Pyrolysis-Gas Chromatography/Mass Spectrometry Analysis of Oils from Different Sources

Regenerated gutter oil (i.e., waste oil) accounts for 10% of the edible oil market, which has caused serious food safety issues. Currently, there is no standard protocol for the identification of the gutter oil. In this study, the pyrolysis-gas chromatography/mass spectrometry (Py-GC/MS) method was employed to analyze eleven oil samples including edible vegetable oils (tea oil, corn oil, olive oil, sunflower oil, peanut oil and blend vegetable oil) and waste oils (used frying oil, lard, chicken fat, inferior oil and kitchen waste grease). Three factors of pyrolysis temperature, reaction time and sample volume were investigated to optimize the analytical parameters. The optimal pyrolysis conditions were determined to be 600°C, 1 min and an injection volume of 0.3 μL. Five characteristic components (tetradecane, z,z-9,12-octadecadienoic acid, decanoic acid-2-propenyl ester, 17-octadecenoic acid, and z-9-octadecenoic acid) were found in all oil samples. The existence of C11-C16 olefins in the pyrolytic products of the animal fats and the other low-quality oils could be utilized to distinguish vegetable oils from gutter oils.


Introduction
In the past ten years, food safety issues related to the reuse of waste oil or grease (i.e., gutter oil) have been frequently exposed [1]. It is estimated that the regenerated waste oil accounts for up to 10% of the cooking oil market, i.e., about 2.5 to 3 million tons of waste oil returns to the dining table every year [2]. As edible oils are a necessity in everyday life, the National Health Department of China began to focus on strengthening the techniques to detect and analyze edible oils.
In addition to the conventional physical and chemical indicators, the current detection/analytical methods of waste oils include various chromatographic methods, spectroscopy, nuclear magnetic resonance, etc. [3][4][5]. However, due to the complicated sources of waste oil, the complex composition, different processing methods, and different refining degrees, there is no single specific indicator or standard to distinct waste oils from edible oils. Consequently, it is imperative to develop a standard analytical method for the detection of the waste oil.
Because of the high boiling point, food oils are hardly to be analyzed directly. Therefore, the oil or grease is usually methylated and then analyzed by gas chromatography (GC) or gas chromatography coupled with mass spectrometry (GC/MS) [6]. In terms of the pyrolysis-gas chromatography/mass spectrometry (Py-GC/MS) technology, oils can be directly pyrolyzed and the small molecules produced by the pyrolysis process are further identified by GC/MS [7]. The obtained pyrolytic products are a very intuitive reflection of the cracked fragments of the oil, which is equivalent to a series of changes in the simulated oil under pyrolytic temperature conditions [8]. The pyrolysis reactor adopts a vertical micro-furnace structure to measure the temperature of the sample in real time. The pyrolysis results demonstrate good reproducibility and overcome the deficiency of easy loss of high-boiling substances, which is conducive to obtaining more accurate analysis results [9].
In this study, eleven different oil samples were collected. The samples included vegetable oils (tea oil, olive oil, peanut oil, corn oil, sunflower oil, and blend vegetable oil), animal fat/oils (lard and chicken fat), and some low-quality oils (used frying oil, kitchen waste grease, and inferior oil). Py-GC/MS was conducted to analyze the pyrolytic products and characteristic peaks of oils from different sources.

Sample Collection and Preservation
The samples of this study mainly included two categories: edible vegetable oils and waste oils (used frying oil, lard, chicken fat, inferior oil and kitchen waste grease). The edible vegetable oils were purchased from the supermarket. The used frying oil and animal oils (chicken fat and lard) were collected from the home kitchen following cooking. The inferior oil with a very low price was purchased from the market. The waste grease was collected from the dining hall of the University. The sample names and the sources are summarized in Table 1. All samples were stored at room temperature.

Pretreatment of Oil Samples
The oil samples of 9-11 (i.e., waste grease, lard, and chicken fat) contained a small amount of water. Therefore, a pretreatment was conducted to remove the moisture from these oils. Firstly, an appropriate amount of oil sample was poured into the centrifuge tube, and then an appropriate amount of anhydrous sodium sulfate was added to the centrifuge tube. The centrifuge tube was vortexed and the water absorption of the sodium sulfate can be observed. In case, if there is no floating matter aggregates, it is still necessary to add a small amount of sodium sulfate until granular particles appeared.
Finally, the centrifuge tube was centrifuged at 3000×G for 20 minutes. Then, the supernatant was carefully collected as the pretreated oil sample.

