Time-dependent categorization of volatile aroma compound formation in stewed Chinese spicy beef using electron nose profile coupled with thermal desorption GC–MS detection

Hui Gong,Zhen Yang,Meng Liu,Zhijia Shi,Jiapeng Li,Wenhua Chen,Xiaoling Qiao

China Meat Research Centre,China Meat Processing and Engineering Center,No.70 Yangqiao,Fengtai Distinct,Beijing 100068,China

Abstract

Keywords:Beef flavor;Gas chromatography–mass spectrometry;Electronic nose;Thermal desorption system;Solid-phase microextraction

1.Introduction

The various methods of cooking beef include barbecuing,broiling,griddling,roasting,grilling,frying,steaming,stewing,and others[1,2].Methods such as barbecuing,grilling,and frying deplete nutrients and even produce harmful substances[3,4].It is well known that harmful compounds,including heterocyclic amines(HCA)[5,6]and polycyclic aromatic hydrocarbons(PAH)[7,8],may form in barbecued meat.Mr Oz[9]studied the HCA and PAH contents of beef cooked on wire and stone barbecues at different cooking levels.The results showed that the total HCA and PAH contents of the samples cooked on a barbecue increased at the well-done and very-well-done levels.

The best-quality beef on the market is often grilled or fried.Oil is most commonly used for frying and is absorbed by the beef.Some studies[10]focused on the effects of frying on the fatty acid(FA)profiles of animal products.In a recent study,the FA profile of king salmon meat was modified by deep-fat frying but remained unchanged after pan frying without oil[11].Therefore,the use of low-temperature cooking without oil is more conducive to the retention of nutrients and prevention of harmful substance formation.Stewing is one of the low-temperature preparation methods performed without oil.Spiced beef stewed at a low temperature without oil is considered a healthy beef product.

Spiced beef cooked in the traditional Chinese way is very popular with consumers because of its good flavor.Some industrial beef products are also gaining worldwide recognition.The quantity of industrially produced spiced beef is small because there is a problem with the heat treatment process.The key con-trol points in heat treatment are the stewing temperature and time.Nonetheless,temperature control is not very strict:a range of 80–120℃ is acceptable.Thus,cooking time has become the most important factor in determining the quality of spiced beef,especially the flavor.The formation process of spiced beef flavor is very complex and includes flavor formation of the beef itself during simmering,flavor migration of spices at different reaction levels.Most previous studies focused on the flavor of the beef itself[10],certain spices[12],or methods of detecting flavor[13].So far, there have been few reports about methods of monitoring the flavor change during cooking time in spiced beef.This study analyzed changes in the flavor of spiced beef after different processing times.Firstly,the main flavor compound species in spiced beef were detected,then the key flavor components were determined,and lastly,the effects of cooking time on the formation and changes in the key flavor substances in spiced beef were studied.The optimum cooking time for spiced beef was determined as a theoretical basis for industrialized production.

In this paper,the two principal methods used to analyze the flavor components of spiced beef were the electronic nose and gas chromatography–mass spectrometry(GC–MS).The electronic nose was used to classify flavor substances and to study changes in the different kinds of flavor substances after various cooking times.The electronic nose has been developed rapidly in recent years and is widely used to analyze volatile profile characteristics of foods,cosmetics,essential oils,and other consumer products[14].The most prominent features of the electronic nose are rapid testing and simple sample preparation[15].GC–MS was used to isolate and analyze the main flavor substances and to provide assistance with gas chromatography–olfactometry(GC-O)[16]and the linear retention index(LRI)[17].GC–MS has become the most basic method of volatile flavor compound determination in recent years[12]and is the basis of LRI,GC-O,and some other methods[16].LRI can help to identify components by comparing experimentally found retention indices with known values.The retention index of a certain chemical compound is the retention time normalized to those of adjacently elutingn-alkanes.In food flavor analysis,GC-O has proven to be a valuable technique to characterize the odor-active compounds[18],impact compounds,and odor of a food sample[19].

