Characterisation of Physicochemical Properties of Propionylated Corn Starch and Its Application as Stabiliser

Plenty of work has been done to expand the usefulness and versatility of starch as a property-enhancing material in food. Modifi cation of starch by means of esterifi cation provides ample opportunities to alter the physicochemical properties of macromolecules. Starch esters exhibit unique properties, namely lower gelatinization temperature, lower retrogradation and lower tendency to form gels (1) as well as enhanced hydrophobic properties (2,3) when compared to the native counterpart. These unique properties make starch esters suitable for use in refrigerated and frozen foods and as an emulsion stabiliser and encapsulating agent. Two common starch esters used in foods are acetylated starch and octenyl succinic (OS) starch. Acetylated starch is prepared with short-chain (C2) while octenyl succinylated starch with medium-chain (C12) fatt y acids. Owing to the diff erent fatt y acid chain lengths, acetylated and OS starch showed diff erent properties, especially emulsion stabilising and encapsulating properties. Comparatively, OS starch showed much bett er emulsion stabilising and encapsulating properties when compared to acetylated starch (4).


Introduction
Plenty of work has been done to expand the usefulness and versatility of starch as a property-enhancing material in food. Modifi cation of starch by means of esterifi cation provides ample opportunities to alter the physicochemical properties of macromolecules. Starch esters exhibit unique properties, namely lower gelatinization temperature, lower retrogradation and lower tendency to form gels (1) as well as enhanced hydrophobic properties (2,3) when compared to the native counterpart. These unique properties make starch esters suitable for use in refrigerated and frozen foods and as an emulsion stabiliser and encapsulating agent. Two common starch esters used in foods are acetylated starch and octenyl succinic (OS) starch. Acetylated starch is prepared with short-chain (C2) while octenyl succinylated starch with medium--chain (C12) fatt y acids. Owing to the diff erent fatt y acid chain lengths, acetylated and OS starch showed diff erent properties, especially emulsion stabilising and encapsulating properties. Comparatively, OS starch showed much bett er emulsion stabilising and encapsulating properties when compared to acetylated starch (4).
From a nutritional point of view, acylated starch appeared to be resistant (3,5). The nondigestible starch fraction was found to be able to reduce postprandial blood glucose and insulin concentration (6), which helps to ISSN 1330-9862 original scientifi c paper doi: 10.17113/ft b.53.03. 15.3907 maintain a healthy glycemic profi le. Annison et al. (5) reported that acylated starch survived intestinal passage and produced short-chain fatt y acids from bacterial fermentation in the large bowel. The reduction of colonic pH by short-chain fatt y acids was found to promote colonic health (7,8). In vitro or in vivo studies also showed that different short-chain fatt y acid profi les were produced when diff erent resistant starch ester was used. Ferguson and Jones (9) conducted an in vitro fermentation on OS starch and showed that it released the highest amount of butyric acid among other short-chain fatt y acids. The release of propionate was found to be interesting as it has been linked to the reduction of cholesterol in serum and liver in rats (10). Animal study revealed that propionate released in caecal digesta was 95.6 μmol from starch propionate, 41.2 μmol from starch acetate, 36.3 μmol from starch butyrate and 12.1 μmol from control starch (5).
Despite the health benefi ts of starch propionate, not much study, however, has been conducted on the characterisation of moderately substituted starch propionate. In this paper, we present a simple method to prepare starch propionate using propionic anhydride and sodium hydroxide as catalyst. Physicochemical properties of starch propionates were then characterised. The potential of starch propionates as a stabiliser was also evaluated.

