Recovery of dairy flavor compounds by pervaporation using poly(ether block amide)

Boya Zhang, Xianshe Feng, IPR Symposium, University of Waterloo, ON N2L 3G1, Canada During dairy products processing, many flavor compounds are lost due to their

high volatility. Even a small amount of loss of flavor compounds may significantly affect products’ sensory quality. Since artificial flavors cannot satisfy consumers’ interest anymore, it is necessary to recover these flavor compounds before commercial processing from raw dairy recourses and then add them back to the final products.

Traditional flavor fractionation processes are based on extraction, distillation, partial condensation, and gas stripping [1,2]. Pervaporation, as a novel separation technique, was introduced as an alternative to the current separation technologies. Its major advantages include: the moderate operating temperature used in pervaporation can reduce the energy consumption; the mild temperature can also protect product integrity from thermal degradation, especially for heat sensitive compounds. In a pervaporation process, a feed liquid mixture contacts one side of a dense membrane, while the preferential permeate diffuses through the membrane and evaporates on the other side by using a vacuum pump. The non-permeated components in the retentate are usually recycled into the feed stream for further recovery.

Polyether block amide (PEBA) 2533 is a copolymer comprising 80 wt.% poly(tetramethylene oxide) and 20 wt.% nylon 12. Its general formula is shown in Figure 1 [3], where PA and PE represent polyamide and polyether segments, respectively. PEBA 2533 has a good selectivity of flavor compounds due to the strong affinity between the polyether segments and the flavor compounds.

Figure 1 Structure of PEBA

The objective of this study is to investigate the pervaporation performance of PEBA 2533 membrane on separating flavor compounds from dairy model solutions. The effects of independent variables like feed concentration and operating temperature, on organic flux and enrichment factor were studied. The coupling effect was also investigated by comparing the pervaporation results obtained in binary and multicomponent systems.

1 Experimental Eight flavor compounds, which represent six categories of dairy flavor

compounds (esters, ketones, aldehydes, acids, sulfur compounds and aromatic compounds), were recovered by pervaporation. The flavor compounds included ethyl hexanoate (EH), ethyl butanoate (EB), 2-heptanone (2-Hep), diacetyl, methyl sulfone (MS), indole, nonanal and hexanoic acid (HA). For single flavor compound

IPR 20

14

pervaporation, each flavor compound was diluted by deionized water to make a dilute flavor–water binary feed solution. For multicomponent pervaporation, the feed solution was prepared by mixing the eight organics with deionized water.

Pervaporation experiments with binary and multicomponent feed solutions were carried out at various flavor concentrations and temperatures. Figure 2 shows the schematic diagram of the experimental set-up for pervaporation separation. The PEBA membrane was mounted into the permeation cell. The feed solution was continuously circulated through the membrane cell and back into the 1000 mL feed tank. Vacuum was applied on the permeate side to generate pressure difference, which providing the driving force. The permeate sample was condensed in a cold trap immersed in liquid nitrogen (-196 °C), and analyzed by gas chromatography.

Figure 2 Schematic diagram of the pervaporation setup

2 Results and discussion 2.1 Influence of feed concentration

Pervaporation separation of flavor compounds from binary feed solutions with PEBA 2533 membrane was investigated at various feed concentrations and temperatures. Figures 3 (a) and (b) show the effect of the feed flavor concentration on the flavor permeation flux and enrichment factor, respectively.

All of the flavor compounds flux increases as the feed concentrations increase. That is because with the feed flavor concentration increasing, the partial pressure difference of the flavor compound between the two sides of the membrane increase, which leads to a higher driving force of the pervaporation process. The enrichment factors change little with the feed concentration, especially beyond 1000 ppm.

Moreover, the flux of the two esters (ethyl hexanoate and ethyl butanoate), one of the ketones (2-heptanone) and the aldehyde (nonanal) increase more significantly than other compounds when feed concentration increase; their enrichment factors are higher as well.

