Sub-critical CO2 extraction of volatile flavour compounds from ghee and optimization of process parameters using response surface methodology

Sub-critical CO2 extraction of volatile flavour compounds from ghee and optimization of process parameters using response surface methodology

Journal Pre-proof Sub-critical CO2 extraction of volatile flavour compounds from ghee and optimization of process parameters using response surface me...

3MB Sizes 0 Downloads 1 Views

Journal Pre-proof Sub-critical CO2 extraction of volatile flavour compounds from ghee and optimization of process parameters using response surface methodology Neha Duhan, J.K. Sahu, S.N. Naik PII:

S0023-6438(19)31073-4

DOI:

https://doi.org/10.1016/j.lwt.2019.108731

Reference:

YFSTL 108731

To appear in:

LWT - Food Science and Technology

Received Date: 9 July 2019 Revised Date:

10 October 2019

Accepted Date: 13 October 2019

Please cite this article as: Duhan, N., Sahu, J.K., Naik, S.N., Sub-critical CO2 extraction of volatile flavour compounds from ghee and optimization of process parameters using response surface methodology, LWT - Food Science and Technology (2019), doi: https://doi.org/10.1016/ j.lwt.2019.108731. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Ltd.

1

Sub-critical CO2 Extraction of Volatile Flavour Compounds from Ghee and

2

Optimization of Process Parameters using Response Surface Methodology

3 4

Neha Duhan, J K Sahu* and S.N. Naik

5

Food and Bioprocess Engineering Laboratory, Centre for Rural Development and

6

Technology, Indian Institute of Technology Delhi, New Delhi – 110 016, India

7 8

*

9

6349 (O); Fax: +91-11-2659-1121 (O)

Corresponding Author: Email: [email protected] (Prof. J.K. Sahu), Tel: +91-11-2659-

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25

1

26

ABSTRACT

27

The volatile flavor compounds of ghee were isolated using sub-critical CO2 extraction. The

28

effect of extraction pressure (3 to 8 MPa), temperature (3 to 38 oC), and time (19 to 221 min)

29

on the extractability of flavor compounds i.e., δ-dodecalctone (C12), δ-tetradecalctone (C14),

30

3-ethyl-3-methyl heptane and total extraction yield as response variables were studied. The

31

experiments were conducted using a central composite rotatable design of the independent

32

variables. The study revealed that the pressure had a positive effect on the extractability of all

33

three compounds. The optimized sub-critical CO2 extraction pressure, temperature and time

34

for maximum recovery of these compounds were observed at 7 MPa, 10 oC and 66 min,

35

respectively. The values of yields of δ-dodecalctone (C12), δ-tetradecalctone (C14), 3-ethyl-

36

3-methyl heptane, and total extraction yield at the optimized conditions were found to be

37

were found to be 216.54, 109.7, 64.82 ng.g-1 and 5.89%(w/w), respectively.

38 39

Keywords: Heat desiccated products; Ghee; Flavour; Volatile Compounds

40 41

1. Introduction

42

Ghee is a heat desiccated fat-rich milk product and popularly consumed in the middle east

43

and Asian countries. Codex (2011) defined ghee as a product exclusively obtained from milk,

44

cream or butter from various animal species by means of processes that result in the almost

45

total removal of moisture and total solids-not-fats (SNFs) with a specially developed flavour

46

and physical structure. As per Food Safety and Standards Authority of India (FSSAI)

47

regulations (2011), ghee is referred as the pure clarified fat rich product derived solely from

48

milk, curd, cream or butter to which no colouring agent or preservative is added. Ghee is

49

distinct from concentrated milk fats such as anhydrous milk fat and butter oil, both in

50

methods of preparation and flavour composition.

2

51

Different methods adopted to prepare ghee, either from milk or cream, are reviewed by

52

Ganguli & Jain (1973) and Sserunjogi, Abrahamsen, & Narvhus, (1998). In all methods, high

53

temperature induces biochemical changes of milk proteins and lactose developing typical

54

volatile flavour profile in ghee. Chemistry of volatile flavour compounds of ghee has been

55

extensively studied by Wadhwa & Jain (1990), Wadodkar, Punjrath, & Shah (2002), and

56

Newton, Fairbanks, Golding, Andrewesc, & Gerrard (2012). The authors reported that the

57

compounds developed due to temperature-induced decomposition of milk fat, proteins, amino

58

acids, lactose and glucose during clarification play a crucial role in the development of ghee

59

flavour. A major group of compounds generated during ghee manufacturing include

60

aldehydes, ketones, free fatty acids (FFAs), carboxylic acids, lactones and alcohols (Wadhwa

61

& Jain, 1990; Newton et al., 2012).

62 63

The contribution of carbonyl compounds (aldehydes and ketones) to ghee flavour was

64

reported by Rao & Ramamurthy (1984) and Yadav & Srinivasan (1992). The sources of

65

carbonyl compounds in ghee are quite diverse. Lactones have a coconut-like flavour which is

66

majorly associated with the characteristic flavour of ghee (Schlutt, Moran, Schieberle, &

67

Hofmann, 2007). δ-lactones and γ-lactones are among the major classes of lactones present in

68

ghee among which, δ-decalactone (C10), δ-dodecalctone (C12) and δ-tetradecalctone (C14)

69

seem to be the most important compounds influencing volatile compounds of ghee flavour

70

(Wadodkar, Murthi, & Punjrath, 1996). Free fatty acids (FFAs) are other groups of

71

compounds contribute to the ghee flavour and their quantities are closely related to the

72

flavour quality (Yadav & Srinivasan, 1984).