In this study,a thermal desorption system(TDS)and a solidphase micro-extraction(SPME)were chosen as the extraction methods before the evaluation,and then TDS was selected for subsequent trials.SPME is a traditional method for the extraction of food flavor substances,and it is now widely used[20].Thermal desorption was first used in the field of environmental science as an extraction method[21],but it has been gradually applied in the field of food flavor analysis because of its high extraction efficiency[22].TDS is a highly optimized system for thermal desorption of volatile organic compounds and semi-volatile organic compounds followed by gas chromatography analysis[17].Comparison of the extraction results determined TDS to be the optimal extraction method.Combined with GC–MS,LRI,and GC-O,the volatile substances and the characteristic flavor compounds in Chinese spiced beef were identified.A method of TDS/GC–MS for detecting flavor substances was established, and the flavor change during the process of beef production was also monitored,providing a new perspective and method for the quality control of Chinese spiced beef.

2.Materials and methods

2.1.Spiced beef processing

Frozen raw meat was trimmed and divided into 10-cm squares when the center reached?2℃.Water was heated to 80℃,salt was added,and traditional Chinese spices were used as flavor adjuncts in 10min of boiling.The spice mixture included Chinese prickly ash,anize,dried tangerine peel,cinnamon,clove,amomum,tsaoko,and fennel.Then,raw meat was added and stewed for 4h.A sample of beef and spiced water was removed once every hour.Meat samples were cut into small pieces 0.3cm square,vacuum packaged,and stored at?80℃ for volatile compound analyses.The spiced water samples were frozen at?20℃ for preparation.

2.2.Volatile compound profiling by electronic nose

An APEN3 electronic nose system(Airsense,Schwerin,Germany)composed of a measuring chamber with 10 sensors,was used for acquisition and analysis of data generated by the PEN3.The PEN3 was controlled by Win Muster Airsense Analytics Inc.proprietary software and was connected to an automatic sampling apparatus(HSS32)that had a carousel of 32 sites for loading samples[23].Two grams of sample were added to a 10-mL electronic nose sample vial.Then,the headspace inside was equilibrated for 1h to remove humidity from the surrounding environment. Preliminary experiments showed that after 30min of equilibration,the headspace reached a steady state.The sample was then warmed to a controlled temperature(50℃)for 5min before the automatic sample program was run.The headspace gas was pumped over the sensors of the electronic nose.During the measurement process,three different phases can be distinguished:concentration(10s),measurement(90s)and stand-by(250s).The electro valves,controlled by a computer program,guided the air though different circuits depend on the measurement phase.Irrespective of phase,air flow was always kept constant though the measurement chamber.During the measurement phase,the bomb pushed the volatiles though a closed loop that included the measurement and concentration chambers.No air entered or exited the loop.The measurement phase lasted for 90s,which was sufficient for the sensors to stabilize.When a measurement was completed,a stand-by phase was activated(250s).The main purpose of the stand-by phase was to clean the circuit and return the sensors to their baseline.During this phase,clean air entered the circuit,crossed first the measurement chamber then the empty concentration chamber,and pushed the remaining volatiles out of the circuit.Each sample was repeated in parallel seven times.Table 1 lists all 10 of the sensors used and their main applications.During the mea-surement phase,a computer recorded the changes in resistance experienced by the sensors.

Table 1Specificity and detection sensitivity of Electronic nose sensors.a

2.3.Headspace volatile sampling

There were two headspace sampling procedures in the experiment:the TDS and SPME extraction methods.For TDS,20g(20mL sample of spiced water)of specimens were placed in a 100-mL glass bottle.One end of the bottle was inserted into a glass tube with high-purity nitrogen gas,and the other was inserted into the Tenax TA extraction tube(Gerstel,Mullheiman derRuhr,Germany).The apparatus was placed in a water bath at 50±0.2℃for10min,then extracted at 50±0.2℃for30min.When the sample was in the water bath,nitrogen was ventilated at a 100mL/min flow rate to release the flavor substances in full.Fig.1 shows the TDS extraction process.The extracted analytes desorbed in the TDS and were then determined by GC–MS.Each sample was repeated in parallel three times.

Fig.1.Schematic diagram of the TDS extraction process.

