analytical approaches for the assessment of free fatty acids in oils and fats

AnalyticalMethods

CRITICAL REVIEW

Publ

ishe

d on

15

May

201

4. D

ownl

oade

d on

16/

10/2

014

13:3

7:13

.

View Article OnlineView Journal | View Issue

Analytical approa

SPoCJfrsiMiCC

completed his postdoctoral studieDr Sherazi was awarded the Best UHEC, Pakistan. Dr Sherazi is anreviewed papers in national and in

aNational Centre of Excellence in Analy

Jamshoro-76080, Pakistan. E-mail: tufail.sh

Tel: +92-22-9213429bDepartment of Pharmacy, Shaheed Mohta

Larkana-77150, PakistancDr M. A. Kazi Institute of Chemistry, Unive

Cite this: Anal. Methods, 2014, 6, 4956

Received 14th February 2014Accepted 15th May 2014

DOI: 10.1039/c4ay00344f

www.rsc.org/methods

4956 | Anal. Methods, 2014, 6, 4956–4

ches for the assessment of freefatty acids in oils and fats

S. A. Mahesar,a S. T. H. Sherazi,*a Abdul Rauf Khaskheli,b Aftab A. Kandhroc

and Siraj uddina

Free fatty acids (FFA) are produced by the hydrolysis of oils and fats. The level of FFA depends on time,

temperature and moisture content because the oils and fats are exposed to various environments such

as storage, processing, heating or frying. Since FFA are less stable than neutral oil, they are more prone

to oxidation and to turning rancid. Thus, FFA is a key feature linked with the quality and commercial

value of oils and fats. The American Oil Chemists' Society (AOCS), the Association of Official Analytical

Chemists (AOAC) and the European Commission (EC) Regulations have established almost identical

standard methods for the assessment of FFA. These methods are based on titration, where oils or fats

need to be dissolved in hot neutralized ethanol or ethanol/diethyl ether using phenolphthalein as an end

point indicator. Titrimetric procedures are, however, laborious and need large amounts of chemicals and

solvents. The cost of chemicals and environmental issues further limit these procedures. In addition,

accurate detection of end points, especially for highly colored crude oil using a colorimetric indicator, is

a difficult task. Despite these disadvantages, the titration method is unfortunately still being used in most

of the edible-oil industries for the determination of FFA. Because of a lack of any comprehensive review

on this very important topic, we have made an attempt to present a review in order to discuss the

various methods available with special emphasis on the instrumental methods because of their high

sensitivity, accuracy and rapidity.

yed Tufail Hussain Sherazi is arofessor at the National Centref Excellence in Analyticalhemistry, University of Sindh,amshoro, Pakistan. His Ph.D. isrom the same institution. Aereceiving his Ph.D., Dr Sherazitarted his career in the edible oilndustry. He was the Generalanager when he le the oil

ndustry and joined the Nationalentre of Excellence in Analyticalhemistry in 2004. In 2006, hes at McGill University, Canada.niversity Teacher Award 2011 byauthor of more than 90 peer-ternational journals.

tical Chemistry, University of Sindh,

[email protected]; Fax: +92-22-9213431;

rma Benazir Bhutto Medical University,

rsity of Sindh, Jamshoro-76080, Pakistan

963

1. Introduction

Vegetable oils or fats are extracted from oilseeds in a crudeform, and are converted to edible oils using various reningstages such as degumming, neutralization, dewaxing, bleachingand deodorization. During processing, a number of parameterssuch as moisture, peroxide value, free fatty acids (FFA), sapon-ication value and iodine value are monitored. Among theseparameters FFA content is considered as a crucial factor and islinked with the quality as well as the economic value of theedible oils.1 Generally, FFA are the hydrolysis products of oiland fat oxidation during long-term storage or processing atelevated temperatures during heating or frying. Initially, theAmerican Oil Chemists' Society (AOCS),2 the Association ofOfficial Analytical Chemists (AOAC)3 and the EuropeanCommission Regulation (EC)4 have recommended standardmethods for the determination of FFA. These methodsare similar to one another and are based on the titration of oil(3.5–56.4 g). Oils or fats are dissolved in hot neutralized 95%ethanol (50–100 mL) or ethanol/diethyl ether (50–150 mL),against a strong base using phenolphthalein as an indicator.Although these titrimetric methods are very simple, they arealso extremely laborious and need large amounts of chemicalsthat are expensive and associated with environmental issues.Furthermore, it is quite challenging to determine the accurate

This journal is © The Royal Society of Chemistry 2014

http://dx.doi.org/10.1039/c4ay00344f

http://pubs.rsc.org/en/journals/journal/AY

http://pubs.rsc.org/en/journals/journal/AY?issueid=AY006014

Fig. 1 Block diagram of the methods developed for the determinationof FFA content.

Critical Review Analytical Methods

Publ

ishe

d on

15

May

201

4. D

ownl

oade

d on

16/

10/2

014

13:3

7:13

. View Article Online

end point during titration for dark crude oils in the presence ofa colorimetric indicator.

To overcome such issues, a number of instrumentalmethods have been developed as potential alternatives to clas-sical methods for more accurate and sensitive measurement ofFFA content in edible oils and fats. In general, these methodscan be categorized into three major groups: colorimetricmethods, electrochemical methods, and spectroscopicmethods. These three groups of methods are discussed in detailin Sections 2.1, 2.2, and 2.3, respectively. The electrochemicaland spectroscopic methods in general involve a variety ofsophisticated instruments with diverse approaches thatcompete with each other in terms of speed of analysis,amenability to automation and reduction of harmful solvents.Some instrumental methods are more accurate and sensitivethan classical titrimetric methods.