73 74

Among various acids, ghee contains about 65% saturated fatty acid, 32% monounsaturated

75

fatty acid (MUFA) and 3% polyunsaturated fatty acid (PUFA) (Gupta, Singh, Gularia, &

3

76

Gupta, 2015). It is well established that a higher intake of dietary saturated fats and

77

cholesterol increases the risk of cardiovascular diseases, obesity and type 2 diabetes (Grundy,

78

2003). Thus, despite being a popular milk product and part of Indian culture, nowadays,

79

many people tend to avoid consumption of ghee in their diets owing to various health

80

constraints.

81 82

Extraction of volatile flavour compounds from ghee for development of various ghee flavour

83

enriched value-added products would be a novel approach to solve the problem. Liquid/sub-

84

critical CO2 has high selectivity towards aroma representative constituents such as esters,

85

aldehydes, ketones and alcohols and thus, was used by many authors for isolation of essential

86

oil and flavours from natural plant based products (Moyler, 1993; Tuan & Ilangantileke,

87

1997; Naik, Lentz, & Maheshwari, 1989; Rout, Naik, Rao, Jadeja, & Maheshwari, 2007;

88

Rout, Naik, & Rao, 2010; Spricigo, Pinto, Bolzan, & Novais, 1999). Extraction temperature,

89

pressure, time, particle size and porosity of the materials are among the major factors that

90

determine the quality and quantity of the extract (Orellana, Smith, & Kitchens, 2013; Moyler,

91

1993; Ahangari & Sargolzaei, 2012). No work has been reported so far that deals with

92

extraction of volatile flavour compounds of ghee employing SC-CO2 technology. Thus, the

93

present study was carried out to bridge the existing gap through usage of SC-CO2 technology

94

for extraction of ghee volatile flavour compounds.

95 96

The first part of the paper describes the identification of major volatile compounds

97

responsible for developing flavour profile of ghee using gas chromatography-mass

98

spectroscopy (GC-MS). The second part of the paper describes the effect of SC-CO2

99

extraction pressure, temperature and time on the recovery of these volatile compounds and

100

optimization of process variables using response surface methodology. Based on the literaure

4

101

and preliminary analysis of volatile compounds of ghee, in the present study, δ-

102

dodecalactone, δ-tetradecalactone, 3-ethyl-3-methyl heptane and total extraction yield were

103

selected as the key compounds for quantification and further optimization.

104 105

2. Materials and Methods

106

2.1. Materials

107

Milk (fat 6.15%, solids-not-fat 8.65%, protein 3.2%, lactose 5.2% and ash 0.6%) collected

108

from a particular cow of Indian breed in Hisar district, Haryana of India was used for the

109

preparation of ghee. Ghee was prepared manually using the method as described by Ganguli

110

& Jain (1973). A constant heating temperature of 135±3 oC was employed during the

111

desiccation process. The ghee samples were kept in sealed containers under refrigerator

112

storage at – 20 oC for further use.

113 114

2.2. Chemicals

115

HPLC grade diethyl ether was procured from Merck India, Mumbai. Anhydrous sodium

116

sulphate was supplied by Fisher Scientific, USA. All standards used in the study were

117

procured from Sigma-Aldrich, USA. CO2 (purity 99.5%) was supplied by Laser Gases, New

118

Delhi. Polytetrafluoroethylene (PTFE) vials were purchased from Borosil India, Mumbai.

119 120

2.3. Analysis of volatile flavour compounds of ghee

121

Fresh ghee sample was analysed to identify the major volatile compounds responsible for the

122

flavour profile of ghee using a GC-MS system. Prior to direct injection to GC-MS, the

123

volatile compounds were separated using a simultaneous distillation extraction (SDE) process

124

taking diethyl ether as an extracting solvent in a modified Godefroot apparatus (Godefroot,

125

Sandra, & Verzele, 1981). The vapour transport arms of the apparatus were thermally

5

126

isolated prior to extraction. After settling of the two-phase system in the demixing part of the

127

apparatus, the sample flask was heated at 90 °C whereas the solvent flask was heated at 45

128

°C. Ice water at 5 °C was circulated continuously through cold finger during the entire

129

process to ensure condensation of vapours. Once the contents of both flasks started boiling,

130

the process was continued for 1 h to allow volatile compounds to be collected in diethyl

131

ether. The diethyl ether extract was added with 10 µl of 0.1% 2-ethylhexyl acetate as internal

132

standard and concentrated to a final volume of 1 ml and then dried over anhydrous sodium

133

sulphate overnight for complete removal of moisture.

134 135

A Shimadzu QP2010 Ultra High-End GC-MS system was used for analysis of the volatiles

136

extracted by SDE process using low polarity, crossbond diphenyl dimethyl polysiloxane

137

based column (30 m x 0.25 mm internal diameter, film thickness 0.25 µm). Helium was used

138

as carrier gas. Initial column oven temperature was kept at 50 °C for 3 min and then

139

programmed to 150 °C at the rate of 3 °C.min-1. Holding time at this temperature was 2 min

140

after which, the temperature was programmed to 210 °C at the rate of 4 °C.min-1. The final

141

temperature was held for 17 min, making the total run time 70 min. The column flow rate

142

was 1.2 mL.min-1. The injector temperature was 260 °C with a split ratio of 1:10 while the

143

interface temperature was 270 °C. MS was scanned at 70 eV over m/z 40–650 at an ion

144

source temperature of 230 °C. Identification of volatile compounds was carried out by

145

comparing their mass spectra and retention indices to those of standard compounds obtained

146

at same operating conditions and using Wiley8/NIST14 database. Retention indices were

147

determined by comparing them with linear retention indices of C7-C14 alkane standards.

