Tropical Tuber Crops

3. Basic studies on the tuber starches.

   Once the extraction of starches in good quantity and quality from most of the tubers was achieved, a detailed study of the physicochemical and functional properties of the different tuber starches was carried out and the results are outlined below

   1. Other components in starch

      Even though starch extracted from the tubers appear white and pure, still it harbors many other biochemical components like moisture, fiber, lipids, sugars, minerals which influence the starch properties.

      The moisture content varies from 6-16% depending on the process used for drying the starch. The fiber content can show considerable variability depending on a number of factors like the sieve used for removal of the fibrous material, varietals variation and age of the crop. These factors are all the more important in cassava and sweet potato, where the fiber content increases with the maturity. The presence of the fiber modifies most of the rheological and functional properties of the starches. The effect of fiber on starch properties is also clear from the fact that cassava flour (containing 2-3% fiber) had different swelling and viscosity properties compared to the isolated starch (having 0.1-0.15% fiber) and neither defatting nor ethanol extraction brought about any major change in the properties of the starch or flour. The total dietary fiber in cassava flour was found to vary between 4.7 to 5.5%.

      For Coleus starch, 0.4% fiber content was observed. It has been found that the extracted starch from cassava tubers subjected to fermentation with inoculum provided culture contained considerable quantity of fiber.

      The tuber starches contain much lower quantities of lipids in them and so the effect of lipids on starch properties is not so pronounced compared to cereal starches. The lipid content in different cultivars of cassava was found to vary from 0.11 to 0.22% in starch and 0.27-0.45% in flour of five varieties. Lipids and surfactants form complexes with amylose chains and the free chains of amylopectin and thus influence starch properties. Based on this principle, it has been possible to improve the viscosity stability of cassava starch by treatment with surfactants.

      Another important component invariably present in starch is phosphorus, since P is involved in starch synthesis. Studies on the P content in cassava starch during growth period did not reveal any major variation with age of the crop for six cultivars over the growth period of 2- 18 months. The P content in different Colocasia cultivars varied from 0.006 to 0.013%. Studies on the P content in six accessions of D. rotundata showed only very minor variability (0.011-0.015%) . Examination of starch from three cultivars of Canna edulis from CTCRI revealed that the P content ranged from 0.05 to 0.08% which is even higher than that found in potato starch. Similarly it has been found that Curcuma starch also harbours high quantity of P (0.045%).

      The high phosphorus content can impart high viscosity to starch and also improve the gel strength. The high viscosity is attributed to the repulsion between the ionic phosphate groups. This was evident in the case of Canna starch which has high P content contributing to high viscosity and good gel strength. Biscuits made from the starch were having good texture and crispness. These starches can also be very useful in food applications requiring good gel strength like jellies etc.

   2. Color and appearance

      Colour is an important criterion for starch quality, especially for use in sago and textile industries. Use of organic acids was helpful in improving colour of sweet potato starch. The starch extracted from Colocasia and Dioscorea tubers by normal process always has an off-colour, which is difficult to remove. However use of ammonia during extraction was found to improve the colour of the starch from aroids especially Colocasia. Though tubers of some varieties of Amorphophallus possess yellow colour in their flesh, the resultant starch is pure white in colour. If the extraction is proper, the colour remains white even after storage for six months.

   3. Granule size and shape

      Cassava starch granules are mostly round with a flat surface on one side containing a conical pit, which extends to a well defined eccentric hilum. Some granules appear to be compound. Under polarised light, a well defined cross is observed. Sweet potato starch is polygonal or almost round in shape and has a centric distinct hilum. Polarisation crosses are less distinct compared to cassava starch. Yam starches have a large variability in shape, round, triangular, oval and elliptical. Some of the elliptical granules are found to possess truncated ends. Colocasia starch granules are mostly round or oval and the hilum is observed only with difficulty. Amorphophallus starch granules are mostly polygonal or round in shape and have faint hilum. We obtained oval and polyhedral shape for Canna edulis starch granules. Curcuma starch granules are elliptical in shape and appear similar to D. rotundata and Canna starch granules (Tab. 9).

      The size of the granules is also quite variable among the tuber starches (Tab. 9). Cassava starch was found to have a size range of 5-40mm. Studies on the starch granule size variation with age of the crop starting from 2^nd month till 18^th month for six varieties revealed that size increase was observed upto 6^th month, but then remained steady . Colocasia granules are much smaller (range of 1-10 mm) and are among the smallest of starches observed in the plant kingdom. The small granule size makes the starch useful in various applications eg. as a filler in biodegradable plastics, in toilet formulations, aerosol etc . Unlike other tuber crops, which do not exhibit any significant variability in size among varieties, Colocasia starch was found to exhibit varietal difference. Studies on 10 varieties revealed a significant difference in average granule size (Fig 7, Tab. 10). Variety C-9 was found to have the highest value of 5.19 mm and C-46 the lowest (2.96 mm). The granules extracted from corms and cormels of four cultivars of Colocasia were compared and it was found that there was very little difference between them.