For SPME,5g(5-mL sample of spice water)of specimens were placed in a 40-mL SPME glass bottle,then conditioned in a thermostatic bath at 50±0.2℃ for 10min.After this sample/headspace equilibration period,the septum covering the vial headspace was pierced with the needle containing the fiber and retracted.Then,the fiber was exposed to the headspace for 30min at 50±0.2℃.SPME Fiber,75μm Carboxen/Polydimethylsiloxane(CAR/PDMS),and a manual holder were purchased from Supelco Park(USA).The extracted analytes desorbed in the injection port of the GC–MS.Each sample was repeated in parallel three times

2.4.Analysis of volatiles by GC–MS

Analyses were performed with a Thermo Fisher Trace1310 gas chromatograph(Thermo Fisher Scientific,Massachusetts,USA)coupled to a quadrupole mass spectrometer TSQ8000(Thermo Fisher Scientific,Massachusetts,USA).

A TDS-3 was installed on the top of a Thermo Fisher Trace1310 GC equipped with a cooled injector system(CIS)(Gerstel,Mullheiman der Ruhr,Germany).Splitless thermal desorption was performed by programming the TDS from 40 to 230 ℃ at a rate of 60 ℃/min.Analytes were cryo-focused in the CIS at?100℃ with liquid nitrogen,and CIS was ramped from ?100 to 230℃ at 10℃/s,with a split ratio of 30:1.

The GC system was equipped with a fused silica capillary column(TG-WAXMS,30m×0.25mm i.d.,film thickness 0.25μm)(Thermo Fisher Scientific,Massachusetts,USA).The oven temperature was programmed as follows:initial temperature 40℃,hold for 3min,then increase to 200℃ at a rate of 5℃/min,then to 230℃ at 10℃/min with 3min of final isotherm.The transfer line temperature was kept at 260℃,and the ion source temperature was 280℃.The detector operated in scan mode from 40 to 600amu with a scanning velocity of 0.2 scan/s and using solvent delay of 3min.

2.5.Statistical analysis

When the electronic nose measurement was completed,the acquired data were properly stored for later use[23].The set of signals detected by all sensors during measurement of a sample constitutes a pattern.The pattern of multiple measurements dealing with the same problem are stored in a pattern file and act as the training set.After the data were recorded,principal component analysis(PCA),linear discriminant analysis(LDA),and loading analysis were applied[24].

After the GC–MS detection method,identification of all volatile compounds was carried out by comparing GC retention time with those of standard compounds and with Nist 98 and Wiley 275 mass spectral libraries.In order to identify unknown spectra,the LRI were also calculated for each peak using as a reference the series of hydrocarbons C6–C25 and compared with literature LRI values.The samples were also analyzed by three experienced assessors by sniffing the port of a sniffer 9000 sys-tem(Brechbühler,Schlieren,Switzerland)until no odors were detected.Aroma compounds were identified by comparison of the assessors’descriptors with those described in Fenaroli’s handbook of flavor ingredients[25].

3.Results and discussion

3.1.Electronic nose determination results

An electronic nose with 10 sensors was used in this research.Each sensor sensitive compound type is shown in Table1.When the response value of each sensor was more than 1,it was meaningful for the study.Therefore,as shown in Fig.2,the response of sensors R2,R6,R7,R8,and R9 were of significance for this experiment.The results showed that these five sensors could be divided into two categories:R6 and R8 are sensitive to aldehydes,ketones,and methyl compounds;R2,R7,and R9 are sensitive to sulfur and nitrogen compounds.R2 was sensitive for nitrogen compounds;the response values were in the wave form.R7 and R9 were sensitive for sulfur compounds,in which the trend for the change in response values was consistent with those of R6 and R8.They were all unified after 1h following a sharp decline in the highest and a slow increase to a stable value tendency.After 1h of cooking,sulfur compounds,methyl aldehydes,and ketone compounds underwent a material transformation process followed by 3h of cooking during which the species were invariant,and these compounds content increased gradually to stable levels.

Fig.2.Flavor response of electron nose sensors for Chinese spiced beef during stewing(up to 4h).R1 through R10 denote the 10 individual sensors.