148

Quantification of the compounds was conducted using 2-ethylhexyl acetate as internal

149

standard and δ-dodecalactone, δ-tetradecalactone, decanoic acid and myristic acid as external

150

standards at three concentrations for calculation of response factors.

6

151

2.4. Experimental design of sub-critical CO2 extraction variables

152

In the present study, SO-CO2 extraction pressure (X1), holding temperature (X2) and time (X3)

153

were the independent variables while concentration (ng.g-1) of δ-dodecalactone, δ-

154

tetradecalactone, 3-ethyl-3-methyl heptane, and total extraction yield (% w/w) were the

155

response variables. From the preliminary trials, the minimum (Xmin) and maximum (Xmax)

156

levels of the extraction pressure, temperature and time were fixed from 3 to 8 MPa, 3 to 38

157

°C, and 19 to 221 min, respectively. A 3 variable-5 level central composite rotatable design

158

of experimental variables was used to carry out experiments. A total of twenty experiments

159

including six replicates at the central point (Table 1) were conducted wherein one experiment

160

was performed in super-critical CO2 at 38 °C. The experimental data were fitted to a second-

161

order polynomial regression equation as follows:

162

=

163

+∑

+∑

+∑







(1)

164 165

where Yn represents the response variables, xi and xj are the coded values of independent

166

variables, ao is a constant, ai, aii and aij are the linear, quadratic, and interactive coefficients,

167

respectively. The real value X can be obtained from its coded value x by using Eq. (2).

168 169

=

+

,

=(



+

)/2 ,

=(





)/

,

=( −

)/ (2)

170 171 172

where Xmax is the maximum value of X and +am is the coded value associated with Xmax. Xmin

173

is the minimum value of X and -am is its coded value. XM is the arithmetic mean of Xmax and

174

Xmin. By varying two variables within the experimental range and holding the other variable as

175

constant at the central point, three-dimensional surface response plots were generated. The

7

176

test of statistical significance, based on the total error criteria, with a confidence level of 95%

177

(p<0.05) was determined using analysis of variance (ANOVA). The experimental design and

178

statistical analysis of the data were carried out using Design Expert Software (Version 10.0.1,

179

Stat-Ease, Inc., Minneapolis, USA).

180 181

2.5. Sub-critical CO2 extraction

182

A sub-critical CO2 extraction system was used for fractionation of volatile flavour

183

compounds from ghee. The system includes a CO2 cylinder, a heat exchanger, a chiller, a

184

high-pressure pump, an extraction vessel and a product collector. A 25 g ghee sample was

185

used in each run. The sample was mixed with 3 mm glass beads, placed inside the vessel, and

186

subjected to the desired extraction temperature and pressure condition in a closed system.

187

CO2 was allowed to enter the extraction vessel through the high-pressure pump. After

188

building up the required temperature and pressure condition, the extraction time was

189

monitored. Upon completion of the extraction time, the product (flavour and other associated

190

compounds) was collected. Volatile compounds of the extracts obtained from the extraction

191

unit were further analysed using SDE-GC-MS process as described in section 2.3. Total

192

extraction yield of the extracted product obtained at various experimental conditions was

193

calculated in terms of % (w/w) with reference to the amount of ghee taken in each case.

194 195

3. Results and discussion

196

3.1. Volatile profile of fresh ghee samples

197

The volatile profile of ghee obtained from fractionation using simultaneous distillation and

198

extraction (SDE), followed by direct injection GC-MS is presented in Table S1. About 41

199

compounds could be identified among which, quantitation of 39 compounds could be made.

200

As evident from the table, the characteristic flavour compounds responsible for developing

8

201

heat-induced distinct flavour of ghee were lactones, fatty acids, esters, aldehydes,

202

hydrocarbons and certain other derived compounds generated as a result of synergic

203

interaction of these compounds with each other(s). Fig. S2 shows GC–MS chromatogram of

204

the volatile components obtained from SDE-GCMS. Some of the selected volatile

205

compounds include δ-tetradecalactone, δ-dodecalactone, tetradecanal, pentadecanal-, 2-

206

pentadecanone, 2-nonadecanone, butyl isodecyl phthalate, glycerol 1-myristate, octadecanoic

207

acid, 2-propenyl ester, octadecanoic acid, 2,3-dihydroxypropyl ester, glycidyl palmitate,

208

glycidyl oleate, tetradecanoic acid, dodecanoic acid, dodecane, 4,6-dimethyl-, heptadecane,

209

8-heptadecene,9-octyl, and 2-palmitoylglycerol. Similar results were reported in the studies

210

conducted by Wadodkar et al., (2002) while analysing various types of ghee. Although

211

lactones have been reported as the major class of compounds significantly affecting flavour

212

of ghee, all the compounds and their interaction with each other(s) decide the volatile flavour

213

profile of ghee (Newton, 2012).

214 215

3.2. Effect of sub-critical CO2 extraction variables on extraction recovery of volatile

216

compounds

217

Table 1 shows the measured values of δ-dodecalactone, δ-tetradecalactone, 3-ethyl-3-methyl-

218

heptane, and total extraction yield at various experimental conditions. Effect of SC-CO2

219

extraction parameters on extraction recovery of these compounds are described in following

220

sections.