Several such methods have been reported during the lastsixty years. In the 1950s, spectrophotometric techniques werewell addressed, and much was also published about FFA usingcolorimetric methods. Aer that, electrochemical techniquesbecame more popular. In the last two decades, however,attention has focused on the use of infrared (IR) spectroscopicmethods for the direct or indirect determination of FFA. Thesemethods were claimed to be environmentally friendly sincechemicals are totally avoided, or are required in negligibleamounts as are the samples. Fig. 1 shows the block diagram ofthe methods developed for the determination of FFA contentduring the last few decades. To the best of our knowledge thereis no comprehensive review available on FFA analysis. There-fore, the present review will explore the many methods used foranalyzing FFA that have been developed in last few decades.

2. Discussion2.1. Colorimetric methods

Various colorimetric methods have been reported by severalresearchers.

2.1.1. Copper soap methods. In 1956, Ayers5 made his rsteffort to determine FFA concentrations by a colorimetric


method. According to his procedure, an aqueous solution ofcopper or cobalt nitrate was used to convert FFA into water-insoluble soaps. Colored soap precipitate was extracted withchloroform and quantication was carried out bymeasuring theabsorbance at 675 and 527 nm for copper and cobalt, respec-tively. Although the method was not proposed for FFA analysis,it provided the basis to develop colorimetric methods for FFAanalysis in edible oils. In the 1960s, Baker6 proposed animproved copper salt-based method, which was designedparticularly for FFA analysis in edible oils. The method involvedthe dissolution of oil in benzene followed by mixing withaqueous cupric acetate solution. The mixture was centrifugedand quantication of FFA was determined by measuring theabsorption of the benzene layer at 640 nm. The results from thismethod showed good correlation with those obtained by titri-metric methods. The only limitation of this method was thedissolution of some copper salts of saturated fatty acids inbenzene. In view of the inconsistency between the workingwavelengths used in these methods, 640 vs. 675 nm,5,6 Bainset al.7 examined the spectra and the effect of the presence ofcopper on the color of oil during the measurement. Heconcluded that 670 nm was the optimum wavelength for usingcopper soaps to analyze non-colored or colored oils. Twelveyears later, Lowry and Tinsley8 reported on a substantial work tooptimize the method in terms of sensitivity and reproducibility.The conclusion of his work was the stabilization of the coppersoap of saturated fatty acids in benzene and improvement in thesensitivity by controlling the pH using pyridine. Because of thepractical and health concerns of benzene, various authors usedalternatives such as toluene,9 isooctane10 and cyclohexane.11

Gyorik et al.12 reported a new analytical tool for the determi-nation of total FFA content in thermally treated cooking oil. Themethod was based on the colorimetric copper-soap methodcombined with the concept of the optothermal window(with 632.8 nm He–Ne laser used as a radiation source). Theresults obtained by this method were compared to thoseacquired by conventional spectroscopy; the correlation betweenthe two methods was high for FFA concentrations exceeding2 mmol mL�1.

2.1.2. Automated ow injection analysis systems. The rstautomated system for FFA analysis in edible oils using thecopper-soap method was introduced in the 1980s.9 The auto-mation was accomplished using a ow injection analysis (FIA)system, which was able to analyze up to 20 samples per h. Incontrast to the previously published methods, based on organicsolvents, the quantication was done in the aqueous phase.However, the FIA system did not receive much attention,because of complications and stability problems. In 1987,improved hardware for the FIA system using a single extractionstep was reported.13 This new system offered enhanced stabilityand a much shorter analysis time (130 samples per h). The onlyshortcoming of the system was poor sensitivity. Puchades et al.14

improved the sensitivity of FIA systems by optimizing theoperating conditions such as pH, ow rate, tubing length/diameter, and proportion of oil to solvent. These optimizationsresulted in a better analysis speed (75 samples per h) and alower limit of detection (LOD;�0.01% FFA) than those obtained

Anal. Methods, 2014, 6, 4956–4963 | 4957


Analytical Methods Critical Review

Publ

ishe

d on

15

May

201

4. D

ownl

oade

d on

16/

10/2

014

13:3

7:13


with the rst automated system. Despite these improvements,the method did not get much attention because of complica-tions in the analysis and measurements that were incompatiblewith values obtained by the official method.

A continuous liquid–liquid extraction system using a single-channel, ow-reversal injection mode was also proposed for thedirect determination of FFA concentration.15 The resultsprovided by the proposed method were in accordance withthose obtained by a manual procedure and comparable to thestandard acid–base titration method, but the samplingfrequency was lower (12 samples per h).

2.1.3. Phenolphthalein-based methods. In 1989, Linareset al.16 made an attempt to automate, and thus overcome theshortcomings of the official method. He used the same reagentsas were used for the official IUPAC method, whilst quantica-tion was based on monitoring the decrease in the phenol-phthalein (PHP) intensity at 562 nm as a result of the reactionbetween potassium hydroxide (KOH) and FFA in the oil sample.The analytical range of this PHP-based method was 0.15–0.81%FFA and the sample throughput was up to�60 samples per h. In1997, a modied method with a similar analytical range and asample throughput of 30–100 samples per h was reported17

using 3–7mL of inexpensive 1-propanol as a diluent and carrier,this was much more economical when compared to the hugeamount of solvents used in the official method.

A direct measurement of FFA in virgin olive oil withoutdilution was reported using a similar system.18 The system wasequipped with an optical ber beam connected to the spectro-photometer set at 562 nm. The system had the same sensitivityas the previous methods with a sample throughput of 12–60samples per h. Aerwards, an indirect titrimetric method wasdeveloped based on the complete extraction of acids from oilinto a reagent consisting of 0.05 mol dm�3 triethanolamine in amixture of 50% H2O + 50% 2-propanol, and indirect titration ofconjugate acids (those in the form BH+) against aqueous alkaliin the presence of PHP indicator.19 The benets of this methodincluded the lack of toxic solvents, analysis at room tempera-ture and the absence of any need for preliminary neutralizationof the acid admixtures.