      Even the distribution of the granule sizes showed only minor difference between the corms and cormels. It was also found that though there was variation among the cultivars, there was no significant variability during the growth period.

      Such varietal difference was not observed in Dioscorea alata, D. rotundata and D. esculenta starches. These starches showed an increase in granule size upto 5 months and thereafter remained steady. One noticeable feature is that D. esculenta granules are very small, while the D. alata granules are very large. The granules of D. esculenta possessed an average size of 2-15 mm, close to Colocasia starch. The small granule size of D. esculenta may be useful in applications similar to those of Colocasia and cereal starches. Starch granules of D. alata and D.rotundata are much bigger, the range is between 6-100 mm for D. alata (average 35 mm) and 10-70 mm for D.rotundata (average 33 mm) starches. No significant varietal difference in granule size was observed for the three yam species viz. D alata, D. esculenta and D. rotundata. There was only minor variation in starch granular size among 10 accessions of Amorphophallus studied.(Tab. 11). Xanthosoma starch granules range from 10-30 mm in size with an average value of 17 mm, and no difference among accessions was observed. Coleus starch granules had a a size range of 5-20 mm.

      Pachyrrhizus starch was found to have a size range of 7-40 mm, and very little difference among varieties. However the largest granule size was observed for Canna edulis starch. The average granule size was found to be over 35 mm for three varieties of Canna. The granule size was in the range of 16-58 mm for Curcuma zedoaria and 14-46 mm for C. malabarica starches.

   4. Spectral features

      The infrared spectrum of starch of different varieties of cassava was found to be similar with peaks at 3600-3200 (broad), 2800 (medium), 1660 (weak), 1480-1250 (medium) and a number of peaks between 1150 and 710 cm ^–1.

      FT-IR of the different tuber starches revealed only very minor differences, in spite of their different crystallinity and granule sizes. The Raman spectra of the tuber starches indicated distinct differences in the peak pattern in the region 800-200 cm^-1 (Fig 8).

      THE CP- MAS ¹³C NMR of the different starches showed typical pattern for the starches and three main peaks were observed. The first peak was at 101-102 ppm corresponding to C₁ and appeared as a singlet or doublet depending on the source of the starch. The next peak appeared at 75-80 ppm corresponding to C₂,_3,5 and was a singlet. The final peak was at 64 and a singlet. The peaks were observed only if the starch granules had 8-10% moisture in them. There was clear correlation between the XRD pattern and the NMR peak pattern. Whereas the starches having pattern ‘A’ showed a doublet for C₁ peak in NMR, those with ‘B’ pattern had clear singlet C₁ peak (Tab. 12; Fig. 9). Thus the structural difference between the two types of starch is evident.

   5. X-ray diffraction pattern

      Starch has a definite crystalline nature and the crystallinity has been assigned to the well-ordered structure of the amylopectin molecules inside the granules. Cassava starch has been found to posses ‘A’ pattern with three major peaks at 2Ø=15.3,17.1 and 23.5°. As already mentioned there was not much variation in absolute crystaniilty among the varieties. The flour also possessed similar pattern and the starchy flour obtained from tubers subjected to inoculum provided fermentation (Fig 2). Sweet potato starch was found to possess ‘A’ pattern. Colocasia, Xanthosoma , Pachyrrhizus, Arrowroot and Amorphophallus starches also possessed ‘A’ pattern. But the edible Dioscorea starches (viz. D. alata, D.esculenta and D. rotundata) possessed ‘B’ patterns similar to potato. It was found that the XRD pattern of extracted starch is same throughout the growth period of D. rotundata (Fig. 12).

      Starch of Canna edulis and Curcuma sp. exhibited ‘B’ XRD pattern. A detailed study of the XRD parameters of the starch extracted from Amorphophallus and Xanthosoma tubers subjected to pre-treatment using different chemicals was carried out. The ‘ d’ spacing, angle intensity and peak intensity were found to be similar for control (water) and chemically pretreated samples of Amorphophallus samples. However shift occurred for the peaks indicating partial change in the crystalline phase. For Xanthosoma starch, higher concentration brought about more significant changes especially with potassium metabisulphite. Heat moisture treatment did not change the XRD pattern of D. rotundata starch.

   6. Molecular weight

      The molecular weight of cassava starch showed only minor differences existed among the different varieties (Tab. 2). D. rotundata starch showed only minor changes in the molecular weight over the growth period . Colocasia esculenta and Amorphophallus paeoniifolius starch from different varieties had almost the same range of reducing values, showing that the tuber starches have nearly equal molecular weights. Studies on the yam and aroid starches extracted from tubers subjected to treatment with different chemicals also indicated only very minor differences in the reducing values among the different treatments. For coleus starch, the reducing value was 1.71, indicating the same range as for other tuber starches while it was between 1.7 to 2.1 for Curcuma starch.