221 222

3.2.1. Effect of SC-CO2 extraction parameters on the recovery of δ-dodecalactone

223

The value of δ-dodecalactone was observed to vary 0 to 288.22 ng.g-1 within the different

224

experimental ranges. δ-dodecalactone is among the most significant compounds associated

225

with typical ghee flavour profile imparting sweet and coconut-fat type aroma (Wadhwa &

9

226

Jain, 1990). Wadodkar et al., (2002) reported a maximum value of 30 ng.g-1 for different

227

ghee samples manufactured by industries. A second order regression equation was developed

228

to analysis the recovery of δ-dodecalactone as a function of the coded value of independent

229

variables and the same is given by Eq. (3).

230 231 232

= 3.90 + 58.68 – 8.84 47.11

+ 4.39

− 11.16

− 21.32

− 30.94 R2 = 0.879

− 0.13

+ 34.79

+ (3)

233 234

where, Y1 refers to δ-dodecalactone content (ng.g-1) in the extract, x1, x2 and x3 are the

235

extraction pressure (MPa), temperature (oC), and time (min), respectively. The linear terms of

236

Eq. (3) show that the recovery of δ-dodecalactone increased with increase in levels of

237

extraction pressure and decreased with increase in levels of both extraction temperature and

238

time. Also, the extraction pressure had the maximum positive effect whereas the time,

239

followed by temperature had a negative effect on the recovery of δ-dodecalactone. Fig. 1

240

show the recovery of δ-dodecalactone as a function of extraction pressure and temperature,

241

pressure and time, and temperature and time, keeping the constant variable at the central

242

point. It is quite interesting to observe that the increase in extraction pressure from 4 to 7

243

MPa coupled with a decrease in temperature increased δ-dodecalactone content in the volatile

244

extract (Fig.1a). Similarly, the increase in pressure from 4 to 7 MPa along with a decrease in

245

extraction time from 180 to 60 min enhanced extractability of δ-dodecalactone (Fig.1b).

246

However, increasing temperature and time had a negative effect on the recovery of δ-

247

dodecalactone content. Increase in the extraction pressure had the maximum positive linear

248

effect whereas extraction temperature and time exhibited a negative linear effect on δ-

249

dodecalactone (Fig.1c). A similar trend was reported by Haan, Graauw, Schaap, & Badings

10

250

(1990) in case of milk fat where an increase in pressure along with a decrease in extraction

251

temperature resulted in the effective separation of δ-dodecalactone from the sample matrix.

252 253

3.2.2. Effect of SC-CO2 extraction on the recovery of δ-tetradecalactone

254

δ-tetradecalactone is one of the major lactones associated with mellow-creamy-fatty flavour,

255

responsible for adding creamy perception to milk-based fat-rich products such as cream,

256

butter and ghee. Obi et al., (2018) reported value of δ-tetradecalactone as 25 µg.g-1 for butter

257

oil. Within the selected range of variables, the values of δ-tetradecalactone in the present

258

study varied from 0 to 503.12 ng.g-1. The effect of coded values of independent variables on

259

the recovery of δ-dodecalactone is shown by Eq. (4).

260 261 262

= 25.75 + 74.42 25.77

+ 77.90

+ 86.70 − 11.28

+ 12.61

+ 13.64



R2 = 0.829

+ 27.67

+ 64.23

+ (4)

263 264

where Y2 refers to the recovery of δ-tetradecalactone (ng.g-1), x1, x2 and x3 are the coded

265

values of extraction pressure (MPa), temperature (oC), and time (min), respectively. Linear

266

terms of Eq. (4) show that the extraction pressure had the maximum positive effect on the

267

recovery of δ-tetradecalactone followed by temperature and time.

268 269

Unlike δ-dodecalactone, the effect of temperature raise resulted in an increase of δ-

270

tetradecalactone content throughout the range. Although pressure increase resulted in the

271

higher recovery of δ-tetradecalactone in the extract, the effect was not as profound as

272

observed in the case of temperature which showed significant linear and quadratic effect

273

(Table 2, Fig.2a). The interactive effect of extraction pressure and time was not profound

274

(Fig.2b). However, the interactive effect of temperature and time played a crucial role in the

11

275

extraction of maximum δ-tetradecalactone as apparent from Fig.2(c) and Table 2. Semi-

276

volatile compounds like δ-tetradecalactone have different recovery rate than other lactones

277

due to the difference in polarity (Sarrazin, Frerot, Bagnoud, Aeberhardt, & Mark, 2011).

278

Moreover, as observed by Schlutt et al., (2007), heat-treatment results in increased

279

concentration of δ-tetradecalactone in dairy cream comparative to all other lactone

280

compounds, which further agreed with the trend observed in the present study.

281 282

3.2.3. Effect of SC-CO2 extraction on the recovery of 3-ethyl-3-methylheptane

283

3-ethyl-3-methyl heptane is branched hydrocarbon, associated with flavour in fat-rich milk-

284

based products like cheese and cream. In the study, the values of 3-ethyl-3-methyl heptane

285

varied from 0 to a maximum value of 142 ng.g-1. Second order regression equation obtained

286

for the quantification of 3-ethyl-3-methyl-heptane in terms of coded values of pressure,

287

temperature and time is given by Eq. 5.