Preparation of starch propionate using alkaline catalyst
A mass of 140 g of corn starch (0.864 mol of anhydroglucose unit) was stirred with 674.8 g (5.185 mol) of propionic anhydride in 2-litre Schott bott le for 5 min in a water bath maintained at 30 °C. The catalyst, 50 % (by mass per volume) of sodium hydroxide solution was added to the mixture using biuret at the rate of 0.25-0.33 mL/min. A mass of 0.15 g of sodium hydroxide solution, 50 % (by mass per volume), was added per each g of starch. The starch suspension was then transferred to a water bath controlled at 70 °C and stirred for 12 h using hotplate (Thermo Fisher Scientifi c, Waltham, MA, USA). A sample of 67.9 mL was drawn hourly over a reaction time of 12 h. The sample was added to water (3 times volume) to terminate the chemical reaction and centrifuged at 2000´g for 5 min. The supernatant was discarded and the sediment was washed with 4.5 volumes of water 16 times until reaching pH=5.0-5.5 (SevenEasy pH meter, Mett ler Toledo, Columbus, OH, USA) and conductivity below 2.0 μS/cm (sensION+ EC5 portable conductivity meter, Hach, Loveland, CO, USA). The starch propionate was dried in an oven at 40 °C for 48 h. Aft er drying, starch propionate was ground manually and sieved through a 250-μm diameter sieve. Starch propionate samples collected hourly over a period of 12 h were coded as SP1-SP12.

Determination of the degree of substitution
Degree of substitution (DS) was determined using the method reported by Bhandari and Singhal (11). A mass of 1 g of native corn starch or starch propionate was accurately weighed into 50 mL of 75 % (by volume) ethanol solution and stirred at 60 rpm on a hotplate for 30 min at 50 °C. The starch slurry was then cooled to room temperature (30 °C) and mixed with 40 mL of 0.5 M NaOH at 60 rpm using hotplate for 2 min. Aft er mixing, the starch slurry was placed in an orbital shaker at 40 rpm and 30 °C for 72 h. Phenolphtalein indicator was dropped to the starch slurry and excess of alkali was back titrated with 0.5 M of hydrochloric acid. Native corn starch was treated in the same manner to obtain blank value. The propionyl content and DS were calculated using the equations proposed by Xia et al. (12). The propionyl content (in %) was calculated using the following equation: where V is titration volume in mL, m is mass in g, c is concentration in mol/L, 57 is the formula mass of propionyl group and 10 -3 is a factor of conversion of mL to L.
The degree of substitution (DS) was defi ned as the average number of sites per α-d-glucose unit that possesses a substituent group, and it was calculated as follows: where 180 is the molecular mass of α-d-glucose, 0.9 is the stoichiometric constant for conversion of glucose to starch, 57 is the formula mass of propionyl group, and 1 is the atomic mass of hydrogen.

Fourier transform infrared spectroscopy
The Fourier transform infrared (FTIR) spectra were obtained with a BIO-RAD FTS 3000 IR Spectrum Scanner (BioRad Laboratories, Hercules, CA, USA).

Nuclear magnetic resonance
The nuclear magnetic resonance (NMR) spectra were recorded using an Ascend 500 spectrometer (Bruker, Ettlingen, Germany) operating at 500 MHz for 1 H NMR. 2D H-H NMR measurement was conducted to obtain the detailed information on the structure of starch propionate. Native corn starch and starch propionates of 7 mg each were dissolved in 0.7 mL of deuterated dimethyl sulphoxide (DMSO) with gentle stirring for 6 h at room temperature to obtain clear solutions.

Powder diff raction
X-ray patterns of native corn starch and starch propionate powders with different DS values were analyzed between 2θ=4 and 45° using a D8 Advance X-ray diff ractometer (Bruker, Berlin, Germany) with Cu-Kα radiation at a voltage of 40 kV and 40 mA.

Measurement of contact angle
Contact angle measurement was performed using a Theta Lite optical tensiometer (Att ension, Espoo, Finland). Att ension Theta soft ware was used to control the drop volume and measure the contact angle. The drop volume of water was fi xed at 30 μL. Sample preparation was carried out according to the method described by Thielemans et al. (13) with slight modifi cations. Starch pellet was prepared by compressing native corn starch or starch propionate under a pressure of 170.9 kg/(m·s 2 ) using a hydraulic bench top laboratory press (Carver, Wabash, IL, USA).