Pyrolysis Coupled with Gas Chromatograph/Mass Spectrometer (Py-GC/MS)
Pyrolysis of oil samples was conducted in a sample cup of Frontier PY-2020iD pyrolyzer (Fukushima, Japan). For each experiment, the pyrolyzer was pre-heated to the desired temperature (300°C, 400°C, 500°C or 600°C), and then purged with ultra-purity helium to remove oxygen. A certain amount of samples (0.1 μL,0.3 μL, or 0.5 μL) was allowed to drop into the pyrolyzer, whereby the sample was pyrolyzed for 30 s, 1 min, 3 min or 5 min. The volatilized products were injected directly into a Shimadzu GCMS-QP2010 gas chromatograph/mass spectrometer (Shimadzu, Japan) equipped with a Frontier Ultra-Allov5 capillary column (Fukushima, Japan).
For GC/MS analysis, the carrier gas of helium (99.999% purity) with a flow rate of 1 mLmin -1 and the split ratio of 50:1 were used. The inlet temperature of GC was maintained at 300°C. The temperature of the GC oven was initially set at 35°C and held at 35°C for 2 min, then ramped to 350°C at a rate of 15 °Cmin -1 and held at 350°C for 10 min. The pyrolytic products were identified by comparison with the NIST mass spectral library (National Institute of Standards and Technology, USA). The distribution of compounds was calculated as the peak area percentage.

Results and Discussions
This study attempted to optimize the detection method of the waste oils, mainly from the three influencing factors of pyrolysis temperature, the sample amount, and the pyrolysis residence time. The pyrolysis temperature refers to the temperature whose sample is pyrolyzed in the pyrolysis furnace, i.e., the temperature before entering the GC column.

Impact of Pyrolysis Temperature
The direct pyrolysis of the waste oils without methyl esterification was performed by Py-GC/MS and the parameters were optimized accordingly. Firstly, the effect of the pyrolysis temperature was studied. Because the smoke point of edible oils starts at 170°C, a lower pyrolysis temperature of 150-200°C was first studied. However, it was found that the pyrolysis at the low temperature was difficult to obtain the volatile effluent, and almost no pyrolytic products appeared. Therefore, the pyrolysis temperature was further increased to 300°C, 400°C, 500°C and 600°C. Taking sunflower oil as an example, the experiments were carried out under the conditions of the sample volume of 1 μL and the pyrolysis time of 1 min. The total ion current (TIC) chromatograms are shown in Figures  1 and 2.
Comparison of Figure 1 with Figure 2 shows that as the pyrolysis temperature rose from 300°C to 600°C, the number of pyrolytic products gradually increased, resulting in more peaks on the TIC chromatogram. The resolution was higher at 600°C which is determined as the optimal pyrolysis temperature in this study.    The comparison with the blank chromatogram after pyrolysis shows that when the injection volume was 0.1 μL and 0.3 μL, the amount of residue in the GC column was relative negligible. When the injection volume increased to 0.5 μL, the amount of residue in the column was more evident. This may affect the analytic results of the following samples. Additionally, the peaks of the TIC chromatogram were not clear for the sample injection of 0.1 μL. Therefore, the optimal injection volume was determined as 0.3 μL in this study.

Optimization of Pyrolysis Reaction Time
Pyrolysis time was investigated at the pyrolysis temperature of 600°C and an injection volume of 0.3 μL. Times studied were 30 s, 1 min, 3 min, and 5 min. The TIC chromatogram in Figure 6 shows very similar results under the reaction time of 0.5 to 5 min. However, when the pyrolysis time was greater than 1 min, the peak intensities of the total ion current were more evident than those of 0.5 min. Accordingly, the optimal pyrolysis time was determined as 1 min.