288 289 290

= 24.18 + 26.50 15.40

+ 19.18

− 4.77

+ 22.44

+ 10.44

+ 11.47

R2 = 0.869

+ 12.66

+ 7.77

+

(5)

291 292

where, Y3 represents 3-ethyl-3-methylheptane content (ng.g-1) in the extract, x1, x2 and x3 are

293

the extraction pressure (MPa), temperature (oC), and time (min), respectively. Linear terms of

294

Eq. (5) revealed that the pressure had the maximum positive effect followed by time, whereas

295

a decrease in temperature had the positive effect on the recovery of 3-ethyl-3-methyl-heptane.

296

The response surface plots represented in Fig.3a-c show a similar trend as inferred by Eq. (5).

297

The extraction pressure and time were found to be the most significant parameters in linear

298

terms whereas the quadratic effect was observed in case of extraction pressure, temperature

299

and time (Table 2). Increase in temperature resulted in decreased content of the target

12

300

hydrocarbon which could be explained on the basis of the highly volatile nature of 3-ethyl-3-

301

methylheptane in CO2. Decreasing the temperature while increasing in pressure increased

302

extractability 3-ethyl-3-methylheptane content as increasing pressure and density of CO2

303

increases solubility of hydrocarbons, separating them from the native matrix (Ozturk &

304

Celiktas, 2017).

305 306

3.2.4. Effect of SC-CO2 extraction on total extraction yield

307

The extraction yield varied from 1.83 to 17.14% (w/w). The maximum yield (15.84%) was

308

obtained at the extraction conditions of pressure, temperature and time at 4 MPa, 31 °C and

309

180 min, respectively. However, interestingly, the quantitative effect on flavour compounds

310

associated positively with ghee, was quite the opposite, resulting in the yield of 39, 78, 43

311

ng.g-1, respectively for δ-dodecalactone, δ-tetradecalactone and 3-ethyl-3-methyl heptane at

312

this condition. Quantification of extraction yield as a function of coded values of independent

313

variables could be best predicted by Eq. (6).

314 315 316

&

= 7.62 − 1.15

0.48

− 1.25

+ 1.94

+ 0.34

− 1.65

R2 = 0.779

− 0.57

− 0.95

+ 1.65



(6)

317 318

where Y4 is the extraction yield (% w/w), x1, x2 and x3 are the coded values of extraction

319

pressure (MPa), temperature (oC), and time (min), respectively. Linear terms of Eq. (6)

320

showed that reducing pressure level and increasing temperature and time resulted in the

321

maximum extraction yield of the extract which has been clearly depicted in response surface

322

plots (Fig.4a-c). Higher total extraction yield in higher temperature could be explained in

323

terms of enhanced extractability of triglycerides in the product stream alongside target

13

324

compounds as observed by Arul, Tardif, Boudreau, McGinnisb, & Lencki (1994) in case of

325

milk fat.

326 327

3.3. Fitting of the regression models and statistical analysis

328

Statistical significance and fitness of the developed regression equations were calculated

329

through analysis of variance (ANOVA) and presented in Table 2. Significance was judged

330

statistically by computing F-value at a p of 0.001, 0.01 or 0.05. The ANOVA show that the

331

model F-values for δ-dodecalactone, δ-tetradecalactone, 3-ethyl-3-methylheptane and total

332

extracttion yield of 6.54, 5.47, 3.64, and 6.26, respectively, were highly significant

333

(corresponding p-value are 0.0023, 0.0069, 0.0283, and 0.0051). The respective values of

334

coefficient of determination (R2) of δ-dodecalactone, δ-tetradecalactone, 3-ethyl-3-

335

methylheptane, and extraction yield were found to be 0.879, 0.829, 0.869, and 0.776 inferring

336

that the developed regression models could predict the response variables adequately. Fig.S3

337

shows the plot of model prediction versus actual experimental data for δ-dodecalactone, δ-

338

tetradecalactone, 3-ethyl-3-methylheptane, and yield, respectively. The observed points on

339

these plots reveal that the actual values were distributed relatively close to the straight line

340

which was in direct consequence of the corresponding values of the correlation coefficient.

341 342

3.4. Process parameters optimization

343

Optimization of independent parameters was carried out with the help of numerical

344

optimization using Design Expert Software. Maximum values of δ-dodecalactone, δ-

345

tetradecalactone, 3-ethyl-3-methyl-heptane, and total extraction yield in the volatile

346

compounds were fixed as criteria for the optimization. The parameters as obtained with

347

respect to the above-fixed responses were: extraction pressure = 7 MPa, temperature =10 °C,

348

and time =66 min. The predicted values of δ-dodecalactone, δ-tetradecalactone, 3-ethyl-3-

14

349

methyl-heptane, and total extraction yield at this optimized condition were 212.03 ng.g-1,

350

107.85 ng.g-1, 66.43 ng.g-1, and 5.71% (w/w) respectively. The actual experimental values of

351

these compounds at the optimized condition were found to be 216.54, 109.7 and 64.82 ng.g-1

352

for δ-dodecalactone, δ-tetradecalactone, and 3-ethyl-3-methyl-heptane, respectively whereas

353

the total extraction yield was observed to be 5.89% (w/w) (Table S2, Fig. S3). As the

354

experimental values were close to the predicted ones, the developed regression equations

355

were found to successfully depict the amount of volatile compounds in ghee.

356 357

3.5. Conclusion

358

Optimization of sub-critical CO2 extraction process variables i.e., extraction pressure,

359

temperature and time was carried out for maximum extractability of δ-dodecalactone, δ-

360

tetradecalactone, 3-ethyl-3-methyl-heptane and total extraction yield in the product.