   7. Amylose content

      The linear component of starch, viz., amylose, imparts definite characteristics to starch and therefore amylose content is an important criterion in determining the starch properties. Amylose content is found to vary considerably among different starches and even genetic modifications have been carried out to obtain starch of amylose contents varying from 0- >75% amylose. It was found that the Blue Values corresponding to total amylose varied from 0.50 to 0.55 for seven cassava varieties, indicating only very minor variation among the varieties [Tab.2]. When the amylose content of six varieties of cassava was compared during growth period, there were only insignificant differences in the amylose content . Examination of the effect of surfactants on the amylose content in cassava starch revealed that though the surfactants reduced the Blue Values, there was no correlation between concentration of surfactant and reduction in Blue Values. Maximum reduction was obtained with cetyl trimethyl ammonium bromide while lowest reduction was with potassium stearate. Sodium lauryl sulphate and cetyl trimethyl ammonium bromide, having bulky hydrophilic groups, might be blocking the entry of the iodide into the amylose helix.

      Colocasia esculenta starch showed wide range in the amylose content and noticeable relationship between the amylose content and granule size was noticed (Tab.10). The variety C-9, which had the largest granule size also possessed the highest amylose content. Amylose content in six D.rotundata varieties ranged from 21 to 24.6% while that in D. esculenta and D.alata in the range 20-26%. Thus the amylose content also varies considerably among the Dioscorea starches. The starch of D. esculenta, D. alata and D. rotundata was also examined for total amylose content in relation to age of crop. The results indicated only very little variation in the amylose content with age of the crop. Xanthosoma starch also had a similar amylose content viz. 15-25% and very little varietal variation was observed. The amylose content of ten cultivars of Amorphophallus paeoniifolius was found to vary very little and ranged from 21.9-23.5%. Starch from three accessions of Canna edulis was found to have an amylose content ranging from 24-30%, the highest being observed for purple accession. Coleus starch had an amylose content of 33% . Thus among the different tuber crops, Canna edulis and Coleus starches have highest amylose contents (Tab 13).

      Soluble amylose is an important component in starch, which plays a significant role in deciding the textural properties. The amylose molecules in the amorphous regions are supposed to make up this fraction and so are easily leached out and hence held responsible for cohesiveness in cooked tubers The soluble amylose contents in the tuber crops starches determined using iodimetry ranged from 25-40% of total amylose. Soluble amylose content in different varieties of cassava was found to be almost equal, even during the growth period. Similar trend was observed for Colocasia, D. alata and Xanthosoma starches. For Amorphophallus starch from different accessions, the soluble amylose content ranged from 9-11% forming nearly 45% of the total amylose content. The soluble amylose content in Coleus starch was 12.8% while it ranged from 10-12% for Canna edulis starch.

      Results on the complexation of tuber starches with the surfactants were quite interesting in that the soluble amylose was suppressed to different levels with different starches and different surfactants. In case of cassava starch, the surfactants had variable effect on soluble amylose portion. The anionic surfactants potassium palmitate and potassium stearate and the neutral surfactants glyceryl monosterate had little suppressive action but the cationic surfactants cetyl trimethyl ammonium bromide and sodium lauryl sulphate had significant effect.

      Increase in concentration of these surfactants led to reduced blue values for soluble amylose. The effect of the cationic surfactant cetyl trimethyl ammonium bromide on the other tuber starches is even more interesting (Tab. 14). Though the surfactant reduced the blue value for soluble amylose of all the starches, the effect on Colocasia and D. esculenta starches was more prominent indicating that the soluble amylose of these two starches may be more anionic in nature or that the amylose helix of these two starches are such as to be able to sterically favour the complex formation. It is worthwhile noting that these two starches have the lowest granule size, but no correlation between complexing properties and granular size could be observed for the other starches. Thus structural differences between starches is evident.

   8. GPC Patterns

      The GPC patterns of debranched starch of different tubers were examined. Starch was treated with isoamylase and the resulting product was passed through Sephadex columns. The fractions were analyzed for total carbohydrates and plotted against elution volumes. The plots (Fig. 13) showed quite lot of difference among the starches in the patterns of peaks demonstrating that the nature of amylose and DP of the starch molecules are diverse. The difference can explain the variation in the properties of the starches.

   9. DSC characteristics.

      Differential scanning calorimeter (DSC) has become an important tool in studying starch gelatinization. The method is simple, fast, requires only small quantities of sample and gives reproducible results. Now the advanced version of DSC, viz., Modulated DSC can give more valuable information regarding starch gelatinization, glass transition and starch-lipid complexes.

      Comparison of the DSC parameters among different tuber starches revealed not only large variation among the different starches, but also between different samples and experimental conditions used like moisture content, rate of heating etc. were used. The DSC data obtained by us by using normal DSC and MDSC is presented in Tab. 15 and Fig. 14. The effect of reheating the sample after one cycle has also been studied. The general trends which emerged from the study of different tuber starches are outlined below. Among the different tuber starches, cassava and sweet potato starches generally had lower T_onsetand T_end values. Highest values were noticed for Colocasia starch and the other starches had values in between. DSC of the starches from the two curcuma varieties showed that the gelatinization peak of C. malabarica starch was a doublet and the splitting of peak may be attributed to structural differences in the starch.