IPR 20

14

(a)

(b)

Figure 3 Effect of feed flavor concentration on (a) the flavor flux and (b) enrichment factor for binary flavor-water solutions using PEBA 2533 membrane. 2.2 Influence of operating temperature

Operating temperature has a significant influence on flavor compound flux. The flavor flux increases obviously as temperature increases. This is because an increase in temperature generally increases both the diffusivity of the permeant and the thermal motion of the membrane polymer chains, which facilitates the movement of permeant. 2.3 Pervaporation with multicomponent feed solutions

The multicomponent feed solutions were prepared by mixing eight flavors and water in different mass ratios. Figure 4 shows the comparison between the flavors

0 20 40 60 80

100 120 140 160 180 200

0 500 1000 1500 2000 2500 3000

Flav

or fl

ux, g

/(m2.

h)

Feed concentration, ppm

Diacetyl 2-Hep EB EH HA MS Indole Nonanal

0

50

100

150

200

250

300

350

400

450

0 500 1000 1500 2000 2500 3000

Enric

hmen

t fac

tor

Feed concentration, ppm

IPR 20

14

enrichment factors obtained from binary and multicomponent systems (the organics concentration: nonanal 15 ppm, indole 100 ppm, EH 100 ppm, other flavors 500 ppm; five operating temperatures were tested, the results shown are average value of the enrichment factors under the five temperatures). It clearly shows that the enrichment factors of most flavors are lower in multicomponent system than in binary system, which indicates coupling effects exist under such flavor concentrations. This happens probably due to the competition and interaction between the flavor compounds when they penetrating the membrane.

Figure 4. The flavors enrichment factors obtained from binary and multicomponent systems (the organics concentration: nonanal 15 ppm, indole 100 ppm, EH 100 ppm, other flavors 500ppm; the results shown are average value of the enrichment factors under the five temperatures). Reference [1] H.E. Karlsson, G. Triigiirdh, Aroma Recovery During Beverage Processing, 34

(1998) 159–178. [2] A. Baudot, M. Marin, Pervaporation of aroma compounds: Comparison of

membrane performances with vapour-liquid equilibria and engineering aspects of process improvement, Food Bioprod. Process. 75 (1997) 117–142.

[3] M.K. Mandal, P.K. Bhattacharya, Poly(ether-block-amide) membrane for pervaporative separation of pyridine present in low concentration in aqueous solution, J. Memb. Sci. 286 (2006) 115–124.

0.1

1

10

100

1000 Indole

HA

EH

Nonanal

EB

2-Hep

Diacetyl

MS

Binary Multi-component

Boya ZhangSupervisor: Prof. Xianshe Feng

May 21, 2014

IPR symposium

DAIRY FLAVOR RECOVERY BY PERVAPORATION USING

POLY(ETHER BLOCK AMIDE) MEMBRANE

IPR Presentation – BOYA ZHANG| 2

FlavorRecovery

FlavorAdded Back

Traditional methods for recovering:

Food, medical and cosmetic industry. 

Flavoring ingredients

Distillation, partial condensation & gas stripping 

Problems: product contamination and degradation，high energy consumption 

Flavor Loss or degradation

BACKGROUND

Processing

IPR Presentation – BOYA ZHANG| 3

BACKGROUND

Pervaporation– an alternative recovery method

Why pervaporation?

High selectivity, low energy consumption, mild operating temperatures and no damage to heat‐sensitive flavors 

FlavorRecovery

FlavorAdded Back

Food, medical and cosmetic industry. 

Flavoring ingredients

Flavor Loss or degradation

Processing

IPR Presentation – BOYA ZHANG| 4

OBJECTIVE

To investigate pervaporation as a mean of recovering flavorcompounds from dairy model solutions, using an organophilicmembrane material: poly(ether block amide) (PEBA).

• To determine how the operating conditions affected thepervaporation of each flavor compound.