361

Numerical optimization suggested the optimum extraction condition for the process in terms

362

of extraction pressure, temperature and time at 7 MPa, 10 °C and 66 min, respectively. The

363

actual experimental values for δ-dodecalactone (216.54), δ-tetradecalactone (109.7), 3-ethyl-

364

3-methyl-heptane (64.82 ng.g-1) , and the total extraction yield (5.89% w/w) were close to the

365

values predicted by model at optimized condition (δ-dodecalactone (212.03 ng.g-1), δ-

366

tetradecalactone (107.85 ng.g-1), 3-ethyl-3-methyl-heptane (66.43 ng.g-1), and total extraction

367

yield (5.71%) ). Therefore, it may be concluded that the model developed was efficient for

368

predicting the outcome in terms of concentration of selected flavour compounds and total

369

extraction yield for experiments conducted within the selected range of extraction pressure (3 to

370

8 MPa), temperature (3 to 38 oC), and time (19 to 221 min).

371 372

Conflict of interest

373

The authors declared no conflict of interest.

15

374

References

375

Ahangari, B., & Sargolzaei, J. (2012). Extraction of pomegranate seed oil using subcritical

376

propane and supercritical carbon dioxide. Theoretical Foundations of Chemical

377

Engineering, 46(3), 258–265.

378

Arul, J., Tardif, R., Boudreau, A., McGinnisb, D.S., & Lencki, R. (1994). Solubility of milk

379

fat triglycerides in supercritical carbon dioxide. Food Research International, 21, 459-

380

467

381 382

Brunner, G. (2005). Supercritical fluids: technology and application to food processing. Journal of Food Engineering, 67(1-2), 21-33.

383

Food Safety and Standards (Food Product Standards and Food Additives) Regulation (2011)

384

Ganguli, N. C & Jain, M.K. (1973) Ghee: its chemistry, processing and technology. Journal

385 386 387

of Dairy Science, 56:19–25 Godefroot, M., Sandra, P. and Verzele, M. (1981). New method for quantitative essential oil analysis. Journal of Chromatography, 203, 325.

388

Grundy, S.M. (2003). Factors determining blood cholesterol levels. In B. Caballero, P.

389

Finglas, & F. Toldra (Eds.), Encyclopedia of Food Sciences and Nutrition. 2nd Edn

390

(1237-1243), Cambridge: Academic Press.

391

Gupta, M., Singh, D., Gularia, P., & Gupta, S. (2015). GC-MS analysis and identifications of

392

chemical constituents of Syzygium aromaticum, Brassica compestris and cow ghee.

393

Journal of Chemical and Pharmaceutical Research, 7(1), 568-572.

394

Haan, A.B., Graauw, J., Schaap, J.E., & Badings, H.T. (1990). Extraction of flavours from

395

milk fat with supercritical carbon dioxide. The Journal of Supercritical Fluids, 3, 15-

396

19.

397 398

Moyler, D.A. (1993). Extraction of essential oils with carbon dioxide. Flavour and Fragrance Journal, 8,235-247.

16

399 400

Murthy, M.K. (1980). Factors affecting the composition, flavour and textural properties of ghee. Indian Dairyman, 32, 765-768.

401

Naik, S.N., Lentz, H., & Maheshwari, R.C. (1989). Extraction of perfumes and flavours from

402

plant materials with liquid carbon dioxide under liquid-vapour equilibrium conditions.

403

Fluid Phase Equilibria, 49,115-126

404

Newton, A.E., Fairbanks, A.J., Golding, M., Andrewesc, P., & Gerrard, J.A. (2012). The role

405

of the Maillard reaction in the formation of flavour compounds in dairy products – not

406

only a deleterious reaction but also a rich source of flavour compounds. Food and

407

Function, 3, 1231.

408 409

Newton, A. (2012). Flavour formation in ghee. Unpublished PhD Thesis. University of Canterbury, New Zealand.

410

Obi, J., Yoshinaga, K., Tago, A., Nagai, T., Yoshida, A., Beppu, F., & Gotoh, N. 2018.

411

Simple quantification of lactones in milk fat by solvent extraction using gas

412

chromatography mass spectrometry. Journal of Oleo Science, 67 (8), 941-948.

413 414

Orellana, J.L., Smith, T.D. and Kitchens, C.L. (2013). Liquid and supercritical CO2 extraction of fat from rendered materials. Journal of Supercritical Fluids, 79, 55–61.

415

Ozturk, E. & Celiktas, O. (2017). Supercritical CO2 extraction of hydrocarbons from

416

Botryococcus braunii as a promising bioresource. Journal of Supercritical Fluids, 130,

417

261–266.

418 419

Rao, D.V., & Ramamurthy, M. K. (1984) Polar carbonyls in cow and buffalo ghee. Indian Journal of Dairy Science, 37, 98-102.

420

Rout, P.K, Naik, S.N., Rao, Y.R., Jadeja, G., & Maheshwari, R.C. (2007). Extraction and

421

composition of volatiles from Zanthoxylum rhesta: Comparison of subcritical CO2 and

422

traditional processes. Journal of Supercritical Fluids, 42, 334–341.

17

423

Rout, P.K., Naik, S.N. & Rao, Y.R. (2010). Composition of absolutes of Jasminum sambac

424

L. flowers fractionated with liquid CO2 and methanol and comparison with liquid CO2

425

extract. Journal of Essential Oil Research, 22(5), 398-406.