Loading analysis was performed,and a plot of the loading factors associated with spiced beef is shown in Fig.3c.Loading analysis is helpful to identify the sensors responsible for discrimination in the current pattern file.The sensor might be switched off for analysis(the response signal was not used)if it has a rather small influence on the identification process[23].The plot shows the relative importance of the sensors in the array.The loading factors associated with the first and second principal components of each sensor are represented.Fig.3c shows that sensors R2,R6,R7,R8,and R9 have a higher influence on the current pattern file,which is identical to the results shown in Fig.2.

Fig.3.The PCA(a),LDA(b),and loading analysis(c)of Chinese spiced beef during stewing(up to 4h).

To investigate changes in the cooking process for Chinese spiced beef by electronic nose,PCA and LDA analyses were applied.PCA and LDA analysis results are shown in Fig.3a and b.These figures show the analysis results on a two-dimensional plane,principal component 1(PC1)and principal component 2(PC2)in Fig.3a,and first and second linear discriminant LDA function 1(LD1)and function 2(LD2)in Fig.3b.PCA is a statistical technique for the reduction of input data dimension and is largely used for feature extraction. It captures the relevantinformation in a set of input data and provides a lower dimension[23].The processed data showed a shift of the different cooking process state that coinciding with the classification by the trained profile panel.The first principal component,PC1,was responsible for 74.65% of the total variation,while 24.82% of the total variance was explained by PC2.Discrimination power(DP)was the parameter used to show the extent of discrimination between groups.A higher DP indicated a higher degree of distinction and a greater difference.The DP of the electronic nose for the 1h sample vs.the other three time points was high(to 2h of 0.81,3h of 0.92,4h of 0.74),the DP for 2h to 3h was 0.77,for 2h to 4h it was 0.66,and for 3h to 4h it was 0.60.These results showed that the power of the electronic nose to distinguish the 1h sample from the other time points was the highest, and among the other three samples, the flavor difference was small,as indicated by lower DP values.It also explained in Fig.2 that the R2 sensor response to the 2 h,3h,and 4h samples fluctuated widely because this sensor has relatively high sensitivity and easily affected by the environment.Greater attention should be paid to the species detected by the R6,7,8,and 9 sensors,which indicate changes in aldehyde,ketone,methyl,and sulfur compounds.

LDA was applied to the same dataset,and it discriminated clearly between the various clusters representing different cooking time states.All spiced beef cooking times were perfectly classified(Fig.3b).In this plot,approximately 91.70% of the total variance of the data is displayed.LD1 and LD2 accounted for 79.77% and 11.93% of the variance,respectively.LDA was the analytical method used to study the flavor change rate.A smaller distance between data clusters indicated a slower rate of change,and a larger distance indicated a faster rate.As can be seen from the long distance in the LD1 from 1h to 2h,the change rate was high;between 2h and 3h the LD2 had a long distance,meaning the rate of change was slowed down;and 3h and 4 h were not separated completely, thus there was no obvious change.Integrated PCA and LDA results found that the spiced beef flavor changed dramatically between 1 and 2 h;after 3h the flavor tended to be stable;and the changes in the vigorous mass were mainly concentrated on aldehydes,ketones,and methyl and sulfur compounds.The reason may be that dramatic changes in the flavor occurred as the Maillard reaction caused changes in volatile substances at first,and accumulated when they get stable.Both PCA and LDA analysis showed that the samples were completely separated from the first hour to the other three time points.In the PCA,it can be seen that the degree of differentiation of samples is more likely to analyze the rate of change in LDA.

3.2.Headspace extraction method chosen

Because of the difference in packing length and the adsorption efficiency,with the same extraction conditions(50℃ for 30min),the results of the TDS extraction method were superior to those of SPME.There results of two extraction methods of TIC chromatography are shown in Fig.4.The highest response of the TDS method was 2-fold higher than that in SPME under the same sample detection conditions for the same compound.The GC–MS analysis results of the SPME method yielded 65 kinds of volatile substances,while TDS revealed 82 kinds of volatile substances.The TDS method was better than the SPME method regardless of the extract species or extraction capacity.Therefore,the TDS method was selected as the extraction method in this study for subsequent research.