Swelling power and percentage of leaching
Swelling power and percentage of leaching were measured according to the method described by Bordenave et al. (14) with slight modifi cations. A mass of native corn starch and starch propionate of 0.1 g (m 1 ) was weighed separately and put into 1.5-mL microcentrifuge tube. A volume of 1.0 mL of distilled water was added and the starch suspension was vortexed with vortex mixer (Labnet International Inc., Edison, NJ, USA) followed by heating at 40 or 90 °C for 5 min.
Aft er that, the tubes were cooled to room temperature for 30 min and subsequently centrifuged at 14 000´g for 5 min (1-15 Microfuge, Sigma Laborzentrifugen GmbH, Osterode am Harz, Germany). Aft er centrifugation, the supernatant in the tubes was transferred into a moist dish and the swelled starch precipitate was weighed (m 2 ). The precipitate and the supernatant were dried separately in an oven at 100 °C for 20 h. The dried supernatant was weighed and labelled m 3 , while dried precipitate was weighed and labelled m 4 .
The swelling power was calculated as follows: Swelling power=m 2 /m 4 /3/ and the percentage of leaching was calculated with the fol lowing equation:

Preparation of oil-in-water emulsions
Olive oil (20 %, by mass) and water (80 %, by mass) were weighed separately. Starch propionate at 15 % (by mass) based on aqueous phase was weighed. Starch propionate was mixed with water and oil. The emulsion was homogenized using an Ultra Turrax T25 homogenizer (IKA Labortechnik, Staufen, Germany) at a speed of 4838.4´g for 5 min. A control emulsion was prepared using native corn starch.

Centrifugation stress study
The oil-in-water (O/W) emulsions were centrifuged at 4000´g for 30 min (Heraeus Multifuge X1, Thermo Fisher Scientifi c). The emulsion stability (ES) parameter pro-posed by Taherian et al. (15) was used to evaluate the stability of the emulsion. The ES was calculated as follows: where h t is the total height of emulsion and h s is the height of sediment layer. Stable O/W emulsion has ES value of 100 %.

Measurement of emulsion viscosity
The viscosity of the emulsion was measured using an SV-10 Sine-wave Vibro Viscometer (A&D, Tokyo, Japan). The emulsion sample was placed in a plastic sample cup and viscosity was measured at 30 °C.

Statistical analysis
The results are presented as mean values±standard deviation (S.D.). The results were analysed using one--way analysis of variance (SPSS soft ware, v. 13.0 for Windows; SPSS Inc., Chicago, IL, USA). When there was a statistically signifi cant diff erence (p<0.05), Duncan's post--hoc test was performed.

Results and Discussion
Eff ect of reaction time on the degree of substitution of starch propionates Fig. 1 shows that when corn starch was treated with propionic anhydride for a period of 12 h, the degree of substitution (DS) increased steadily with reaction time. The range of DS values of starch propionates was between 0.47 and 0.94. This trend may be att ributed to an increase in diff usion of acylating reagent into and adsorption onto starch molecules with prolonged reaction time (16). Besides, the progressive increase in DS with time could be partly att ributed to alkaline treatment. In the presence of alkali, protons of the hydroxyl groups may dissociate, leaving negative charges on starch molecules. The negatively charged starch molecules may then repel each other causing dissociation of the double helices of amylopectin (17), which leads to the exposure of more reactive sites on the macromolecule for the acylating reagent to react upon.  temperature of 123 °C, reaction time of 60 min, and starch/ acetic anhydride mass ratio of 1:4. In contrast, the present study produced starch propionate with DS values of 0.47-0.94, which are considered moderate, at a condition of 0.15 g of NaOH per g of starch, reaction temperature of 70 °C, reaction time of 1-12 h, and starch/propionic anhydride mass ratio of 1:4.82. Diff erences in the substitution degree of acylated starches could be due to diff erences in reaction conditions such as temperature and type of anhydride. Increasing temperature produced starch with higher DS, indicating that higher temperature favoured acceleration of the forward reaction (19,20). In the present study, starch propionate with lower DS was obtained because a lower reaction temperature was used. Other authors showed that acyl anhydride with longer chain lengths of fatt y acids tends to form starch ester with lower DS (20,21). The reaction occurred at slower rate due to the steric eff ects of larger molecular size of reagent and a lower diff usion rate in the starch matrix.
The common practice to prepare propionylated starch is to use propionic anhydride. Various catalysts have been used such as 1-methylimidazole (5), sodium hydroxide (18), formamide (22) and pyridine (23). The selection of catalyst was based on its suitability and safety in the preparation of starch propionate as food, cosmetic and pharmaceutical additives. Sodium hydroxide was chosen as catalyst since it is water soluble and can be easily washed off aft er the esterifi cation process.