Saad et al.20 reported single-line and two-line manifold owinjection (FI) methods using PHP and bromothymol blue (BTB)as indicators. The method was based on monitoring changes inthe absorbance of the indicators at 562 nm for PHP and 627 nmfor BTB. Various FI parameters were optimized such as carrierand reagent concentrations, length of the reaction coil, ow-rate, injected volume, and the size of the mixing chamber. Thesingle-line manifold with PHP as the indicator was recom-mended for the determination of samples with an acidity higherthan 0.4%, while for lower acidities (<0.4%) a two-line manifoldwith BTB was recommended. The method was linear over therange 0.4–10.0% FFA for single-line manifold and 0.11–0.50%FFA for the two-line manifold. Sample throughputs were 35–74and 21–46 samples per h for single-line and two-line manifolds,respectively.

A non-aqueous single-line manifold system was built bymodication of an high-performance liquid chromatography(HPLC) for FIA.21 Oil samples were injected without pre-

4958 | Anal. Methods, 2014, 6, 4956–4963

treatment into an n-propanol solution containing KOH andPHP. The linear concentration range was calculated as0.09–1.50% and 0.07–1.40% FFA for corn and sunower oils,respectively. The benets of this method were its simplicity andhigh sample throughput, and its results were comparable withthose obtained by the AOCS method.

Makahleh and Saad22 reported a single-line FIA methodattached to a capacitively coupled contactless conductivitydetector. Methanol mixed with 1.5 mM sodium acetate (pH 8) ina ratio of 80 : 20 (v/v) was used as a carrier stream at a ow rateof 1.0 mL min�1. This method offers a high sampling rate of40–60 samples per h and is an inexpensive automated systemthat needs minimum human intervention over long periods oftime. Good agreement was found between this method and thestandard non-aqueous titrimetry method. In 2012, a visual andgreen titrimetric method was described using PHP as an indi-cator23 and for which the sample was dissolved in a water–ethanol mixture (1 : 1 v/v) and titrated with a 0.02 mol L�1

aqueous NaOH solution. The results obtained by this approachwere comparable to those obtained with the AOCS titrationmethod.

2.1.4. Miscellaneous spectrophotometric methods. A newapproach for the determination of FFA concentrations in a non-aqueous medium without titration was reported in the litera-ture.24,25 The method utilized phenol red as a fatty acid indi-cator, which was solubilized in reverse micelles formed bysodium bis(2-ethylhexyl) sulfosuccinate in isooctane. Quanti-cation was achieved by monitoring the changes in absorbanceat 560 nm (disappearance of the red color). The method waseconomical (small amounts of samples and solvents), but wasfound to be lacking in accuracy. Therefore, no any additionalefforts were made for its improvement. Rejeb and Gargouri26

reported a newmethod for the determination of olive oil acidity.The method was based on the measurement of polyunsaturatedFFA reacting with lipoxygenase. The concentrations of enzy-matically produced hydroperoxy-fatty acids were easily deter-mined by UV spectrophotometry at 234 nm. The proposedmethod was successfully applied for the determination ofacidity in olive oil samples and good agreement was obtainedbetween the results obtained with this enzymatic method andthose obtained using classical titration methods. This enzy-matic method was accurate, simple, sensitive, inexpensive, andreliable for determination of acidity of olive oil with a highsample throughput. In 2012, Fedosov et al.27 developed a novelmicrotitration method for determining the concentration ofFFA in small samples of oil (5–150 mg). The method was basedon the optical signal of pyranine (aqueous pK 7.3), the absor-bance and uorescence of which are sensitive to pH. All reac-tants were dissolved in a medium with universal solubility,which allowed accurate optical measurements on conventionalequipment. FFA standards stipulated the selection of theneutralization point of titration curves but comparison ofuorescence and absorbance of pyranine pointed to the some-what better “signal to noise” properties of the uorescent signalwhen working with heavily pigmented oil samples. Blindexamination of different experimental mixtures (FFA ¼ 0.15–40%) revealed a good agreement between the results obtained




Publ

ishe

d on

15

May

201

4. D

ownl

oade

d on

16/

10/2

014

13:3

7:13


using the pyranine method and those obtained using the offi-cial method. Furthermore, it was reported that the small size ofoptical cells could allow further minimization of the pyranineassay if adapting it to an automated procedure.

2.2. Electrochemical methods

2.2.1. Voltammetric techniques. Voltammetry is animportant class of electrochemical methods widely used inanalytical chemistry and various industrial processes. In vol-tammetry, information about an analyte is obtained bymeasuring the current as the potential is varied.28 It iscommonly used for compounds that are readily oxidized/reduced to provide highly selective and sensitive quantitativemethods. Electrochemically, FFA are weakly active and can bereduced or oxidized. The voltammetric behavior of quinone inthe presence of acids in buffered protic solvents was rstreported by Takamura and Hayakawa.29 It was reported that thereduced form of quinone (dissolved in ethanol) gave a verystrong voltammetric signal in the presence of acids. For analysisof FFA in edible oils, the behavior of quinone has been utilizedby various authors30–34 to develop HPLC and FIA methods. TheFIA system coupled with voltammetric detection gave resultswith a high sensitivity and better reproducibility, and themethod had a low solvent consumption and an excellentthroughput of 60 samples per h. Unfortunately this techniquedid not get much attention. Li et al.35 proposed a new methodusing a polypyrrole-modied electrode and a linear potentialsweep voltammetric mode. In this method, the behavior ofnaphthoquinone in the presence of FFA was investigated in anethanol–1,2-dichloroethane (3 : 1) solution containing 0.1 MLiClO4. A linear calibration graph was obtained in the range of5.0 � 10�6 to 6 � 10�3 M for FFA (R ¼ 0.993), with a sensitivityof 2.41 � 10�2 A L mol�1 and a LOD of 1.2 � 10�6 M. Themethod was accurate, superior in sensitivity and capable ofanalyzing 45 samples per h.