Fig.4.TIC chromatography of the TDS and SPME extraction of Chinese spiced beef.

3.3.Chinese spiced beef–main flavor composition analysis

As shown from the results of the TDS method,82 kinds of volatile compounds were detected by comparative mass spectra using the NIST 08 library search system,as shown in Table 2.Twenty-six kinds of volatile compounds were identified by comparison of retention indexes with literature data.Thirty-six kinds of volatile compounds had been reported in the literature as having flavor.Thirty-five kinds of volatile compounds were determined by olfactometry.The following chemical structures were identified:aldehyde(10),hydrocarbon(18),ester(5),alcohol(12),phenol(3),ether(4),heterocyclic(7),terpene(14),ketone(7),and acid(2).The relative concentrations are shown in Fig.5;ethers and aldehydes were the most abundant compounds.Under comprehensive inspection,3-methyl-butanal,pentanal,hexanal,ρ-xylene,heptanal,limonene,γ-terpinene,octanal,linalool,4-terpinenol,α-terpineol,naphthalene,and(E)-anethole were the core flavor components.The most represented of the above-mentioned compounds were aldehydes,alcohols,terpenes,and ethers.Alcohols,terpenes and ethers originated from spices,and aldehyde came from beef,including correlated reactions.

Table 2 Volatile compounds in finished spiced beef product as detected by TDS/GC–MS,respective chromatographic linear retention indexes(LRI)and olfactory identification(GC-O).

Table 2(Continued)

Fig.5.Relative concentrations of flavor compounds in Chinese traditional spiced beef.

3.4.Change law during stewing of the volatile flavor compounds

The changes that occurred in volatile flavor compounds of beef and spiced soup during 4h of stewing are shown in Fig.6a and b,respectively.As shown in Fig.5,ether showed the greatest response of the volatile flavor compounds in Chinese spiced beef,followed by aldehydes and ketones.The response of the remaining compounds(hydrocarbons,alcohols,terpene,phenols,heterocyclic,esters,and acids)was low and slightly changed.GC–MS results for the volatile flavor compounds of Chinese spiced beef were coincident with the sensor response results of the electric nose.In the spiced soup as shown in Fig.6b,aldehyde was the highest responder of the volatile flavor compounds,followed by alcohol,phenol,and ether.

Fig.6.Change in volatile flavor compounds in 4h in stewed beef(a)and spiced soup(b).

As seen in Table 2,only four ethers were detected in Chinese spiced beef.However,their content was very high:as Fig.5 shows,a stable relative content of 58.5% was present in the 4h sample.As shown in Fig.6a,the value in the 1h spiced beef sample was the highest and was followed by a radical decrease and then a gradual increase to stability.In the 1h spiced soup sample,ether was also the highest.The reason for the change was related to the source of the ethers. The majority of ether was derived from spices and were very easily volatilized.In 1h they could be fully released;thus,both beef and spiced soup were high in ether at the 1h time point.But when emitted into the air,the response in spiced soup was lower and then stable.On the other hand,the ether in beef turned large surface adsorption into internal absorption and then increased to maximum stability in the 4h sample.

Another important compound was aldehydes.There were 10 aldehydes detected in Chinese spiced beef.As Fig.5 shows,the relative amount was 10.6% in the 4h sample.Most of the straight-chain aldehydes were derived from the oxidation of unsaturated FAs in the samples,such as hexanal,octenal,pentanal,or heptanaletc in previous reports[26].For exam-ple,hexanal was derived from ω-6 unsaturated FAs in terrestrial animal tissues, octenal was derived from linoleic acid, and undecanal was derived from oleic acid[27].The aroma note of hexanal was described as intense grass-like,pentanal was like almonds,heptanal was fatty,and octanal was like orange peel.In the first 2h,aldehydes in beef were increased and then stabilized in the following 2h.Yet in spiced soup,aldehydes showed a gradual upward trend.This was probably due to aldehydes mainly being rooted in beef and the related Maillard reaction.For spiced soup,aldehydes from the spices were released in 1h,and sustained release occurred in the following hours due to the beef and related Maillard reaction produced.As a result,aldehydes in the spiced soup were the amount of the spice itself and the beef with related reactions.