Pyrolysis of Oils from Different Sources
The oil samples including tea oil, olive oil, peanut oil, corn oil, sunflower oil, vegetable blend oil, used frying oil, lard, chicken fat, inferior oil and kitchen waste grease were pyrolyzed at 600°C and a volume of 0.3 μL for 1 min. The TIC results are shown in  The TICs of all oil samples were quite complicated in terms of the number of peaks and the peak shape. Because vegetable oils or animal oils are essentially fatty acid glycerides, the resulting TICs after pyrolysis were very similar. Nevertheless, the TICs of oil samples from different sources could be distinguished by either the retention time for different compounds or the peak height/area for the same compound.
A specific peak, named as Peak 1 was observed at the retention time of 9.5 min. This peak was identified as tetradecane by searching through the NIST library. The comparison of Peak 1 of different oil samples is listed in Table 5. The area of Peak 1 of all edible vegetable oils was less than 2.0E+05, and the peak height was less than 1.50E+05. And the similarity of all edible vegetable oils in this peak was less than 92%, while the results of animal oils, used frying oil, inferior oil, and kitchen waste grease showed opposite trends. This feature may be employed as an evaluation indicator to distinguish vegetable oils from lard, chicken fat, kitchen waste grease, and inferior oil.
Two other distinct peaks appeared between 14 and 16 minutes were marked as Peak 2 and 4, respectively. These two peaks showed obvious higher peak intensities. A smaller peak between Peak 2 and 4 was marked as Peak 3. To be more specific, Peak 3 could be distinguished into two very close small peaks, labeled as Peaks 3-1 and 3-2. The height of these peaks of various oil samples is summarized in Table 6. For most vegetable oils, the height of Peak 2 was shorter, but the height of Peak 4 was higher. In terms of the peak height ratio of these two peaks, the ratio of H#4/H#2 was the largest for vegetable oils. For animal oils and other low-quality oils, this ratio was small. For example, the height of Peak 2 of the inferior oil was slightly higher than that of Peak 4 with a ratio of 0.20. However, corn oil and kitchen waste oil did not conform to the above rules. This ratio (2.09) for corn oil was not as large as other vegetable oils, while kitchen waste grease had a sufficient height difference with a ratio of 4.25. The height of Peak 3-2 of all oils and fats peaks was relatively close. But the height of Peak 3-1 was obviously different, i.e., the peak heights of all edible vegetable oils were less than 3.00E +05 and others were greater than 3.00E+05. Therefore, edible vegetable oils can be distinguished from other fats.

Analysis of Pyrolytic Products of Oils from Different Sources
Because the structure of the pyrolytic products following Peak 4 was relatively complex and the similarities of the corresponding chemicals were low, this study specifically analyzed the pyrolytic products prior to Peak 4 and compared the similarity of various oils. The main ingredients (about 90%) are listed in the following Tables 7-17.       According to these results, during the first 6.5 minutes, the pyrolytic products of all oil samples were quite similar, most of which were small-molecule chemicals such as 2-acrolein, hexene, heptane, aldehydes, and olefins. Moreover, these substances had a higher similarity, mostly over 90%.
For animal fat/oils, inferior oil, and kitchen waste grease, pentadecane (C15) was observed at the retention time of 10.4 min, and the similarity was higher than 90%. Other vegetable oils did not show pentadecane in the pyrolytic products.
Peak 2 was identified as z,z-9,12-octadecadienoic acid, while Peak 3-1 was identified as decanoic acid-2-propenyl ester. Due to its low strength, Peak 3-2 was identified as 17-octadecenoic acid, but the potential was low. For the used frying oil, animal fat/oils, and inferior oil, Peak 4 was mainly z-9-octadecenoic acid. But Peak 4 of vegetable oils could also be a mixture of z-9-octadecenoic acid and z,z-9,12octadecadienoic acid.
As shown in the mass spectrum, not all olefins having a carbon number higher than 11 (undecane) were present in the pyrolytic products of vegetable oils. For example, dodecane, tridecane, and pentadecene were absent from the products of tea oil, olive oil, and peanut oil. But the products from animal fats, used frying oil, and inferior oil contained all kinds of C11-16 olefins (Table 18). The possible reason is that these oils have been used and recovered, wherein the C16-C18 fatty acids were degraded to a certain degree. So, the pyrolytic products of these low-quality oils contained all kinds of olefins. This can be used as a key indicator to distinguish inferior oils and animal fats from vegetable oils.

CONCLUSIONS
The pyrolysis conditions of oil samples were optimized as the pyrolysis temperature of 600°C, the sample volume of 0.3 μL, and the reaction time of 1 min. According to the TIC of Py-GC/MS, when the retention time was less than or equal to 6.5