Heteroelements frequently make up several percent of fossil materials like petroleum. Compounds containing the heteroatoms sulfur,nitrogen and oxygen can cause undesirable chemical reactions in fuels,resulting in fuel discoloration,gum formation,plugging of filters,etc [1, 2]. In many cases,heteroatom-containing compounds (HACC) have been found to be more problematic in the environment than similar hydrocarbons,for instance with respect to toxicity [3, 4]. As a result,investigators aim to identify HACC in petroleum fractions as a pre-analytical step. Without the knowledge of the presence of these compounds or not,traditional GC identification techniques of HACC in petroleum fractions,however,waste efforts,chemical materials and time. Presence,configuration and concentration are three analytical factors that affect HACC analysis in petroleum condensate,volatile and crude oils. The key-step in HACC analysis is to identify,and if possible quantify,the HACCs in such complex mixture.
Condensate oil contains a considerable concentration of polycyclic aromatic sulfur heterocycles (S-PAHs) [5] that are emitted to the environment as carcinogens in the case of incomplete fuel combustion [3]. According to oil spill pollution research S-PAHs are more resistant to microbial degradation,accumulated to more degree,than their poly-aromatic compound (PAC) analogs particularly in fresh water environment [6-8]. In addition,sulfur dioxide (SO2) is emitted during complete combustion of gasoline or condensate containing these sulfur compounds. These emissions have hazardous effects on human life,soil and air and are main contributors to acid rain.
The composition of condensate oil,as a gasoline-like fraction,differs from that of crude oil. It must therefore be shown whether the analytical methods applied to crude oil are suitable for oil condensate or not. Nitrogen and oxygen containing compounds exist different configurations in crude oil [2, 3, 9, 10]. What about the presence of these compounds in the light petroleum fractions as the condensate and volatile oils? Answers to all these questions and others require difficult analytical tests and attempts that waste chemical materials,efforts and time. Therefore,simple analytical technique is needed to identify these compounds in given petroleum fraction as pre-informative analytical method for other selective techniques.
Today mass spectrometry (MS) is an indispensable analytical tool for petroleum analysis and that includes the identification by means of MS techniques such as mass spectral library searches or/and fragmentations pattern interpretation [11]. These MS techniques can be used as pre-informative tool for many selective analytical methods (atomic emission (AED),sulfur chemiluminescence(SCD),nitrogen-phosphorus detectors (NPD),etc) to characterize the petroleum fraction. Traditional use of GC/MS involves hydrocarbon type analysis after extraction by different techniques [12-17] to overcome HACC-hydrocarbons detection interference. For this purpose,GC/MS techniques have been used for petroleum fraction analysis: crude oils [18],middle distillate [14] and jet fuels [19]. Several extraction techniques have been used without prior knowledge of the presence and/or configuration of a target HACC in the given petroleum fraction. Particularly,condensate oils have different analytical behavior than crude oils,so it can not resort to the expectation. This search analytical step about the presence and/or configuration of target HACC waste chemical materials,efforts and time. Furthermore,the result is negative in many cases. Another difficulty is the detection of lower concentration of some HACC by commonly available selective analytical techniques due to hydrocarbon interference,e. g. sulfur detection quenching effect,or the extraction technique overlooked these trace concentrations. Non-availability of standards for this huge number of petroleum isomers is another identification problem.
Kind and Fiehn [11] discussed the importance of MS software development for the advancement of structure elucidation of small molecules. NIST (National Institute of Standards and Technology) program AMDIS (Automated Mass Spectral Deconvolution and Identification System) as free software package is traditionally used for deconvolution of GC-MS spectral data [20] but one report on the use of AMDIS for the identification of 2,8-dimethyl[b,d] dibenzothiophene in a middle distillate by selective-ion monitoring (SIM) technique has appeared [21]. Reprocessing to enhance the deconvolution performance and overcome false positive of AMDIS has been published [22]. GC-MS-AMDIS technique is used prominently for metabolomics analysis as multi-components complex mixture [23]. This pre-identification step can very easily produce compositional information that cannot be obtained by other routine methods for petroleum analysis. To our knowledge,despite the availability of AMDIS,it suffers from complete disregard as identifier especially in the oil analysis. The aim of this study is to use AMDIS to identify parts per million (ppm) amounts of HACCs in a petroleum condensate oil as a pre-informative step for other selective detection techniques. For this purpose,we wish to test the validity of AMDIS as an identification tool rather than use it as a deconvolution software in petroleum analysis. The study aims also to examine the efficiency of AMDIS as identifier program and whether there are any defects in deconvolution program or not [22]. The proposed identification procedure can be used instead of expensive selective detection techniques if propose a GC/MS quantitative analytical method.
The petroleum condensate samples were collected directly from the separator at the head of the well. The condensate oil samples were kindly provided by El-Hamara Company,Egypt. The volatile oil sample was from VE-Gas Petroleum Company.