Fourier transform infrared spectroscopy of native starch and starch propionates
FTIR spectra of native corn starch and starch propionates SP1, SP6 and SP12 are presented in Fig. 2. Absorption bands at 3423.51, 2928.17, 1650.13 and 1239.09 corresponded to hydrogen-bonded hydroxyl, CH 2 symmetrical stretching vibrations, scissoring of two OH bonds of absorbed water molecules and C-O-H deformation, respectively (12,24,25).
All starch propionates gave a new peak at 1743.48 cm -1 , indicating stretching of ester carbonyl C=O (12,18,23) and successful propionylation. Two new bands were also seen at 2878.26 and 2986.96 cm -1 in the spectra indicating methyl and methylene C-H stretching of propionyl groups in esterifi ed starch (24,26,27).

Nuclear magnetic resonance spectroscopy of native starch and starch propionate
The 1 H NMR spectra of native starch and starch propionate SP12 (DS=0.94) are shown in Fig. 3. For native starch, 1 H-chemical shift at 5.4963 ppm was assigned to OH-3, at 5.3985 ppm to OH-2, at 5.1051 ppm to H-1, at 4.5731 ppm to OH-6, at 3.6518 ppm to H-3 and at 3.5843 ppm to H-5 (2,28,29). Aft er propionylation, all starch samples exhibited characteristic peaks similar to native starch. The new 1 H-chemical shift s were observed at 2.3470 and 1.0324 ppm, corresponding to CH 2 and CH 3 protons, respectively (30). These observations indicated a successful propionylation. This result was also corroborated by the two new bands evident at 2878.26 and 2986.96 cm -1 in the FTIR spectra, indicating methyl and methylene C-H stretching of propionyl groups in starch propionate. This shows that propionylation possibly occurred at OH-2, OH-3 and OH-6 of the anhydroglucose units.    (2). The crystallinity of starch was formed by the double helices of amylopectin chains (31,32). According to Xu et al. (18), the intra-and intermolecular hydrogen bonds account for highly ordered crystalline structure in starch granule. The peak intensities of SP1 were found to be lower than the characteristic peaks of the native corn starch. Only part of the hydroxyl groups was replaced with propionyl group. The remaining hydroxyl groups may still form hydrogen bonds, resulting in the decrease of diffraction peak intensities in SP1 (18). Further propionylation caused progressive changes in the X-ray diffraction patterns as shown by the formation of broad peaks in SP6 and SP12. These changes indicated the disruption of starch granule crystallinity through the replacement of the hydroxyl groups with propionyl groups on the starch molecules upon esterification (2,18). Table 1 shows contact angle values of various starch propionates. It is evident that the contact angle values progressively increased with propionylation time. The gradual increase in contact angle from samples SP1-SP11 was due to an increase in the hydrophobicity which was caused by propionylation of the starch. It can be deduced that in the process of propionylation, hydrophilic hydroxyls were substituted with more hydrophobic ester groups. The same observation was reported by Chi et al. (2). Contact angle value of native corn starch was not measured because when water was dropped onto the surface of the pellet, it quickly disintegrated. It was concluded that native corn starch is too hydrophilic for the measurement of contact angle. Table 2 shows swelling power values of native corn starch and various starch propionates. Native corn starch exhibited swelling power of 2.3 and 10.9 g/g at incubation temperatures of 40 and 90 °C, respectively. At 40 °C, 0.4 % of starch polymers leached out of native corn starch. Leaching increased to 1.6 % when native corn starch was incubated at 90 °C. As shown by Bordenave et al. (14), native corn starch showed a swelling power of 7.84 g/g and leaching of 1.72 % when the experiment was conducted at 23 °C for 30 min. Native corn starch has been reported to exhibit swelling power of about 12.81 g/g and leaching >5 % when the starch was incubated at 90 °C for 1 h (33). In the present study, native corn starch had much lower swelling power and percentage of leaching, which was probably due to the short incubation period of 5 min.   Aft er propionylation, corn starch showed a higher swelling power compared to native corn starch at incubation temperature of 40 °C. According to some authors, acylation could weaken the intermolecular hydrogen bonding within starch granule, making it readily swollen when heated at a lower temperature (34). It was observed that swelling power and percentage of leaching of starch propionates increased from SP1 (DS=0.47) to SP5 (DS=0.73). The weak intermolecular hydrogen bonding present within starch propionate was substantiated by the X-ray patt erns seen in Fig. 4. On the other hand, from SP5 (DS=0.73) to SP11 (DS=0.92) swelling power decreased but percentage of leaching reached plateau because of the enhanced hydrophobicity of starch propionates when more and more hydroxyl groups were replaced by propionyl groups.