There were seven ketones detected in Chinese spiced beef.As Fig.5 shows,the relative amounts was 7.3% in the 4h sample.Aliphatic ketones,especially methyl ketones,arise from autoxidation or beta oxidation of fatty acids and are considered to be among the precursors that contribute to the fatty aromas associated with cooked meat[28].2-butanone was derived from oxidized tallow and was implicated in the buttery aroma note of cooked meat.Ketone content increased gradually in the process of cooking for 4h,and the content was less in soup.Mainly because ketone was the product of oxidation of tallow,and gradually accumulated in the beef.

There were 12 alcohols detected in Chinese spiced beef.As Fig.5 shows,the relative amount was 6.5% in the 4h sample.Alcohols were derived mainly from the oxidative decomposition of fat.2-ethyl-1-hexanol was the predominant alcohol[29].The flavor note of straight-chain primary alcohols was like that of a greenish,woody,and fatty- floral[30].Some other alcohols may be from the spiced soup,as well as some flavor-like camphoraceous,aniseed,or flower.The alcohol content of spiced beef was not high,and there was a rise in the declining trend;the alcohol content of the soup was higher.This showed that most of the alcohol from the spiced beef was from fat oxidation and a certain degree of evaporation, and the majority of alcohols in the spiced soup were from the spice itself.

There were 18 hydrocarbons detected in Chinese spiced beef.As Fig.5shows,the relative amount was 7.2% in the 4 h sample.Hydrocarbons did not contribute much to the flavor of spiced beef;there was no detailed analysis.There were 14 terpenes detected in Chinese spiced beef.As Fig.5 shows,the relative amount was 5.7% in the 4h sample.Besides the meat aroma,terpenes that were mainly from the spiced soup gave Chinese spiced beef a unique flavor.Other types of compounds were of low content and did not have a large contribution to the flavor,so they are not mentioned here.

In the above analysis,the data are only based on the GC–MS experimental results.For the common flavor compounds,sulfur or nitrogen in the meat detection response is too low and is not shown in this analysis.Further research could use a flame photometric detector or other analytical technology.In this study,during the beef cooking process,the contents of flavor substances in real-time change–control of especially two kinds of volatile aldehyde and ether can be used as one of the key points in quality control.

The electronic nose method first divided Chinese spiced beef flavor substances into large classes through different types of sensors and distinguished the 4h samples.Combined with the method of TDS-GC–MS,the species of flavor substances were divided more detail, and the change trend was consistent with the results of the electronic nose.The electronic nose method was not only a means of flavor detection but could also a compensate for the weakness of the GC–MS method,such as time consuming,high cost.So,it can be used as an effective and fast quality control method for the processing of Chinese spiced beef.

4.Conclusion

This was a study of volatile compound detection by GC–MS with two different headspace sampling methods(TDS and SPME)and electronic nose to quickly identify the different cooking process responses of Chinese spiced beef.Results indicated that Chinese spiced beef had a variable volatile composition. Based on the instrumental analysis,82 kinds of volatile compounds in Chinese traditional spiced beef were identified,and differences in the composition of volatile components from four different sampling time points were observed,indicating that the flavor quality of spiced beef samples clearly varies with cooking time. It was also found that the electronic nose can identify spiced beef of different cooking time.Volatile profiling by GC–MS with TDS and the responses from the electronic nose in the combination with multivariate statistical analysis should be promising tools for control of the cooking process of spiced beef.

This research did not receive any specific grant from funding agencies in the public,commercial,or not-for-profit sectors.

Acknowledgement

This work was part of the project“Research and Development of Nutrition and Health Processing for Halal Beef and Muttons”,and was financially supported by Ningxia Hui Autonomous Region Technology R&D Support Program as well as the “13th Five-Year Plan”(No.2016YFD0400703)of National Key Research and Development Program of China.