2.2.2. Electrical conductivity methods. Recently a newapproach for the determination of FFA in edible oils based onthe change of the electrical conductivity (EC) value of a 0.04 MKOH solution during a KOH–FFAmixed reaction was reported.36

The changes of the EC value were determined for the KOHsolution layers, and the concentration of FFA was determinedfrom a calibration graph. Various parameters were optimized,such as the concentration of KOH in the solution, holding time,the types of edible oils, setting temperature, and the ratio of oilto KOH solution (m/v). The method was validated using stan-dard addition and the AOCS method. The results and analyticalperformance obtained from using the EC procedure was betterthan those obtained by the AOCS method.

2.2.3. pH methods. In general, the low polarity and highviscosity of oils have made it impractical to use pH methodsfor the determination of FFA concentration. Lapshina et al.37

carried out pH measurements by extracting FFA from oil into apolar solvent. In order to facilitate extraction, the authorssuggested the use of the weak base triethanolamine (TEA) inan appropriate solvent mixture (diethyl ether 80%, chloroform19% and water 1%) to convert carboxylic acids into their salts.


The method was workable but showed poor reproducibilitybecause of the instability of the electrode and the basic reagentin the solvent mixture. In the late 1990s, the method wasmodied by replacing TEA with a 1 : 1 (v/v) mixture of iso-propanol and water containing 0.2 M TEA and 0.02 MKNO3.38–41 Rather than extracting FFA, this solvent mixtureeffectively makes an emulsion in the range of pH values thatwere measured. In 1998, Velasco-Arjona and Luque de Castro42

proposed a fully automated method using a robotic station forthe determination of the acid value in olive oil, without titra-tion, based on pH measurements of the emulsion. The resultsobtained showed excellent correlation with those obtainedusing potentiometric titration demonstrated the convenienceof the automated method. This method was able to analyze 15samples per h. Various publications on this approach werereported for acidity measurements in oil seeds.43–45 A newconcept of ow titration was proposed using sequential injec-tion analysis with a diode array spectrophotometric detectorcombined with multivariate curve resolution using alternatingleast squares in the presence of alizarine as an indicator.46 Thetotal waste generated in this method was 3.33 mL (0.41 mL ofsample, 0.25 mL of indicator and 3 mL of carrier). The resultsof this method compared well with those from titrimetricmethods and the frequency of the analysis was 12 samplesper h.

2.3. Spectroscopic methods

2.3.1. Infrared methods2.3.1.1. Direct infrared methods. The possibility of utilizing

Fourier-transform infrared (FTIR) spectroscopy for FFA analysishas been examined by a number of scientists over the last twodecades. In 1991, Lanser et al.47 rst reported a semi-quantita-tive method for the analysis of FFA in soybean oils by computer-assisted FTIR using deconvolution. He collected the spectra ofoils using a transmission cell, and observed the spectralchanges in the carbonyl region in the range of 2000–1600 cm,1

and correlated them to the FFA content obtained by the stan-dard AOCS method. The method was based on developing acalibration plot using oleic acid as a standard in FFA-freesoybean oil. A major issue preventing the method from beingmore quantitative was the spectral variation of the ester bandcaused by the carboxylic acid C]O absorption appearing on theshoulder of the triglyceride ester linkage absorption which wasfound in various oils.

In 1993, a method was developed based on the measurementof the carboxylic acid C]O band at 1711 cm�1 using a trans-mission approach. This method was accurate and covered ananalytical range of 0.2–8% FFA.48 To avoid the matrix effectsnoted earlier47, the ratioed out calibration method was used,and sample spectra were compared against the spectrum of aFFA-free oil of the same type of oil rather than using a back-ground spectrum of air. This approach gave results which werecomparable in precision and accuracy to those of the AOCSreference titration method. By using macroprogramming themethod was completely automated; this approach carries out ananalysis in less than 2 min.

Anal. Methods, 2014, 6, 4956–4963 | 4959



Publ

ishe

d on

15

May

201

4. D

ownl

oade

d on

16/

10/2

014

13:3

7:13


Aer the previously described preliminary studies, a newprocedure was reported for FFA in olive oil based on attenuatedtotal reectance (ATR) FTIR spectroscopic measurements usingpartial least-squares (PLS) regression.49 The method developedwas applicable to samples of different categories of olive oil withan analysis time of 5 min per sample. Che Man et al.50,51

described transmission FTIR spectroscopic methods using 200mm NaCl windows for analysis of the low levels of FFA in palmolein and 100 mm BaF2 windows for analysis of the higherconcentrations of FFA in crude palm oil. The PLS calibrationmodels were developed to determine the correlation of FTIR andchemical methods. The method developed drastically reducedthe analysis time to less than 2 min per sample. Verleyen et al.52

followed up on the work previously reported by Lanser et al.47

with more rigorous research and developed effective calibrationsfor different oils (corn, soybean, sunower, palm, palm kernel,and coconut oils). It was concluded that the calibration modelsdeveloped for the six oils differed signicantly and indicated theneed to develop a calibration model for each specic oil or fat.

Inon et al.53 developed a chemometric-based method usingPLS multivariate calibration and net analyte signal (NAS) pre-processing for the determination of FFA in commercial olive oilsamples of different types and origins. The LOD, sensitivity andselectivity of the methodology developed were evaluated bycomparison to the NAS and an LOD value of 0.072%, with asensitivity of 0.077 per unit concentration, were obtained.