The conditions were described in detail in reference [5]. The GC-MS analysis was carried out on a GCQ-Finnigan MAT chromatograph with an ion trap mass analyzer. The conditions were: initial temperature 60℃ for 1 min,ramp at 5 ℃/min to 300 ℃,held for 20 min. The capillary column was a DB-5 from J&W (30 m×250 μm×0.25 μm).
For spectrum reprocessing and identification,two softwares were used: off line version of X-calibur-GCQ-Finnigan MAT chromatograph and the NIST program,AMDIS. AMDIS program was obtained from NIST (Gaithersburg,MD) as part of the NIST’98release of the NIST/EPA/NIH mass spectral library. The simple AMDIS identification approach was based on the following steps: (i) open MS spectrum raw format,X-calibur raw format in our case,by AMDIS program; (ii) analyze GC-MS data for retention time reading and peak area calculations (analyze menu); (iii) select the target m/z value (option menu) and (iv) match unknown/known spectra (Library menu) to confirm the true peak identification.
Contents of alkylphenols condensate oil were determined via ferrocene carboxylic acid chloride(FCC) derivatives formation according to the published procedure in reference [24]. Briefly,in a 5-mL sample vial,0.150 mmol (about 15-20 mg) of the petroleum condensate oil was mixed with a solution of 0.160 mmol (40.9mg) of FCC in 0.82 mL of dichloromethane and a solution of 0.221 mmol (27.4 mg) of 4-(dimethylamino)pyridine (DMAP) in 0.55 mL of dichloromethane. After 1 min,the DMAP and the excess of FCC were removed by separation on a minicolumn (50 mm×5 mm i. d. of alumina). The esters were eluted with 5 mL of dichloromethane.
Briefly,the sample oils were applied to the columns. The components were adsorbed onto neutral alumina and eluted with n-hexane,toluene,chloroform,and methanol to obtain aliphatic hydrocarbons,aromatic hydrocarbons,nitrogen compounds,and hydroxyl aromatic compounds,respectively [25].
This issue was described in detail in reference [5]. Briefly,silica column was used to separate aliphatic and aromatic fractions of condensate petroleum oil. Pd (Ⅱ)-silica column was used to separate S-PAHs from aliphatic fraction.
The chromatogram of the pre-isolated aliphatic hydrocarbon fraction of petroleum oil condensate [5] is shown in Fig. 1. S-PAHs were separated from crude oil distillates within the aromatic fraction [5] by cyclohexane-dichloromethane elution on silica columns. For oil condensate,the pre-isolated aromatic fraction was injected using the traditional approach into a gas chromatograph installed with an atomic emission detector to identify S-PAHs. However,the fractions of 11 condensate samples were free from S-PAHs and,unexpectedly,S-PAHs were detected in the aliphatic fraction [5]. Therefore,we adopted a pre-informative theoretical identification approach. Fig. 1 shows the hydrocarbons distribution of pre-isolated aliphatic fraction for oil condensate. AMDIS could identify S-PAHs in the GC-MS spectrum of oil or its aliphatic fraction (Fig. 1). In the analytical scheme,m/z entries to the AMDIS-mass spectrum of the oil or its fractions can be replaced with a number of aromatic or aliphatic fraction injections to identify S-PAHs. For example,ppm concentrations of methyl (C1-),ethyl (C2-) and propyl (C3-) dibenzothiophenes (DBTs) were identified in the aliphatic fraction using X-calibur software (Fig. 1),which is comparatively simpler than the traditional technique. The aliphatic fraction represents the main composition of condensate oil so the direct identification of S-PAHs in this fraction reduces many experimental isolation steps.
The accuracy of AMDIS for detecting ppm concentrations of C1-,C2-,and C3-DBTs in the same aliphatic fraction is shown in Fig. 2,and the results were consistent with the compositional results by X-calibur. Simulated carbon-12(C12) and sulfur-181 (S181)-selective AED traces for the aliphatic and S-PAH fractions of a petroleum sample,respectively,are shown in Fig. 2 [5, 26].
The chromatogram of the S-PAH fraction detected experimentally by AED-S181 is shown in Fig. 3 [5],which illustrates that the same chromatogram details are obtained as with AMDIS (Fig. 2). The S-PAH fraction was separated from the aliphatic fraction or condensate oil using the nano-Pd (Ⅱ) SPE technique [5, 27],validating the AMDIS identification of S-PAHs in the condensate oil. Moreover,the comparison between Figs. 1 and 3 confirms this result where the identified S-PAHs groups in the aliphatic fraction theoretically are the same groups identified in pre-isolated PAHs fraction experimentally. The concentrations of S-PAHs ranged from 380 to 3 430 ppm according to our previous quantitative study [5]. Furthermore,each group of S-PAHs isomers was eluted within two bracketing n-alkanes starting from n-C19. Accordingly,Kovats retention indices could be calculated using these bracketing n-alkanes. To the best of our knowledge,this is the first example of calculating Kovats retention indices using bracketing n-alkanes,with all previously published Kovats retention indices for S-PAHs based on bracketing poly-aromatic compounds.
