 506 120 78 105 336 62 158 246 436 315 38 0 670 154
7 180 1 370 11 700 5 630 8 350 2 530 9 900 54 200 3 000 4 330 21 900 30 500 50 900 31 400 38 800 47 500 10 800 20 500 45 500 11 100 21 500 14 700 32 200 31 700 30 900 4 580 48 700 37 600
V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11 V12 V13 V14 V15 V16 V17 V18 V19
23 31 684 1759 20 966 252 242 2073 881 1210 736 1279 807 263 638 1149 174 675
22 48 217 687 21 268 403 130 665 292 437 235 374 473 150 235 446 90 518
119 1700 5322 4306 212 5603 2213 5449 2902 3981 5146 3537 3191 2715 1121 608 4123 418 3935
54 357 1063 1055 68 915 760 991 994 760 862 600 705 668 338 124 862 150 1109
175 1534 4053 1980 380 3651 1717 5010 1760 3094 3994 2614 1470 1248 805 370 3096 362 3082
44 351 1005 557 92 785 415 1195 642 876 1010 751 271 236 191 77 871 93 1030
42 364 1476 486 121 1247 430 2070 531 1298 1642 1183 223 223 174 73 1097 99 1148
24 65 326 164 36 266 83 581 259 538 618 518 53 61 79 18 294 41 480
0 32 274 90 0 137 42 456 149 466 473 444 23 34 38 9 175 35 479
4 200 15 400 43 000 33 900 4 840 41 200 25 200 32 300 28 200 44 700 42 600 43 100 20 300 25 000 11 300 8 130 31 400 7 010 28 500
a
Compound class abbreviations refer to names listed in Table 1. Concentrations (parts per million, ppm) have been determined with GC GC.
ATR-intensity correction was not applied. Although the high viscosity of several of the crude oils would make it reasonable to perform the IR measurements at elevated temperatures, all IR measurements were carried out at room temperature (20 °C) for practical reasons and to obtain a high screening velocity. A cover plate was used to prevent evaporation of light ends during measurement. Gas Chromatography. GC GC analysis was performed on a double column Hewlett-Packard P 6890 gas chromatograph (Agilent Technologies) equipped with a CIS4 PTV injector, a sulfur chemiluminescence detector, and a liquid nitrogen cryogenic modulation assembly (Zoex Corp.). The first column was a nonpolar DB-1, dimethylpolysiloxane, 10 m, 0.25 mm i.d., 0.25 μm Df. (J&W Scientific) and the second one a medium polarity stationary phase BPX-50, 50% phenyl(equiv.)polysilphenylene-siloxane, 2 m, 0.10 mm i.d., 0.10 μm Df. (SGE). The modulation capillary was comprised of DPTMDS fused silica tubing, 2 m (1 m in loop), 0.10 mm i.d. (BGB Analytik Vertrieb, Germany). The initial oven temperature for the first dimension column was 40 °C. After an initial hold of
compounds and concentrations, has been used. All samples were stored in a refrigerator at 4 °C and brought to ambient conditions 24 h prior to analysis. Next, samples were homogenized by shaking the sample can every 10 min for 1 h. Experimental protocols on further pretreatment, preparation, and spectral recording have been used throughout the study to ensure the acquisition of reproducible, high quality data. Details on these protocols can be found in ref 41. Modeling of the IR spectra has been carried out for 10 different sulfur classes, as listed in Table 1, i.e., the total sulfur content and 9 sulfur speciation groups, commonly used in GC GC analysis. IR Spectroscopy. IR measurements have been carried out at room temperature on a Bruker Tensor-27 Fourier transform infrared (FT-IR) spectrometer equipped with a DTGS detector. The sample compartment was flushed with dry air to reduce interference of H2O. Spectra were recorded with a horizontal ATR accessory (FastIR, Harrick Scientific Products) with a ZnSe crystal as the internal reflection element. The spectral resolution was 4 cm-1 for all spectra, and 50 scans were accumulated with medium apodization for each spectrum. 559
Energy Fuels 2010, 24, 557–562
: DOI:10.1021/ef900908p
de Peinder et al.