Because of the practical limitations of transmission spec-troscopy for the quantication of relatively high concentrationsof FFA, a simple and direct methodology was developed for thedetermination of such FFA levels (>30%) by single bounce ATR-FTIR spectroscopy in poultry feed lipids,54 deodorizer distillatesand crude oils.55 The accuracies of the methods were checked bycomparison to a conventional AOCS titrimetric procedure. Yuet al.56,57 reported a new approach for the automated determi-nation of FFA in edible oils using a spectral reconstitutiontechnique. The data obtained by this technique were accurateand reproducible. The automated method developed wascapable of analysing �90 samples per h, a rate appropriate forthe throughput required by commercial contract and high-volume process control laboratories.

2.3.1.2. Indirect infrared methods. For fats and oils that mayhave undergone signicant thermal stress or extensive oxida-tion, an indirect FTIR spectroscopic method was developed foranalysis of FFA based on ATR which was used in order to removethe matrix effect.48 KOH/methanol (1%) was used to extract theFFA and convert them to their potassium salts. The carboxylateanion absorbs at 1570 cm�1, well away from the interferingabsorptions of carbonyl commonly present in oxidized oils. Thegreater sensitivity and accuracy of the indirect methodcompared to the direct determination of FFA were shown toovercome the matrix effects, although use of this methodcomplicated the analysis. An automated FIS was developed forthe rapid determination of the FFA content in olive, sunower,and corn oils.58 The method was based on the indirectapproach48 to avoid the matrix effects linked with directmeasurements of the carboxylic acid band at 1711 cm�1. Thisautomated method was able to analyze �40 samples per h.

4960 | Anal. Methods, 2014, 6, 4956–4963

A transmission spectroscopic method using potassiumphthalimide as a weak base to convert the FFA present in oils totheir carboxylate salts without causing oil saponication wasreported.59 The method was automated, and capable of analyzingabout 60 samples per h. In the mid-2000s, a sensitive and rapidmethodology for the quantication of low levels (<0.005%) of FFAin edible oils using sodium hydrogen cyanamide to convert FFAto their carboxylate salts was described.60,61 The feasibility ofemploying a portable variable lter array (VFA) IR spectrometerequipped with a transmission ow cell for FFA analysis wasreported by Li et al.62 The indirect approach was employed toconvert the FFA to their salts as described earlier.60,61 Based onthis measurement, use of a VFA-IR spectrometer was aneconomical means for analysis of FFA levels in crude and renededible oils and the instrumentation was capable of analyzing�20–30 samples per h. Aryee et al.63 determined FFA content insh oils employing the same approach. It was reported that theFTIR method is a exible and viable instrumental alternative tothe standard titrimetric method.

2.3.2. Near-infrared methods. The near-infrared (NIR)methods for the analysis of FFA in oils and fats were reportedbefore the mid-infrared methods. Frankel et al.64 used NIRspectroscopy for the rst time to determine the quality of thesoybeans and its oil stored at different moisture levels. Heobserved that NIR analysis at 2260 nm gave results whichshowed a good correlation coefficient of 0.864 compared withthe results from titratable FFA methods. Furthermore, it wasreported that NIR analysis is most suitable and rapid method toevaluate hydrolytic deterioration in stored soybeans. Thismethodology was used to evaluate the factors affecting thequality of soybeans used for food in domestic and foreignmarkets.

Aer the initial work of Frankel et al.,64 no further work wascarried out on NIR for 10 years. The majority of the methodswere developed for the mid-IR range. In 1997, a direct methodfor the determination of FFA in sh oil was reported.65 PLS andmultiple linear regression (MLR) were used for calibration, andit was shown that a mathematical treatment using the rstderivative provided better results for NIR spectra than thatusing the second derivative and N-point smoothing. The cali-bration equation for FFA obtained from PLS was found to bebetter than that obtained from MLR.

Che Man et al.66 developed a calibration for the determina-tion of the content of FFA in crude palm oil and its fractionsbased on the NIR reectance approach. For the preparation of aFFA calibration graph, oil was hydrolyzed with 0.15% (w/w)lipase in an incubator at 60 �C (200 rpm). The optimized cali-bration models were constructed with MLR analysis based onthe C]O overtone regions from 1850–2050 nm. Calibrationswere validated with an independent set of 8–10 samples. Themethod developed was rapid, and its accuracy was generallygood for the quality control of raw material. The method wasable to analyze 12 samples per h. The hydrolytic degradation oflipids in sh oil was monitored by an NIR monochromator, intransectance mode, in the range of 1100–2500 nm.67 For FFAanalysis, PLS regression calibration was performed. The accu-racy of the calibration model developed was tested using a




Publ

ishe

d on

15

May

201

4. D

ownl

oade

d on

16/

10/2

014

13:3

7:13


validation set. It was reported that NIR spectroscopy with PLSregression could be successfully used to monitor hydrolyticdegradation of lipids in the sh oil stored under industrialconditions.

A novel method was developed for the direct estimation of FFAcontent in sunower seeds with a high oleic acid content.68 Forthe calibration, a sample set of different varieties of sunowersfrom the harvests of 2004 and 2005 was used. The NIR spectro-scopic calibration developed was calculated with a modied PLSalgorithm, standard normal variate (SNV), detrend scattercorrection and the second derivative of the spectra of groundsunower seeds. The results obtained by this method demon-strated the efficiency and cost effectiveness of the NIRmethod forthe evaluation of FFA content in sunower seeds.