AMDIS pre-identification reduces the injection trails of the given fractions on GC detectors and,therefore,reduces the number of analytical steps. There is a difference in analytical behavior between petroleum fluids that settle in crude oil aromatic fractions containing S-PAHs and condensate oil aliphatic fractions containing S-PAHs. It is well known that traditional analytical methods used to detect biomarkers in a crude oil require several preparations and are time-consuming [28],where the presence of biomarkers in a given petroleum fraction or not is the analytical question. The comparatively simpler AMDIS identification of pentacyclicterpanes (m/z 191) in a crude oil than traditional methods is shown in Fig. 4. Theoretically,the following biomarkers were identified: 1=18a(H)-22,29,30-trisnorhopane (Ts); 2=17ct(H)-22,29,30-tgimorhopase (Tm); 3=unknown; 4=17tz(H),21p(H)-norhopane (C29H); 5=17ct(H),21p(H)-hopane (C30); 6=homohopanes (C31); 7=CS2 bishomohopanes (C32) and 8,9=C33. There was a good correlation between the experimental [28] and AMDIS biomarker identification results,and it can be seen that the biomarker identification in the given oil can be simplified to a single injection of a very dilute sample of crude oil in GC/MS. These results confirm the validity of the AMDIS identifier program.
The traditional identification of NCCs in petroleum condensates is time-consuming and wastes chemicals. Therefore,we used AMDIS to detect NCCs in petroleum oil condensate for several NCC m/z values [2, 10, 13, 14]. AMDIS detection of m/z=149 in the aliphatic and S-PAH fraction chromatograms for the petroleum condensate sample is shown in Fig. 5. Two main peaks (shown in red) at 35.84 and 48.34 min were detected. The chromatogram obtained by AED nitrogen-388a (N388a) of the S-PAH fraction contains one peak at 35.84 min. Further analysis confirmed that this peak is 2-methylbenzothiazole (Mr=149.213; boiling point (B. P.)=238 ℃) because: (i) the peak detected by AED for both S181 and N388a in the S-PAH fraction indicated that the compound contains both S and N atoms; (ii) a library search of the best-fit m/z=149 corresponded to 2-methylbenzothiazole [29],which eluted within C2-DBT m/z=226; and (iii) in contrast to NCC,this compound eluted on the Pd (Ⅱ) column [1],indicating that this compound configuration permits the formation of a weak,but not a strong,S-Pd (Ⅱ) bond.
Our results raise the possibility of studying the separation of benzothiazole isomers on the Pd (Ⅱ) column,similar to the separation of S-PAHs. The other peak (m/z=167) was not detected by AED-N338a in the S-PAH fraction,indicating that the compound does not contain sulfur and nitrogen atoms but probably only contains nitrogen atoms. Furthermore,NCC did not elute on the Pd (Ⅱ) column. According to a library search,the best fit m/z=167 corresponds to carbazole (Mr=167; B. P.=354 ℃),which co-eluted within C7-DBT m/z=282. These findings were confirmed empirically by successful separation of the two compounds as fractions [25],as shown in Fig. 6.
Alkylphenols were identified in the oil condensate using the same approach rather than traditional analytical methods based on SPE,a suitable derivatization method,and GC analysis. For 11 alkylphenol [2] m/z entries,AMDIS identified one compound in the aliphatic fraction of the oil condensate (Fig. 7) corresponding to a peak of M+=191 (m/z) at 23.56 min. According to a library search for the best fit,M+=191 (m/z) corresponds to 2,4-ditertbutyl phenol (Mr=206; B. P.=253 ℃),which co-eluted with an n-C17 alkane. This was confirmed empirically by AED-iron 302 detection of 2,4-ditertbutyl phenol in oil condensate as ferrocene carboxylic esters [24, 30]. AED-iron 302 detection of two alkylphenol ferrocene carboxylic acid esters is shown in Fig. 8,which are phenol as standard and 2,4-ditertbutyl phenol in the condensate sample.
Traditional methods for the ppm detection and identification of HACC in petroleum fractions waste chemicals,time,effort,and,in many cases,produce negative or inconclusive results. Only the MS format of the sample is required for theoretical identification of ppm HACC using AMDIS in any given petroleum fraction. This can be regarded as a pre-informative analytical step for the selective detection of ppm concentrations of HACC in petroleum fractions. AMDIS can be used to determine biomarkers in oil fractions instead of the traditional time-consuming procedure. AMDIS identified ppm concentrations of HACC in the chromatograms of petroleum oil and overcame the deficiencies of standard methods. Our approach is recommended to protect different areas of the gas chromatograph from repeated injections and chemicals and time wastage in preliminary experiments.
2016,
Vol. 34 Issue (9): 925-932
PDF
2016,
Vol. 34 Issue (9): 925-932
PDF