      Thus it is evident that there is considerable variability in the gelatinization temperatures among the different tuber starches. The difference in the gelatinization temperatures can be traced to the variation in the starch intermolecular bonds. High temperature of gelatinization can be indication of higher stability of the starch crystallites in the starch molecules, which means more heating is required to swell the granules. In addition, a number of other factors like varietal differences, environmental conditions and the experimental protocols like level of moisture, sample preparation, rate of heating and instrument used – all contribute to the values.

      The range of gelatinization is also quite different among the different starches. In our study, we obtained highest range for cassava starch (12.90°) and lowest for Xanthosoma starch (4.7°). Higher range has also been attributed to higher level of crystanillity which imparts higher structural stability so that the water molecules need longer time to penetrate the crystalline areas. There does not appear to be any relation between the XRD pattern and gelatinization range, as both cassava and Xanthosoma have ‘A’ pattern but the range of gelatinization are far different. An ‘A’ XRD pattern indicates closer packing and should have a higher range of gelatinization, but such an effect is not observed. In addition , the T_onset does not appear to be influencing the range of gelatinization in any regular pattern. Similarly granule size and gelatinization range do not show any relationship. In the Brabender viscographic curves, the yam starches show a longer range which does not appear in the DSC. The difference in the water: starch ratio between Viscography and DSC may be reason for this anomaly.

      Gelatinization enthalpy is another important parameter obtained by DSC. The value depends on a number of factors like crystallinity, intermolecular bonding etc. For cassava starch values of 12.4 and 16.6 J g^-1 were obtained by us in Normal DSC and Modulated DSC respectively. The enthalpy of gelatinization of five varieties of cassava varied from 10.6-13.8 J g^-1.

      Gelatinization enthalpy was also found to depend on the genetic and environmental factors. Effect of variety and environmental conditions was also evident.

      The gelatinization enthalpy was 12.9 J g^-1 for Colocasia starch and 16.6 J g^-1 for Amorphophallus starch. For D. alata starch the values were 11.6 and 15.4 J g^-1 . For D. rotundata starch values of 10.28 and 15 J g^-1 were recorded. For Xanthosoma starch the enthalpy was 9.1 and 15.22 J g^-1 in the two studies. For Canna starch, an enthalpy value of 17 J g^-1 was recorded. There does not appear to be any relationship between the enthalpy and other factors like amylose content, granule size and XRD patterns. Since the amylopectin has a more crystalline nature, higher amylopectin content has been considered to contribute to higher enthalpy of gelatinization. However such an effect is not evident in any of the studies. Again the differences do not reflect the structural differences among the tuber starches. The results also show that unlike the gelatinization temperatures, which show wide variation, the enthalpy of gelatinization is within in a small range for the tuber starches.

      DSC has been quite useful in studying retrogradation properties of starches. Starch retrogradation is another aspect in which lot of vagueness exists. Originally it was assumed that retrogradation occurs by association of the amylose chains. But using waxy starches, it has been established that amylopectin can also take part in the retrogradation by the association of the outer chains Retrogradation parameters of tuber starches have been examined by DSC. The values for T_onsetshowed very wide range from 37° to 58°, highest and lowest being for arrowroot and Xanthosoma starches respectively. It is well known that retrogradation brings about drastic reduction on T_onset of starches. The highest reduction was observed for Amorphophallus and lowest for D. esculenta starches. The range also varied widely from 12.2°C for cassava to 43°C for D. alata. The wide range in values indicates that the retrograded starch contains recrystallised amylopectins of different crystanillity. Among the starches, D. alata starch appears to have the maximum variability in crystalline structure. . The reduction was to the extent of 2.5 fold to 7 fold. No relationship could be derived from the properties of retrograded starches with those of the nonretrograded starches.

      1. MDSC in starch gelatinization

         The tuber starches were studied in detail using Modulated DSC (MDSC) . MDSC uses a continuous heating-cooling cycle which is helpful in identifying reversible and non-reversible processes occurring in polymers. Gelatinization of starch is a non-reversible process whereas melting of the amylose-lipid complex is reversible one. In the normal DSC, the latter is studied by subjecting the starch to cooling after the first run and then re-heating it so that the gelatinization peak does not appear during the second heating. Since MDSC separates the two processes, MDSC should be better in studying the two processes. It was with this presumption that MDSC studies were carried out on various starches.

         MDSC was run on a Seiko SII 6200 DSC equipment provided with a built-in software. The samples were directly weighed into coated aluminum pans. Double distilled water was added to get a water-starch ratio of 2:1 and empty aluminum pan was used as reference. The heating cycle used was as follows: first heating from 15°C to 150°C /min, cooling to 30°C/min; second heating from 30 to 130C at 3°C/min and final cooling to 30°C /min. Using the built-in software the thermogram was resolved into reversible, irreversible and total peaks. The gelatinization onset (T_o ), Gelatinization end(T_e ) and gelatinization enthalpy (DH) were determined from the thermographs and corrected for dry weight. The results were expressed as mean of at least three values.