Eff ect of propionylation on the swelling and leaching behaviour of starch
At 90 °C, a lower swelling power was evident in starch propionates when compared with native corn starch. Native corn starch with intact granules can swell to its maximum when heated in water. Starch propionates having hydroxyl groups replaced with propionyl groups became more hydrophobic than native corn starch, thus resulting in a lower degree of swelling. Swelling power of starch propionates decreased signifi cantly from SP1 (DS=0.47) to SP9 (DS=0.86), but there was no signifi cant diff erence between SP9 (DS=0.86) and SP11 (DS=0.92). The weakened granular structure of starch propionates contributed to a signifi cantly higher percentage of leaching than the native corn starch. The increase in percentage of leaching in starch propionates from SP1 (DS=0.47) to SP11 (DS=0.92) was not signifi cantly diff erent. The fi ndings showed that percentage of leaching was independent of DS and not signifi cantly aff ected by the degree of propionylation.
The ability of starch propionate to swell and thicken at low and high temperatures shows that it has the potential for use as a cold swelling starch and thickener.

Physical stability of emulsions with starch propionates
In a preliminary study, emulsions were prepared with 10, 15 and 20 % (by mass) of starch propionates (SP1-SP12). Emulsions prepared with 10 % of starch propionates separated immediately aft er centrifugation. At 15 % of starch propionates, O/W emulsions prepared with SP3-SP9 (DS=0.64-0.86) were stable while others separated. At 20 % of starch propionates, emulsions were stable. Hence, 15 % of starch propionate was selected for the study as this fraction could discern the eff ect of DS of starch propionates on the stability of O/W emulsions. Stability of emulsions prepared with 15 % of starch propionate at diff erent DS is shown in Fig. 5. Emulsions were best stabilised with starch propionates SP3-SP9 (DS=0.64-0.86). A control emulsion was prepared using native corn starch. When emulsion was prepared with native corn starch granules at 15 %, it separated quickly aft er homogenization. The good emulsion stabilising properties of starch propionate indicated that it is bett er compared to starch acetate. Starch acetate is not known to be a good emulsion stabiliser (4). Thus, the starch propionate produced in the present study could be a promising emulsion stabiliser.

Viscosity of emulsions with starch propionates
The viscosity of emulsions prepared with starch propionates at diff erent DS levels is shown in Table 3. There is an increase in the viscosity of the emulsion from SP1 (DS=0.47) to SP5 (DS=0.73), followed by a decrease aft er SP5. These changes could be due to the swelling of the starch propionate granules as shown in Table 2. Crosbie (35) reported that viscosity of starch was correlated with its swelling power. According to McClements (36), in order to be an effective emulsifi er, biopolymers must rapidly adsorb to the surface of the oil droplets during homogenization and provide a membrane to prevent aggregation of oil droplets. All biopolymers are capable of providing physical barrier via thickening ability to prevent droplets from getting suffi ciently close together to coalesce. Hence, in the present study, the mechanism of emulsion stabilisation was due to viscosity enhancement caused by starch propionate, which prevented the oil droplets from coming close together.

Conclusions
In conclusion, starch propionates with degree of substitution (DS) values of 0.47-0.94 were produced throughout a 12-hour propionylation process. Contact angle values were found to increase proportionately with an increase in DS values, indicating increasing hydrophobicity of the starch propionate. Samples SP1-SP11 showed higher swelling power than the native corn starch at 40 °C, indicating cold swelling and thickening properties. Swelling power was found to decrease with increasing DS. Stable oil-in-water (O/W) emulsions were produced at DS values of 0.64-0.86 and viscosities between 1236.7 and 3330.0 mPa·s. This shows that starch propionate with the above characteristics has a great potential to be used as a cold swelling starch, thickener as well as O/W emulsion stabiliser in food, pharmaceutical and cosmetic industries.