426

Sarrazin, E., Frerot, E., Bagnoud, A., Aeberhardt, K., & Mark R. (2011). Discovery of new

427

lactones in sweet cream butter oil. Journal of Agriculture and Food Chemistry, 59,

428

6657–6666.

429

Schlutt, B., Moran, N., Schieberle, P., & Hofmann, T. (2007). Sensory-directed identification

430

of creaminess-enhancing volatiles and semivolatiles in full-fat cream. Journal of

431

Agriculture and Food Chemistry, 55, 9634–9645.

432

Spricigo, C.B., Pinto, L.T., Bolzan, A., & Novais, A.F. (1999). Extraction of essential oil and

433

lipids from nutmeg by liquid carbon dioxide. Journal of Supercritical Fluids, 15, 253–

434

259.

435 436 437 438 439 440

Sserunjogi, M. L., Abrahamsen, R. K., Narvhus, J. 1998. A review paper: current knowledge of Ghee and related products. International Dairy Journal, 8, 677-688. Tuan, D.Q., & Ilangantileke, S.G. (1997). Liquid CO2 extraction of essential oil from Star Anise fruits (Illicium verum H.). Journal of Food Engineering, 31, 47–57. Wadhwa, B.K., & Jain, M.K. (1990). Chemistry of ghee flavour-a review. Indian Journal of Dairy Science 43, 601-607.

441

Wadodkar, U.R., Murthi, T.N., & Punjrath, J.S. (1996). Isolation of ghee volatiles by vacuum

442

degassing, their separation and identification using gas chromatography/mass

443

spectrometry. Indian Journal of Dairy Science, 49, 185-198.

444

Wadodkar, U.R., Punjrath, J.S., & Shah, A.C. (2002). Evaluation of volatile compounds in

445

different types of ghee using direct injection with gas chromatography-mass

446

spectrometry. Journal of Dairy Research, 69, 163-171.

18

447 448

Yadav J.S., & Srinivasan, R.A. (1992) Advances in ghee flavour research. Indian Journal of Dairy Science, 45, 338-348.

449

Yadav, J.S., & Srinivasan, R.A. (1984). Qualitative and quantitative changes in flavour

450

characteristics of ghee made from ripened cream. Indian Journal of Dairy Science, 37,

451

350-356.

452

19

List of Tables Table 1 Full factorial central composite experimental design of independent variables and their corresponding response values Table 2 Analysis of variable (ANOVA) and regression parameter values for the fitted model

1

Table 1 Full factorial central composite experimental design of three independent variables and their corresponding response values S.

Pressure,

Temperature

Time

δ-

δ-tetradecalactone

3-ethyl-3-methyl

Myristic acid

Palmitic

Total extraction

no.

X1

X2

X3

dodecalactone

(ng.g-1)

heptane

(ng.g-1)

acid

yield (% w/w))

(mPa)

(°C)

(min)

(ng.g-1)

1.

4

10

60

0

0

88.27

336.537

1688.2

2.05

2.

7

10

60

235.78

191.57

96.83

29.47

742.75

8.56

3.

4

31

60

4

42.12

13

52.8

1567.93

12.31

4.

7

31

60

50.342

67.25

56.04

1434.62

4270.25

9.35

5.

4

10

180

0

0

94.848

364.8

2509.82

6.26

6

7

10

180

7.86

81.24

142.01

1530.18

1855.26

7.62

7.

4

31

180

39.02

78.03

43.35

112.71

498.53

9.84

8.

7

31

180

65.76

434.81

139.57

1309.79

1771.44

7.48

9.

3

20.5

120

0

0

0

12.8

73.6

17.14

10.

8

20.5

120

288.22

215.05

99.25

975.98

1974.01

6.32

11.

5.5

3

120

34.05

6.81

29.13

20.43

78.32

2.15

12.

5.5

38

120

12.54

503.12

91.47

1894.86

2848.82

9.26

(ng.g-1)

2

(ng.g-1)

13.

5.5

20.5

19

3.07

2.13

0

0

1783.97

1.83

14.

5.5

20.5

221

17.94

3.32

83.72

681.72

1064.44

5.23

15.

5.5

20.5

120

3.75

25.22

28.218

226.06

474.98

7.53

16.

5.5

20.5

120

3.5

24.13

26.8

225.96

473.12

7.44

17.

5.5

20.5

120

3.23

25.83

28.42

225.12

475.34

7.87

18.

5.5

20.5

120

3.45

24.27

27.6

225.05

473.12

7.77

19.

5.5

20.5

120

3.55

26.08

26.28

225.11

475.92

7.63

20.

5.5

20.5

120

3.51

25.94

28.18

227.01

474.18

7.69

3

Table 2 Analysis of variable (ANOVA) and regression parameter values for the fitted model Response