         Some of the graphs obtained for the different starches are given in Fig. 15. The graphs corresponding to total (S) contains both the reversible and irreversible changes taking place during gelatinization.. The starch gelatinization is an endothermic process, which involves absorption of water by the starch during swelling, and it is an irreversible process. A second peak corresponding to the melting of the starch-lipid complex, is obtained for starches having lipids in them and is due to the complex formed between starch and lipid, and the peak depends on a number of factors like concentration, lipid type and heating conditions.

         In modulated DSC, the heating takes place in stages, in which there are continuous cycles of heating and cooling and it is claimed that it is able to separate the reversible and irreversible processes taking place during the runs. It is therefore expected that it should be possible to separate the starch gelatinization and starch-lipid complex melting even if they overlap. So during the splitting of the total peak into K (kinetic) and C modes, the irreversible transformation viz. gelatinization should appear in the K graph while the reversible viz. starch-lipid melting should be only present in the C graphs. However the result shows that such a clear separation does not take place. Whereas the DH for K is invariably lower compared to S(Total) , the C graph invariably contains some peak corresponding to the starch gelatinization which is not expected to be there (Fig.15). This indicates that the separation of S into K and C does not strictly follow the reversible and irreversible pattern. It is also quite possible that the so-called gelatinization of starch is not fully irreversible and may contain reversible processes also. This is invariably true for all the starches examined The absence of peak for starch-lipid melting during the second heating cycle proved beyond doubt that the tuber starches do not contain any naturally bound lipids in them. The result is interesting in that Pacchyrrhizus plant produces pods and so it can also be considered as a leguminous crop. The starches from leguminous crops usually contain native lipids and show the starch-lipid melting endotherm. The absence of such a peak in the DSC of Pacchyrhizus starch shows that the starch has more similarity to tuber starches rather than legume starches.

      2. Starch-surfactant complexation in determination of amylose content in various starches

         The starch–surfactant complexation has been used for determination of amylose in the tuber starches. The principle behind this method is based on the comparison of the enthalpy values of the starch-lipid melting with standard amylose. The method has many advantages over the other common methods. Iodimetric method is very sensitive to pH and often provides inconsistent results due to the difficulty experienced in dissolving starch to get uniform solution. GPC method is long and is also not so reliable. The DSC method is very simple and requires only very small quantities of the sample . The principle behind the method is that when a lipid or surfactant is added to starch, it binds with the amylose fraction of starch. The measure of enthalpy of melting of complex can give the estimate of amylose content in starches. It has also observed that whenever the complex is cooled and reheated, the enthalpy during second heating is higher compared to first heating. In addition the T_onset is also elevated.

         This indicates that during first heating when starch gelatinises, the amylose portion is released and hence can form complex with the lipid or surfactant. So if the process is repeated, the complexation becomes more strong. In modulated DSC, there is continuous heating-cooling cycle during each run and hence it can be considered as a series of heating and cooling operations. So this method should give better results. This is the rationale behind use of MDSC in amylose determination. Another drawback in the earlier methods is use of the lipid lysolecithin which is quite costly. So use of easily available surfactants viz.

         Sodium dodecyl sulphate(SDS) and cetyl trimethyl ammonium bromide (CTAB) was tried.

         The experimental technique involves weighing out 2-5 mg starch into the pans, addition of 5% solution of the surfactant, so that the starch : water ratio is 1:2. After sealing, the pans are allowed to equilibrate for 1 hour. The pans are transferred to the equipment and an empty pan is used as reference. The heating sequence is as follows. First heating from 15 to 150°C at 3 °C /minute, cooling to 30 at 30° C/min, reheating from 30 to 150°C at 3°C/min and final cooling to 15°C at 30°C/min. The amplitude was fixed at 2°C/min and frequency 0.017Hz so that effectively, the sample is heated continuously by 3°C, immediately cooled by 2°C and so on. Each run takes nearly an hour and half. A sample of potato amylose was used as standard. The enthalpy of gelatinization was calculated using standard software. The results were calculated on the basis of at least three runs.

         Since the starches were obtained from different sources and it is not always safe to rely on reported data, the amylose contents in the starches were determined for all samples by standard iodimetric method. This involved dissolving the starch in DMSO-Urea solution, followed by determination of absorbance of the iodine-starch complex obtained from the solutions. The total amylose content was determined by precipitating out the starch from the solution using alcohol followed by re-dissolving in DMSO-urea solution and determination of absorbance. In order to compare with data obtained by GPC, the analysis of all samples was carried out at the Food Science department of Uppsala university based on the standardized procedure developed there. Starch was debranched using -amylase followed by fractionation in a Sepharose column. For debranching, 5l isoamylase from Pseudomonas amyloderamosa was used and after reaction, the enzyme was inactivated in a boiling waterbath for 5 min. The samples were injected on a Sepharose CL 6B column (1.6 X 70 cm. ) using 0.25 M KOH as eluent at a flow rate of 13 ml per hour. Two ml. fractions were collected and the elution profile was obtained by the phenol- sulphuric acid method