5 min, the oven was programmed at a rate of 2.5 °C/min to 320 °C, which was maintained for 20 min. The secondary oven chamber for the second dimension column had an initial temperature of 90 °C. After an initial hold of 5 min, it was programmed at a rate of 2.5 °C/min up to 370 °C, which was maintained for 20 min. The hot-pulse duration was set to 500 ms, and the modulation time was 10 s. Samples were injected either pure or, when viscosity at 60 °C or S-content did not allow so, diluted with toluene and/or cyclohexane. Concentrations of components in parts per million sulfur (ppm S) were calculated by means of a classified internal standard. Chemometrics. Modeling was performed using the PLS Toolbox (Eigenvector Research, Inc.) for MatLab (The MathWorks, Inc.) on the IR spectra of the 47 crude oils. Principle component analysis (PCA) of the total data set was applied to obtain representative subsets for calibration and validation. A group of 28 spectra was selected for calibration (samples C1-C28). The remaining 19 spectra were used for validation (samples V1-V19). As input for modeling of the nine different sulfur compound classes, the concentrations as determined with GC GC have been used. Modeling for the total sulfur content was carried out on the data as determined according to ASTM method D2622. Prior to modeling, a baseline correction was applied to the IR spectra by subtracting a third degree polynomial fit using the regions 4000-3500, 2500-2000, 1900-1800, 1560-1520, 1000-990, and 650-600 cm-1. Subsequently, the region 2500-1800 cm-1 was removed from the spectra since no absorbance bands were observed in this region. Next, preprocessing of the IR spectra was optimized for all 10 sulfur classes by systematic varying preprocess parameters like scaling, smoothing, region selection, and spectrum derivative options. This resulted in a selection of 18 different preprocessing methods, based on previous modeling experience with this data set,41-44 which are listed in Table 2. MC refers to mean centering and was applied in all cases. The spectral range was either 3500-650 or 1800-650 cm-1. For scaling, either the option “none”, multiplicative signal correction (MSC), or standard normal variate (SNV) with and without detrending, second, or third order polynomial, was applied. The Savitzky-Golay (SG) smoothing and differentiation parameters were varied from 25 to 49 points, using a second order polynomial and none, first, or second derivative. As an example, preprocessing method 13 comprises a SG smoothing with 25 points using a second order polynomial and taking the first derivative followed by SNV, detrending with a second order polynomial, and MC on the 3500-650 cm-1 region. For each of the 18 preprocessing methods, PLS modeling was carried out for the 10 sulfur classes, which resulted in 180 models. From these, the 10 models with the lowest root-meansquare-error-of-prediction (RMSEP) value for each of the sulfur classes were selected for concentration prediction.
Figure 1. GC GC plot of crude oil C21. S-compound classes and the internal standard have been indicated. White colors represent high concentrations and black colors low concentrations.
Figure 2. Overlay of 28 spectra of crude oils as used for calibration of the PLS-models.
point range from ambient to 370 °C, whereas the ASTM method includes also higher-boiling compounds as well as elemental and inorganic sulfur. Infrared Spectroscopy. As reported before,41-43 the IR spectra of crude oils are very similar, particularly after intensity normalization and preprocessing. This is illustrated in Figure 2, showing the overlay of the 28 baseline corrected crude oil spectra of the calibration set C1-C28. All spectra are dominated by strong absorption bands of aliphatic C-H stretching (3000-2800 cm-1) and bending (1470-1350 cm-1) vibrations. Small differences are present in the fingerprint region (1300-650 cm-1). The absorption bands in this region can be merely attributed to aromatic skeletal modes. In general, specific C-S, S-H, and/or S-S vibrations are not very IR active because of the small dipole moment change during the vibration of these structural elements.45 However, for example, thiophene rings exhibit several sharp bands related to ring stretching (1550-1350 cm-1) and dC-H out of plane vibrations (800-690 cm-1).45,46 Data Analysis. PLS modeling of the 10 sulfur concentration classes, using 18 different preprocessing methods, resulted in 180 models. For these 180 models, the RMSEP
Results and Discussion Gas Chromatography. The concentrations in ppm of the 9 different sulfur compound classes and S total, as determined with GC GC analysis and ASTM method 2622, respectively, are presented in Table 3. Calibration samples C1-C28, used for building the models, as well as samples applied for validation (V1-V19) represent a wide range of concentrations for the different sulfur speciation classes, which validates PCA of the IR spectra for this study. To illustrate the results from GC GC, a retention timeintensity plot for crude oil C21 is shown as an example in Figure 1. Note that the physico-chemically determined amount of “total S” differs from the summed concentrations as measured with GC GC. This is due to the fact that the latter method only covers compounds that elute in the boiling
(45) Lin-Vien, D., Colthup, N. B., Fateley, W. G., Graselli, J. G., Eds. The Handbook of Infrared and Raman Characteristic Frequencies of Organic Molecules; Academic Press, Inc.: San Diego, CA, 1991. (46) Infrared and Raman Interpretation Support software, IRIS 3.0, Thiophenes, http://www.vibspec.com.
560
Energy Fuels 2010, 24, 557–562
: DOI:10.1021/ef900908p
de Peinder et al.