Source

Sum of square

df

Mean square

δ-dodecalactone

Model

103100.5

9

11455.61

Error

0.14

5

0.028

Cor Total

118456.88

19

Model

323743.06

9

35971.45

Error

3.72

5

0.74

Cor Total

390137.6

19

Model

28471.95

9

3163.55

Error

0.000

5

0.000

Cor Total

37180.70

19

Model

174.51

9

19.39

Error

0.12

5

0.025

Cor Total

227.94

19

δ-tetradecalactone

3-ethyl-3-methylheptane

Total extraction yield

Response Model term

δ-dodecalactone Standard error

p-value

δ-tetradecalactone Standard error

F-value

p-value

Std Dev

CV%

7.46

0.0021

39.19

100.54

5.42

0.0072

81.48

91.71

3.63

0.0284

29.51

52.29

3.63

0.0285

2.31

30.55

3-ethyl-3-methyl heptane

p-value

Standard error

4

Standard error

Total extraction yield Standard error

p-value

Intercept

15.98

33.23

12.04

0.94

218.78

x1

10.60

0.0002

22.05

0.0071

7.99

0.63

145.15

0.0185

x2

10.60

0.4241

22.05

0.0028

7.99

0.63

145.15

0.0131

x3

10.60

0.3172

22.05

0.3501

7.99

0.63

145.15

0.1819

x1x2

13.85

0.1549

28.81

0.6461

10.43

0.82

189.65

0.0043

x1x3

13.85

0.0496

28.81

0.3595

10.43

0.82

189.65

0.4703

x2x3

11.74

0.0308

28.81

0.0499

10.42

0.82

189.65

0.0046

x1x1

10.32

0.0010

21.46

0.2575

7.77

0.61

141.30

0.0586

x2x2

10.32

0.6793

21.46

0.0046

7.77

0.61

141.30

0.0089

x3x3

10.32

0.9904

21.46

0.6107

7.77

0.61

141.30

0.0106

5

List of Figures Fig.1 Response surface plot of δ-dodecalactone as a function of (a) pressure and temperature, (b) pressure and time and (c) temperature and time, keeping the constant variable at central point Fig.2 Response surface plot of δ-teradecalactone as a function of (a) pressure and temperature, (b) pressure and time and (c) temperature and time, keeping the constant variable at central point Fig.3 Response surface plot of 3-ethyl-3-methyl heptane as a function of (a) pressure and temperature, (b) pressure and time and (c) temperature and time, keeping the constant variable at central point Fig.4 Response surface plot of total extract yield as a function of (a) pressure and temperature, (b) pressure and time, and (c) temperature and time, keeping the constant variable at central point

1

Delta-dodecalactone (ng/g)

300

200

100

0

-100

31

7 28

6.4

25

5.8

22 19

Temperature (°C)

5.2

16

4.6

13 10

Pressure (mPa)

4

(a)

Delta-dodecalactone (ng/g)

300

200

100

0

-100

180

7 6.4

150 5.8

120

Time (min)

5.2 90

4.6 60

(b)

2

4

Pressure (mPa)

Delta-dodecalactone (ng/g)

300

200

100

0

-100

180

31 28

150

25 22

120

19 16

90

Time (min)

13 60

Temperature (°C)

10

(c) Fig. 1 Response surface plots of δ-dodecalactone as a function of (a) pressure and temperature, (b) pressure and time, and (c) temperature and time, keeping the constant variable at central point

Design points below predicted value

Delta-tetradecalactone (ng/g)

600 500 400 300 200 100 0 -100

31

7 28

6.4

25

5.8

22 19

Temperature (°C)

5.2

16

4.6

13 10

(a)

3

4

Pressure (mPa)

Design points below predicted value

Delta-tetradecalactone (ng/g)

600 500 400 300 200 100 0 -100

180

7 6.4

150 5.8

120

5.2 90

Time (min)

Pressure (mPa)

4.6 60

4

(b)

Delta-tetradecalactone (ng/g)

600 500 400 300 200 100 0 -100

180

31 28

150

25 22

120

Time (min)

19 16

90

13 60

Temperature (°C)

10

(c) Fig. 2 Response surface plot of δ-teradecalactone as a function of (a) pressure and temperature, (b) pressure and time, and (c) temperature and time, keeping the constant variable at central point

4

3-ethyl-3-methylheptane (ng/g)

160 140 120 100 80 60 40 20 0

31

7 28

6.4

25

5.8

22 19

Temperature (°C)

5.2

16

Pressure (mPa)

4.6

13 10

4

(a)

3-ethyl-3-methylheptane (ng/g)

160 140 120 100 80 60 40 20 0

180

7 6.4

150 5.8

120

Time (min)

5.2 90

4.6 60

(b)

5

4

Pressure (mPa)

3-ethyl-3-methylheptane (ng/g)

160 140 120 100 80 60 40 20 0

180

31 28

150

25 22

120

Time (min)

19 16

90

13 60

Temperature (°C)

10

(c) Fig.3 Response surface plot of 3-ethyl-3-methyl heptane as a function of (a) pressure and temperature, (b) pressure and time, and (c) temperature and time, keeping the constant variable at central point

6

Total extraction yield (%)

20

15

10

5

0

31

7 28

6.4

25

5.8

22 19

Temperature (°C)

5.2

16

Pressure (mPa)

4.6

13 10

4

(a)

Total extraction yield (%)

20

15

10

5

0

180

7 6.4

150 5.8

120

Time (min)

5.2 90

4.6 60

4

(b)

7

Pressure (mPa)

Total extraction yield (%)

20

15

10

5

0

180

31 28

150

25 22

120

Time (min)

19 16

90

13 60

Temperature (°C)

10

(c) Fig.4 Response surface plot of total extract yield as a function of (a) pressure and temperature, (b) pressure and time, and (c) temperature and time, keeping the constant variable at central point

8

High lights •

Effect of sub-critical CO2 extraction on volatile flavour compounds of ghee was studied.



Extraction pressure, temperature and time were studied as independent variables.



SC-CO2 pressure had the highest effect on recovery of the volatile flavours.



Maximum recovery of the volatile flavours was observed at 7 mPa, 10°C and 60 min.

1

Conflict of interest

The authors declared no conflict of interest.