Ng et al.69 described a transmission-based method using PLSand forward stepwise multiple linear regression (FSMLR) tech-niques for FFA in frying oils. The PLS model provided improvedresults compared to those obtained by the FSMLR model whenthe rst derivative as spectral treatment was used in the regionof 700–1100 nm. The best correlation coefficient between theresults of the NIR and the results of the wet chemical methodsfor FFA was 0.943. In the longer wavelength region (1100–2500nm), the FSMLR model performed better than the PLS model.The method developed was capable of analyzing 20 samples perh. Many authors reported the use of NIR calibration for FFAquantication in the oxidation70 and interesterication of oil71

to replace the traditional methods. Furthermore, it was reportedthat NIR spectroscopy might be adapted for real-time qualitycontrol purposes. In 2012, Yu et al.72 reported an automaticdetermination of acid value in edible oils by using an NIRspectrometer coupled with a continuous samples injection cell.The coefficient of determination (R2) and the root mean squareof the standard error of performance (RMSEP) of the calibrationof acid value were 0.9873 and 0.114 mg g�1, respectively, in therange of 5500–4600 cm�1 with baseline correction points at6524 cm�1 and 4823 cm�1 and SNV. By this method, an analysisrate of 90 samples per h could be achieved.

2.3.3. Raman spectrometry. Fourier transform Ramanspectrometry in combination with PLS regression was used fordirect, reagent-free determination of the FFA content in oliveoils and olives. Oils were directly investigated in a simple owcell, while olives were investigated in a dedicated sample cup,which was rotated eccentrically to the horizontal laser beamduring spectrum acquisition. External and internal (leave-one-out) validation were used to assess the predictive ability of thePLS calibration models for FFA content in oil and olives in therange 0.20–6.14% and 0.15–3.79%, respectively.73 The resultsobtained by the proposed methods could be used for screeninghigh-quality olives before processing, as well as for the on-linecontrol of the production of olive oil.

El-Abassy74 reported a visible Raman spectroscopic methodfor determining the content of FFA in extra virgin olive oil usingPLS. For the calibration, oleic acid was used to increase the FFAcontent in extra virgin olive oil by up to 0.80%. The calibrationcurve of actual FFA (%) obtained by titration versus predictedvalues based on the Raman spectral regions at 1600–945 cm�1

provided a robust and precise model with high accuracy for the


prediction of the percentage of FFA obtained. Although themethods were simple and accurate, the high cost of the instru-ment restricted the use of this technique in routine analysis.

2.3.4. Nuclear magnetic resonance spectroscopy. Nuclearmagnetic resonance (NMR) spectroscopy is increasingly beingapplied to the analysis of food and food components. In theeld of edible oils and fats, NMR provides useful informationon the lipid class composition (the relative amount of tri-acylglycerols, diacylglycerols, FFA, sterols, phospholipids, andother minor components). The determination of FFA by NMRwas rst studied on the lipids extracted from albacore tuna(Thunnus alalunga) using high-resolution 13C-NMR spectros-copy.75 FFA carbonyl resonances were detected at the lower eldof the 13C-NMR carbonyl region and used for its quantitativedetermination. NMR data agreed with data obtained using UVspectrophotometry, and showed that NMR is a suitable methodto follow lipolytic alterations. It was mentioned that the methodcould be used to monitor the changes in sh lipids and oilsduring processing or storage. Sun and Moreira76 studied thecorrelation between NMR proton relaxation times and theincrease in FFA in soybean oil degradation. Soybean oil wasdegraded by heating only and by frying corn tortilla dough for10 min h�1 up to 50 h. Both AOCS and NMRmethods were usedto determine the extent of oil degradation. Results showed thatFFA increased as heating time increased. The longitudinalrelaxation time (T1) and the transverse relaxation time (T2)decreased as the oil degradation increased. There was a linearrelationship between the NMR and AOCS measurements. Anumber of methods have been reported for investigations ofFFA in vegetable oils by 13C-NMR,77,78 31P-NMR,79 and 1H-NMR.78,80 Although the methods developed were easy, rapid andsimple, and thus could be used as a promising tool for onlinequality control, they did not attract much attention from theindustry because of the high price of the NMR equipment andthe technical skills required.

2.4. Thermogravimetric methods

Thermometric methods are also titrimetric methods that usethe change of temperature resulting from a chemical reaction todetermine the end point of the titration. The principle is thesame as that for potentiometric auto-titration. The only differ-ence is the type of sensor (detector); usually, thermistors areused. Vaughan and Swithenbank81 rst reported considerableadvances in determining the end point of the acid–base reac-tion in non-aqueous media. A novel titrimetric procedure wasreported for the determination of FFA levels in edible oils usingan automated thermometric titration with KOH in isopropanolas titrant.82 The end point was indicated by the stronglyexothermic base-catalyzed reaction between acetone and chlo-roform. The procedure was accurate, sensitive, easy to use, fast,and highly reproducible but did not get much attention.

3. Conclusion

The instrumental techniques described in this paper haveconsiderable merits over classical titration methods in terms of

Anal. Methods, 2014, 6, 4956–4963 | 4961



Publ

ishe

d on

15

May

201

4. D

ownl

oade

d on

16/

10/2

014

13:3

7:13


speed, automation and reduction of costly as well as toxicchemicals. Colorimetric methods have poor sensitivity andneed frequent calibration. Electrochemical methods requiretitrant standardization while accuracy and reproducibilitydepend on the conditions of the electrodes. Optimization ofseveral parameters is also very essential. None of the methodsdescribed, however, are specic to FFA as they all measure totalacidity rather than that of FFA. Among the spectroscopicmethods, FTIR methods have proved to be most benecial interms of simplicity, economy, its environmentally friendlynature and speed of analysis, especially when using the directapproach for the determination of FFA content in vegetable oils.However, selection of the method for FFA analysis depends onthe availability of equipment, skilled workers, purpose of studyand number of samples.

Conflict of interest

The authors declare that there are no conicts of interest.

Acknowledgements

The authors would like to thank the National Centre of Excel-lence in Analytical Chemistry, University of Sindh, Jamshoro,Pakistan for the nancial support.