         The results are presented in Table 16. The results obtained by the MDSC method have been compared with those obtained using iodimetry and Gel Permeation Chromatography. The starch samples used in the study are of different origins including cereal, tuber and pea starches. In addition genetically modified starches from barley, maize and potato have been compared. So a wide range of starches has been examined. Among the cereal starches, there was good agreement between values obtained by GPC, iodine staining and the present method. The values obtained for CTAB were higher than those obtained on using SDS. For tuber starches, both the surfactants gave very good results. On the whole, the results indicate a very good match between the amylose contents determined by iodimetry and GPC with those obtained by the present method. A comparison between the two surfactants do not show any major difference though it appears that CTAB is more sensitive for low amylose samples. The length of the hydrophobic chain is very important in deciding the complex formation. It has been found that 12 to 18 carbon chain length is optimal for complexation with amylose. Both the surfactants have this range and the role of the hydrophilic head needs further examination. The MDSC method for determination of amylase reported for the first time thus appears quite efficient since it gives acceptable vales for a wide range. The surfactants used are widely available. Unlike iodine staining which requires defatting to get more accurate results, the present method does not need any such step.

      3. Influence of chain length on the starch gelatinization

         The influence of chain length of the lipid used for complexation with starch on the gelatinization characteristics was studied using MDSC. It is well-known that for efficient complexation, a minimum chain length of the lipid or surfactant is required. Though complexation has been reported for even 6 carbon system, the optimum length is 12 to 18. The strength of the complex can be gauged from the gelatinization temperatures as well as gelatinization enthalpy. The higher the gelatinization temperature and gelatinization enthalpy, the stronger will be the complex. Earlier work has been mainly on pure amylose and complexation was mainly studied using Glyceryl derivatives of various fatty acids and data has often been compared with the results from different lipids or surfactants.. It was decided that a comparison will be more relevant if the surfactant used is same and for this purpose lysolecithins having different chain length in the fatty acid portions were used for complexation studies. A large number of starches from different sources was examined. These included cereal starches, tuber starches and genetically modified ones. Lysolecithin having the fatty acid residues (C6:0), (C-10:0), (C-14:0) and (C-18:0) were used for the study. A 1% solution of the surfactant was prepared in double distilled water. Starch was weighed accurately (1-5mg) into the pans, 2 times the weight of the surfactant solution was added , the pans were sealed and MDSC was run as described earlier. The T_onset and T _end and H were determined from the thermograph after the resolution of the graphs into reversible and irreversible graphs. At least three runs were made for each sample. Results are presented in Table 17 and Fig. 16.

         No	Starch	C-6			C-10			C-14			C-18
         		Tonset	Tend	DH	Tonset	Tend	DH	Tonset	Tend	DH	Tonset	Tend	DH
         1.	Cassava	51.6	60. 1	11.7	63.4	76.0	11.6	65.4	77.4	10.5	64.3	76.5	8.8
         		–	–	–	–	–	–	93.7	103.4	1.88	106.0	114..8	1.8
         		–	–	–	73.5	81.0	1.3	98.4	104.9	2.4	108.7	116.4	2.75
         2.	Xanthosoma	74.2	81.5	13.13	72.0	79.2	12.8	75.6	82.6	11.38	74.1	81.5	9.3
         		–	–	–	–	–	–	92.3	101.5	2.3	105.0	115.8	3.0
         		–	–	–	72.7	79.8	1.86	98.2	103.5	2.82	108.5	116.4	3.75
         3.	Colocasia	76.7	80.9	11.7	77.2	85.1	11.6	79.5	86.6	9.36	79.5	86.9	12.0
         		–	–	–	–	–	–	95.1	102.0	0.53	104.2	112.5	0.5
         		–	–	–	80.2	87.3	0.38	99.5	105.2	0.98	108.4	115.4	1.15
         4.	D. alata	71.2	80.3	12.9	70.3	80.0	12.3	72.2	80.6	8.72	71.5	80.2	10.6
         		–	–	–	–	–	–	94.1	102.0	0.09	104.2	115.5	3.35
         		–	–	–	72.1	78.5	1.52	94.1	104.2	2.36	109.2	116.5	3.6
         5.	D.esculenta	66.1	71.1	13.13	63.2	75.2	9.18	62.4	73.0	6.5	64.9	73.3	6.6
         		–	–	–	–	–	–	94.5	105.2	1.35	104.0	114.0	1.75
         		–	–	–	71.1	78.6	1.21	98.6	104.5	1.5	109.8	115.4	1.48
         6.	C.edulis	64.2	71.6	9.08	63.1	71.2	12.91	64.3	71.2	8.12	63.9	71.5	12.67
         		–	–	–	–	–	–	92.4	103.3	3.78	104.4	116.5	4.7
         		–	–	–	73.3	78.8	1.6	98.2	104.2	2.98	110.5	117.5	3.98
         7.	Amorphophallus	78.0	84.8	12.26	75.1	83.0	14.48	77.9	85.0	12.16	77.15	84.3	9.38
         		–			–	–	–	94.1	102.3	2.0	104.5	116.7	3.38
         		–	–	–	71.2	78.6	2.28	98.0	104.3	2.9	109.1	116.8	3.5
         8.	D.rot	74.5	83.1	11.8	72.1	81.9	13.2	73.3	84.7	9.65	72.4	81.3	8.45
         		–	–	–	–	–	–	93.9	101.3	1.25	103.8	115.5	2.52
         		–	–	–	73.7	80.2	1.49	97.5	103.0	2.71	109.1	116.2	3.2
         9.	Pacchy.	67.0	73.1	8.83	62.6	72.9	9.47	64.4	74.5	9.45	63.7	75.3	8.10
         		–	–	–	–	–	–	94.2	103.1	2.37	104.7	113.8	2.35
         		–	–	–	75.7	85.1	0.66	99.0	105.4	2.53	106.7	118.2	3.24
         10	Arrowroot	69.1	77.3	12.31	68.8	78.0	12.61	71.3	79.5	7.85	70.2	78.6	8.37
         		–	–	–	–	–	–	97.3	102.4	2.15	106.3	115.9	3.0
         		–	–	–	71.8	79.2	1.82	97.5	103.9	2.62	109.1	116.6	3.4