Figure 3. RMSEP values divided by the standard deviation for the 10 sulfur speciation classes for 18 different preprocessing methods. Table 4. Results of Optimized PLS Models to Predict the Concentrations of 10 Different Sulfur Compound Classes in Crude Oilsa preprocessing (method S class number) STD Ar-S BT NBT DBT NDBT BNaT NBNaT DNaT S total
15 10 18 15 9 7 9 11 13 9
mean concn (ppm) 586 328 2161 666 1976 585 773 281 227 21587
STDEV RMSECV RMSEP (ppm) (ppm) LVs (ppm) 585 302 1864 625 1553 451 722 253 243 16351
448 229 769 306 632 213 367 95 131 5403
6 4 6 6 10 8 5 8 4 7
537 147 700 228 383 187 331 194 199 2520
a Mean concentration, preprocessing method, STDEV, RMSECV, and LV values refer to the calibration set and RMSEP values to the validation set.
values obtained for each speciation class were divided by the standard deviation of the calibration values to express the relative error. These relative errors have been plotted as a function of the preprocessing method for each sulfur speciation class in Figure 3. The figure illustrates that, independent of the applied preprocessing method, some classes (e.g., NBT) are better predicted than others (e.g., STD). In our opinion, this demonstrates the ability of the models to extract structure related correlations from the IR spectra. Next, the models with the lowest RMSEP values for each of the 10 classes were selected for further evaluation. This is summarized in Table 4, showing for each class the applied preprocessing method, the mean concentration value for the calibration set, the corresponding standard deviation, the root-mean-square-error-of-validation (RMSECV) value, the number of latent variable (LVs) that was used for the model, and the RMSEP values obtained for the validation set. In addition, the corresponding plots of the predicted versus the measured concentrations for the calibration set (b) and the validation set () are shown in Figure 4. First of all, the results confirm the conclusion from our previous papers41-43 that the prediction of the total sulfur concentration of crude oils by means of PLS modeling of the IR spectra is a valuable alternative for ASTM method 2622. The models to predict the dibenzothiophenes (DBT) is promising followed by the related benzothiophene compound classes BT, NBT, and NDBT. The correct prediction of DBT concentrations is particularly interesting in view of the fact that these compounds are the major sulfur containing species left in fuels after hydrodesulfurization. Moreover,
Figure 4. Prediction plots of PLS modeling the concentration of 10 sulfur speciation classes of crude oils based on their IR spectra. Calibration spectra (b) and validation spectra (). The corresponding preprocessing methods are listed in Table 4.
the models for the speciation of DBT together with BT, NBT, and NDBT might be useful, as this type of compound is known to hamper efficient crude oil processing and refining. Furthermore, we conclude that the models for the remaining classes STD, Ar-S, BNaT, NBNaT, and DNaT are less useful for concentration prediction. The differences in the predictive power of the models can be explained by the assumption that vibrations related to benzothiophene structures are well represented in the IR spectra, whereas other sulfur containing functional groups lack specific sulfur related absorption bands. 561
Energy Fuels 2010, 24, 557–562
: DOI:10.1021/ef900908p
de Peinder et al.
Closer examination of the correlation plots in Figure 4 reveals that the concentration distribution for some of the sulfur groups (e.g., Ar-S and NBT) is not homogeneous. This is due to the fact that the PCA to obtain calibration and validation subsets was based on spectral variation and not on differences in sulfur specie concentrations. It cannot be excluded that, in some cases, PCA based on this parameter might lead to better models. Finally, it should be noted that in refineries effluents have lower average sulfur concentrations. This requires the development of dedicated models. However, as demonstrated in this study, several of the models developed for crude oils also perform well at low sulfur concentrations. Therefore, we believe that PLS modeling of the IR spectra of refinery effluents might have similar potential for sulfur speciation purposes.
sulfur speciation, however, is limited. From the nine different sulfur compound classes that are usually determined with standard GC GC analysis, the models to predict the concentration of DBT and the related benzothiophene compound classes BT, NBT, and NDBT perform reasonably well. However, the models for the remaining classes STD, Ar-S, BNaT, NBNaT, and DNaT are less useful. As such, PLS regression modeling is not as widely applicable for sulfur speciation as GC GC. On the other hand, it can be a fast, clean, and nonelaborate method for qualitative and quantitative on-site or even in situ screening of crude oils on (di-) benzothiophenes, a class of compounds which is known to be detrimental in crude oil processing and a predominant sulfur-residual in fuels.
Conclusions
Acknowledgment. This work was carried out by financial support of Shell Global Solutions International B.V., The Netherlands. Dr. D. Petrauskas and Dr. F. Salvatori are gratefully acknowledged for providing the crude oil samples and permission to use the corresponding data.
PLS modeling of the IR spectra of crude oils is a valuable alternative to ASTM method 2622 to predict the total sulfur content of these materials. The application as a tool for
562