References

1 R. D. O'Brien, Fats and Oils: Formulating and Processing forApplications, Technomic Publishing Company, Lancaster,PA, 1998.

2 Official Methods and Recommended Practice of the American OilChemists Society, Champaign, IL, 4th edn, 1989.

3 Official methods of analysis of the Association of OfficialAnalytical Chemists, USA, 15th edn, 1990.

4 Determination of the free fatty acids in EuropeanCommission Regulation (EEC) No. 2568/91, Off. J. Eur.Communities: Legis., 248, 5.9.91, p. 6.

5 C. W. Ayers, Anal. Chim. Acta, 1956, 15, 77–83.6 D. A. Baker, J. Am. Oil Chem. Soc., 1964, 41, 21–22.7 G. S. Bains, S. V. Rao and D. S. Bhatia, J. Am. Oil Chem. Soc.,1964, 41, 831–832.

8 R. R. Lowry and I. J. Tinsley, J. Am. Oil Chem. Soc., 1976, 53,470–472.

9 L. Ekstrom, J. Am. Oil Chem. Soc., 1981, 58, 935–938.10 D. Y. Kwon and J. S. Rhee, J. Am. Oil Chem. Soc., 1986, 63, 89–

92.11 M. Brenardez, L. Pastoriza, G. Sampedro, J. J. Herrera and

M. L. Cabo, J. Agric. Food Chem., 2005, 53, 1903–1906.12 M. Gyorik, Z. Ajtony, O. Doka, A. Alebic-Juretic, D. Bicanic

and A. Koudijs, Instrum. Sci. Technol., 2006, 34, 119–128.13 J. S. Canham and G. E. Pacey, J. Am. Oil Chem. Soc., 1987, 64,

1004–1007.14 R. Puchades, A. Suescun and A. Maquieira, J. Sci. Food Agric.,

1994, 66, 473–478.15 Z. Zhi, A. Rios and M. Valcarcel, Anal. Chim. Acta, 1996, 318,

187–194.

4962 | Anal. Methods, 2014, 6, 4956–4963

16 P. Linares, M. D. Luque de Castro and M. Valcarcel, Anal.Chim. Acta, 1989, 225, 431–436.

17 P. G. Nouros, A. G. Constantinos and M. G. Polissiou, Anal.Chim. Acta, 1997, 351, 291–297.

18 E. Mariotti and M. Mascini, Food Chem., 2001, 73, 235–238.19 E. Kardash and Y. I. Tur'yan, Croat. Chem. Acta, 2005, 78, 99–

103.20 B. Saad, C. W. Ling, M. S. Jab, B. P. Lim, A. S. Mohamad Ali,

W. T. Wai andM. I. Saleh, Food Chem., 2007, 102, 1407–1414.21 H. F. Ayyildiz, H. Kara and S. T. H. Sherazi, Lipids, 2011, 46,

1181–1190.22 A. Makahleh and B. Saad, Anal. Chim. Acta, 2011, 694, 90–94.23 J. A. Aricetti and M. Tubino, J. Am. Oil Chem. Soc., 2012, 89,

2113–2115.24 P. A. Walde, J. Am. Oil Chem. Soc., 1990, 67, 110–115.25 P. A. Walde and C. Nastruzzi, Food Chem., 1991, 39, 249–256.26 I. B. Rejeb and M. Gargouri, Anal. Lett., 2011, 44, 1454–1462.27 S. N. Fedosov, J. Brask and X. Xu, J. Am. Oil Chem. Soc., 2012,

89, 2155–2163.28 C. G. Zoski, Handbook of Electrochemistry, Elsevier Science,

2007.29 K. Takamura and Y. Hayakawa, Microchem. J., 1972, 17, 546–

555.30 F. Kusu and K. Takamura, J. AOAC Int., 1994, 77, 1686–1689.31 T. Fuse, F. Kusu and K. Takamura, Bunseki Kagaku, 1995, 44,

29–33.32 T. Fuse, F. Kusu and K. Takamura, J. Pharm. Biomed. Anal.,

1997, 15, 1515–1519.33 T. Fuse, F. Kusu and K. Takamura, J. Chromatogr. A, 1997,

764, 177–182.34 A. Kotani, F. Kusu and K. Takamura, Anal. Chim. Acta, 2002,

465, 199–206.35 S. G. Li, H. Zhang and W. T. Xue, Eur. J. Lipid Sci. Technol.,

2007, 109, 1088–1094.36 X. Yu, C. Yang, S. Du and J. Gao, Food Anal. Meth., 2012, 5,

1453–1458.37 T. M. Lapshina, Y. I. Turyan and S. I. Danilchuk, J. Anal.

Chem. USSR, 1991, 46, 833–840.38 Y. I. Turyan, O. Y. Berzin, I. Kuselman and A. Shenhar, J. Am.

Oil Chem. Soc., 1996, 73, 295–301.39 O. Y. Berezin, Y. I. Turyan, I. Kuselman and A. Shenhar,

J. Am. Oil Chem. Soc., 1996, 73, 1707–1711.40 O. Y. Berezin, Y. I. Turyan, L. Kogan, I. Kuselman and

A. Shenhar, J. Am. Oil Chem. Soc., 1997, 74, 1339–1341.41 I. Kuselman, Y. I. Turyan, O. Y. Berezin, L. Kogan and

A. Shenhar, J. AOAC Int., 1998, 81, 873–879.42 A. Velasco-Arjona and M. D. Luque de Castro, J. Am. Oil

Chem. Soc., 1998, 75, 1849–1853.43 I. Kuselman, Y. I. Turyan, I. Burenko and B. Anisimov,

Talanta, 1999, 49, 629–637.44 Y. I. Turyan and E. I. Garibyan, J. Am. Oil Chem. Soc., 2008,