• To compare the pervaporation behavior of single flavor componentand multiple flavor components, in order to determine whetherthere is any “coupling effect” between permeating species.

IPR Presentation – BOYA ZHANG| 5

Flavor pervaporation process

Organophilic Membrane

Vacuum

Esters Ethyl butanoate (EB)

Ethyl haxanoate (EH)

Ketones 2-heptanone (2-Hep)

Diacetyl

Aldehyde Nonanal

Acid Hexanoic acid (HA)

Sulfur compound Methyl sulfone (MS)

Aromatic compound Indole

Flavor compounds

WaterFeed Permeate

20% 80%

PEBA 2533

1. M.K. Mandal, P.K. Bhattacharya, Poly(ether-block-amide) membrane for pervaporativeseparation of pyridine present in low concentration in aqueous solution, J. Memb. Sci. 286 (2006)115–124.

IPR Presentation – BOYA ZHANG| 1

Evaluation of pervaporation performance

• Flux (Ji)

• Enrichment factor (βi)

Permeate flow rate of component i per unit membrane area.

i CPi

CFi

Ratio of component i’s concentrations in permeate and in feed.

Ji Ni

A

Pervaporative enrichment process is effective only when βi > 1.

IPR 20

14

7 16 16 20

153

210 247

400

0

50

100

150

200

250

300

350

400

450

En

rich

men

tfa

cto

r

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Fla

vor

flu

x,g

/(m

2 .h

)

Binary feed system: single flavor + waterFlavor concentration: 50 ppm; Temperature: 36

EF=1

• All the flavorcompounds can besuccessfullyconcentrated bypervaporation withthe PEBA membrane.

• Flavor compounds withhigher hydrophobicitiesare enriched moresignificantly with thePEBA membrane.

IPR Presentation – BOYA ZHANG| 7

BINARY SYSTEMPERVAPORATION

0

50

100

150

200

250

300

350

400

450

En

rich

men

tfa

cto

r

PEBA 2533

PDMS

Comparison with published PDMS pervaporation studies

IPR Presentation – BOYA ZHANG| 8

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Fla

vor

flu

x,(g

/m2 .

h) PEBA 2533

PDMS

In general, PEBAperformed betterthan PDMS for flavorenrichment.

BINARY SYSTEMPERVAPORATION

1. C.C. Pereira, J.R.M. Rufino, A.C. Habert, R.Nobrega, L.M.C. Cabral, C.P. Borges, Aromacompounds recovery of tropical fruit juice bypervaporation: membrane material selection andprocess evaluation, J. Food Eng. 66 (2005) 77–87.

2. A. Overington, M. Wong, J. Harrison, L. Ferreira,Concentration of dairy flavour compounds usingpervaporation, Int. Dairy J. 18 (2008) 835–848.

OPERATING CONDITIONS

Effect of feed concentrationFlavor flux

With an enhancement of feed concentration, MS and indole fluxes almostmaintain the same; the flux of the other six compounds increase .

Temperature: 36

IPR Presentation – BOYA ZHANG| 9

0

5

10

15

20

25

0 200 400 600

Fla

vor

flu

x, g

/(m

2 .h

)

Feed flavor concentration, ppm

EH

0.0

0.5

1.0

1.5

2.0

0 20 40 60

Fla

vor

flu

x, g

/(m

2 .h

)

Feed flavor concentration, ppm

Nonanal

0

1

2

3

4

5

6

0 1000 2000 3000

Fla

vor

flu

x, g

/(m

2 .h

)

Feed flavor concentration, ppm

DiacetylHAMSIndole

020406080

100120140160180200

0 1000 2000 3000

Fla

vor

flu

x, g

/(m

2 .h

)

Feed flavor concentration, ppm

2-Hep

EB• When the concentration of

feed flavor is increasing, βEB

and βEH increase as well,while the other flavors’enrichment factors decrease.