         Various interesting points emerge from the results. As is evident from the table, the chain length really influences both the gelatinization temperatures and the enthalpy. The peaks corresponding to starch gelatinization, starch-lipid complex melting during first heating and during second heating are affected. A comparison of the values for cassava starch is as follows 64-76°C for starch gelatinization, 106-114°C for starch-lipid complex melting for first heating and 108-116°C for second melting with C-18. For C-14, the values are 65-77, 92-101 and 98-103°C.

         In case of C-10, the gelatinization temperatures is.63-76°C. Though there is no peak corresponding to starch-lipid melting during first heating a peak corresponding to melting of starch complex appears during second heating . The absence of the peak for the first heating is due to the masking of this peak by the starch gelatinization peak. This is confirmed by the enthalpy values for the starch gelatinization.

         Similar results were obtained for all the tuber starches. It was also observed that there was no peak for starch-lipid complex melting when C6:0 lysolecithin was used showing that there is no effective complexation when the chain length is only 6 and this effect is clearer when tuber starches are used. The peak for starch-lipid complex melting was observed only for Glacier variety of Barley during first heating and the result with this starch can be explained on the basis of high amylose content in this variety. The following conclusions can be drawn. There is no effect of chain length on the starch gelatinization temperature since there is only very small difference between the values.

         Increase in chain length leads to higher temperatures for the starch-lipid complex melting indicating the strength increases with increase in chain length. This is further confirmed by the results for the C6 system where there is no peak at all for the starch-lipid complex melting.

         The results with the enthalpy also showed similar trends. For tuber starches, the DH showed a steady increase with reduction in chain length. The DH values for starch-lipid complex melting show reducing values with reduction in chain length for most starches. However for the tuber starches, the first starch-lipid complex melting peak is absent and hence there is no DH values for these. The values for the second peak follow clearly the trend of reduction with reducing values of chain length.

         The data thus clearly indicate the following. The longer the chain length, the complexation is stronger. This is shown by the higher gelatinization temperature and gelatinization enthalpy of the starch-lipid complex melting.

         These show higher values during second heating as expected proving that amylose leaches out during gelatinization of starch. The complexation formed with C-6 is very low or absent or even if formed, it is very weak. The role of the hydrophilic head is also important. Some differences exist between the cereal and tuber starches in their complexation with C-6 lysolecithin indicating differences in their internal structures. The results will be useful in identifying the correct lipid or surfactant to be used to get optimum complexation with starch and thereby modification in properties.

      4. Effect of native lipids on the gelatinization of different starches

         It has been observed that durum wheat provides better quality noodles compared to normal wheat flour. It was felt that the native lipids present in these may be influencing the starch gelatinization differently, since it is basically the starch which gives the texture, structure and shape to the noodles. To examine this aspect, it was decided to study the effect of native lipids from wheat, durum and rye on the gelatinization characteristics of starch of D.alata and pachyrhysus. These two root starches were taken since D alata starch has a ‘B’ XRD pattern and Pachyrhysus ‘A’ pattern and so the study can throw some light on the effect of natural lipids on starches with different structural properties and whether complexing with lipids can make them suitable for noodle production.

         The total lipids were isolated by extraction of the corresponding flour at room temperature with ethanol and removal of the solvent in vacuum, The polar lipids were separated from the total lipids by column chromatography. Accurately weighed quantity of The lipids (200mg) was allowed to equilibrate with water, starch (200 mg) was added and the mixture was thoroughly mixed to ensure uniform distribution. For DSC studies, a small quantity of the material was accurately weighed and transferred into the pan, taking care that all the material is completely inside the pan, it is sealed and placed into the sample holder. The DSC was run as follows: heating 15 to 150°C at 5°C/min, cooled to 30°, reheated to 130°C at 5°C/min and cooled again. The gelatinization parameters were directly read out from the thermograph. Each sample was run at least three times using polar lipids and total lipids.