85, 91–92.45 E. O. Y. I. Turyan, Food Chem., 2012, 132, 1562–1565.46 V. del Rıo, M. S. Larrechi and M. P. Callao, Talanta, 2010, 81,

1572–1577.47 A. C. Lanser, G. R. List, R. K. Holloway and T. L. Mounts,

J. Am. Oil Chem. Soc., 1991, 68, 448–449.




Publ

ishe

d on

15

May

201

4. D

ownl

oade

d on

16/

10/2

014

13:3

7:13


48 A. A. Ismail, F. R. van de Voort and J. Sedman, J. Am. OilChem. Soc., 1993, 70, 335–341.

49 E. Bertran, M. Blanco, J. Coello, H. Iturriaga, S. Maspoch andI. Montoliu, J. Am. Oil Chem. Soc., 1999, 76, 611–616.

50 Y. B. Che Man and G. Setiowaty, Food Chem., 1999, 66, 109–114.

51 Y. B. Che Man, M. H. Moh and F. R. van de Voort, J. Am. OilChem. Soc., 1999, 76, 485–490.

52 T. Verleyen, R. Verhe, A. Cano, A. Huyghebaert and W. DeGreyt, J. Am. Oil Chem. Soc., 2001, 78, 981–984.

53 F. A. Inon, J. M. Garrigues, S. Garrigues, A. Molina-Dıazand M. Dela Guardia, Anal. Chim. Acta, 2003, 489, 59–75.

54 S. T. H. Sherazi, S. A. Mahesar, M. I. Bhanger, F. R. Van DeVoort and J. Sedman, J. Agric. Food Chem., 2007, 55, 4928–4932.

55 S. Naz, S. T. H. Sherazi, F. N. Talpur, S. A. Mahesar andH. Kara, J. AOAC Int., 2012, 95, 1570–1573.

56 X. Yu, S. Du, F. R. van de Voort, T. Yue and Z. Li, Anal. Sci.,2009, 25, 627–632.

57 X. Yu, F. R. van de Voort, J. Sedman and J. Gao, Anal. Bioanal.Chem., 2011, 401, 315–324.

58 M. J. A. Canada, A. R. Medina and B. Ledle, Appl. Spectrosc.,2001, 55, 356–360.

59 A. Al-Alawi, F. R. van de Voort and J. Sedman, J. Am. Oil Chem.Soc., 2004, 81, 441–446.

60 A. Al-Alawi, F. R. van de Voort and J. Sedman, Spectrosc. Lett.,2005, 38, 389–403.

61 A. Al-Alawi, F. R. van de Voort, J. Sedman and A. Ghetler,JALA, 2006, 11, 23–29.

62 Y. Li, D. L. Garcia-Gonzalez, X. Yu and F. R. van de Voort,J. Am. Oil Chem. Soc., 2008, 85, 599–604.

63 A. N. A. Aryee, F. R. van de Voort and B. K. Simpson, ProcessBiochem., 2009, 44, 401–405.

64 E. N. Frankel, A. M. Nash and J. M. Snyder, J. Am. Oil Chem.Soc., 1987, 64, 987–992.


65 Z. Hui-Zhen and L. Tung-Ching, J. Agric. Food Chem., 1997,45, 3515–3521.

66 Y. B. CheMan andM. H. Moh, J. Am. Oil Chem. Soc., 1998, 75,557–562.

67 D. Cozzolino, I. Murray, A. Chree and J. R. Scaife, Food Sci.Technol., 2005, 38, 821–828.

68 C. R. Moschner and B. Biskupek-Korell, Eur. J. Lipid Sci.Technol., 2006, 108, 606–613.

69 C. L. Ng, R. L. Wehling and S. L. Cuppett, J. Agric. FoodChem., 2007, 55, 593–597.

70 J. A. Gerde, C. L. Hardy, C. R. Hurburgh Jr and P. J. White,J. Am. Oil Chem. Soc., 2007, 84, 519–522.

71 L. P. Houmoller, D. Kristensen and H. Rosager, Talanta,2007, 71, 868–873.

72 X. Yu, J. Zhang, Q. Li, C. Xu and J. Gao, Trans. Chin. Soc. Agric.Mach., 2012, 43, 150–154.

73 B. Muik, B. Lendl, A. Molina-Dıaz and M. J. Ayora-Canad,Anal. Chim. Acta, 2003, 487, 211–220.

74 R. M. El-Abassy, P. Donfack and A. Materny, J. Am. Oil Chem.Soc., 2009, 86, 507–511.

75 R. Sacchi, I. Medina, S. P. Aubourg, I. Giudicianni,L. Paolillo and F. Addeo, J. Agric. Food Chem., 1993, 41,1247–1253.

76 X. Sun and R. G. Moreira, J. Food Process. Preserv., 1996, 20,157–167.

77 S. Ng, J. Am. Oil Chem. Soc., 2000, 77, 749–755.78 R. Kumar, V. Bansal, A. K. Tiwari, M. Sharma, S. K. Puri,

M. B. Patel and A. S. Sarpal, J. Am. Oil Chem. Soc., 2011, 88,1675–1685.

79 F. M. Dayrit, O. E. M. Buenafe, E. T. Chainani and I. M. S. DeVera, J. Agric. Food Chem., 2008, 56, 5765–5769.

80 J. K. Satyarthi, D. Srinivas and P. Ratnasamy, Energy Fuels,2009, 23, 2273–2277.

81 G. A. Vaughan and J. J. Swithenbank, Analyst, 1965, 90, 594–599.

82 K. S. Thomas, J. Am. Oil Chem. Soc., 2003, 80, 21–24.

Anal. Methods, 2014, 6, 4956–4963 | 4963


analytical approaches for the assessment of free fatty acids in oils and fats

Documents