• Feed concentration affect EHand nonanal’s enrichmentfactors more significantly.

0

50

100

150

200

250

300

0 500 1000 1500 2000 2500 3000

Enr

ichm

entf

acto

r

050

100150200250300350400450

0 500 1000 1500 2000 2500 3000

Enr

ichm

entf

acto

r

Feed flavor concentration, ppm

DiacetylHAMSIndole2-Hep

Effect of feed concentrationOPERATING CONDITIONS

Temperature: 36

IPR Presentation – BOYA ZHANG| 10

Enrichment factor

EH

EB

Nonanal

Effect of temperature

0

1

2

3

4

5

6

7

20 30 40 50 60 70

Fla

vor

flu

x, g

/(m

2.h

)

Temperature,

DiacetylHAMSIndole2-HepEBEHNonanal

0

200

400

600

800

1000

1200

20 30 40 50 60 70

Wa

ter

flu

x,

g/(

m2

. h)

Temperature,

Flavor flux Water flux

OPERATING CONDITIONS

Flavor concentration: 50 ppm

Flavor and water fluxes increase with increasing temperature.

IPR Presentation – BOYA ZHANG| 11

0

10

20

30

40

50

60

70

80

90

Ac

tiv

ati

on

en

erg

y, J

/mo

l Flavor

Water

Activation energyOPERATING CONDITIONS

J J0 expEp

RT

Arrhenius-type equation:

IPR Presentation – BOYA ZHANG| 12

Where Ep is the activation energy of permeation, J is permeation flux, J0 is the pre-exponential factor,R is gas constant,

and T is the absolute temperature.IP

R 2014

0

10

20

30

40

50

60

70

80

90

Ac

tiv

ati

on

en

erg

y, J

/mo

l Flavor

Water

1

10

100

1000

20 40 60 80

En

ric

hm

en

tfa

cto

r

Temperature,

DiacetylHAMS2-HepEBEHNonanal

Effect of temperatureOPERATING CONDITIONS

Flavor conc.: 50 ppm

Indole

IPR Presentation – BOYA ZHANG| 13

If the activation energy of one flavor is higher than water, theenrichment factor of the flavor will increase as temperature goeshigher, vice versa.

Enrichment factor

Comparison: multicomponent and binary systempervaporation

Enrichment factor

0.01

0.10

1.00

10.00Diacetyl

MS

EB

EH

2-Hep

Nonanal

Indole

HA

BinaryMulti-component

Flavor flux

0.1

1

10

100

1000Diacetyl

MS

EB

EH

2-Hep

Nonanal

Indole

HA

MULTICOMPONENTPERVAPORATIONMulticomponent feed system: eight flavors + water Flavor conc.: 50 ppm; Temp: 36

Coupling effect-- is mainly attributed to the competition between flavor compounds when they penetrate the membrane. 

IPR Presentation – BOYA ZHANG| 14

CONCLUSIONS

IPR Presentation – BOYA ZHANG| 15

• PEBA 2533 has good pervaporation performance on concentrating dairyflavors, especially hydrophobic ones.

• The increase in feed concentration has positive effects to the flavorfluxes; however, its impact on enrichment factor varies from compound tocompound, depending on the nature of the flavor molecule.

• Temperature affects the enrichment of dairy flavors as well. Thetemperature dependence of flavor and water fluxes obeys Arrhenius-typerelationship . Flavors and water’s activation energy determines thechanging trend of enrichment factor.

• In multicomponent pervaporation, the “coupling effect” betweenpermeating species decreases the flavors enrichment.

ACKNOWLEDGEMENT

IPR Presentation – BOYA ZHANG| 16

• Financial support

• Prof. Xianshe Feng

• Members in Feng’s group: Dihua, Yifeng, Shuixiu, Xincheng, Kai, Yijie, Guan.

IPR Presentation – BOYA ZHANG| 17

IPR 20

14