         The results showed that the lipids form strong complex with the starches examined. However the results do not show any higher value for polar lipids compared to total lipids. It was expected that the polar lipids should form stronger complexes and should give higher values for both gelatinization temperatures and enthalpy.

         Such trend was not observed . But the results clearly showed that the tuber starches can form complexes with lipids and there is no inhibition for such complex formation and hence lipids can be used to modify the starch properties.

   10. Gelatinization and Pasting temperatures

       Gelatinization of starch takes place over a definite range of temperature known as gelatinization temperature. Gelatinization temperature can be measured microscopically, by DSC and also by using a Viscograph. However use of viscograph provides the pasting temperature rather than gelatinization temperature. The former may be defined as the temperature at which a perceptible increase in viscosity occurs and is always higher than gelatinization temperature. Among different tuber starches, cassava starch has the lowest gelatinization temperatures. No relationship between granule size and gelatinization temperature was observed. Pasting temperature of H-165 starch determined using a viscograph was slightly lower than those for most of the other varieties, and M4 starch had the highest range of pasting temperature . These values are quite close to DSC values of 66 and 78°C for T_onset and T_peak respectively. When cassava starch was subjected to steam pressure treatment at different pressures, there was progressive increase in pasting temperature by 2 to 9° C depending on the time of treatment and pressure used. The pasting temperature rise was higher for longer time of treatment and higher pressures . Increase in pasting temperatures was also observed on treating the starch with surfactants. It was found that different types of surfactants generally increased the pasting temperatures, but most pronounced effect was noticed on treatment with potassium palmitate and potassium stearate. On esterification to acetate, the pasting temperatures were reduced due to weakening of associative forces. Examination of gelatinization temperature of cassava starch in non-aqueous solvents indicated that the values were enhanced tremendously in glycerol and ethanol, while in DMSO and formalin, only slight increase was noticed. The large increase in the first two solvents can be traced to steric factors.

       The pasting temperature of sweet potato starch varied between 66.0 to 86.3°C. Pasting temperatures of different cultivars of Colocasia esculenta and Xanthosoma sagittifolium starches have been determined using the Brabender Viscograph. There was only very slight difference between varieties, but the pasting temperatures of these two starches were distinctly higher than those of cassava and sweet potato starches. The gelatinization temperature of starch of three accessions of Canna edulis was 74-85 to 80-95°C by the Brabender Amylograph, but 74-75°C by the RVA. Pasting temperatures of starch of Amorphophallus paeonifolius extracted from 10 accessions were compared and again no significant variability was noticed and the starch gelatinised in the same range as the other aroid starches viz. ., 81-85 to 82-85°C. However the range was lower compared to cassava or the yam starches. For Pacchyrrhizus starch, values of 74-79°C were obtained and only minor variation existed among varieties . The values for curcuma starch was 81°C by RVA while it was 65-85°C for coleus starch determined microscopically. Yam starches generally gelatinised over a temperature range of around 20°C and gelatinization continued even after 95°C showing strong intermolecular linkages. The large range for yam starches may be attributed to the presence of phosphate linkages in these starches (similar to potato starch) since it was found that the yam starches invariably contained higher quantity of phosphorus.

   11. Viscosity

       An important and useful property of starch is that it provides a viscous paste when heated in presence of water. It is this viscosity that accounts for the use of starch in textile, paper, adhesive and food industries.

       Detailed Brabender viscographic studies on cassava starch from different varieties and at different starch concentrations have been carried out and the results indicate variation in the viscosity parameters. There was no correlation between viscosity and granule size, peak viscosity and setback viscosity.

       In a Redwood Viscometer No. 1, a 2% paste of cassava starch had a value over 50 seconds and only minor variation among varieties was observed. When cassava starch was subjected to steam pressure treatment, the viscosity fell steadily with increase in pressure and time of treatment. At a pressure of 15 psi for 150 minutes, the viscosity fell from 58.5 seconds to 30 seconds. The same trend was observed in the Brabender Viscosity results also, where the peak viscosity dropped from 430 BU to just 30 BU. It was also observed that the reduction in viscosity was linearly related to the severity of treatment.

       It was observed that different surfactants affected the viscosity of cassava starch differently. Viscosity of cassava starch in various non-aqueous solvents has been compared. Ethanediol depressed the Redwood viscosity from 55 seconds to 17.5 seconds, while glycerol increased it to 175 seconds. DMSO and Formalin also increased the viscosity, but to a much lower extent.

       In addition to peak viscosity, the breakdown in viscosity is another important criterion that decides the applicability of starch in food and industry. In this respect, cassava starch is considered inferior to maize starch, because its viscosity is rapidly reduced on heating under shear leading to a ‘long’ and cohesive texture for its paste, which is not desirable in food and textile applications. Much effort has been paid to strengthen the starch paste viscosity. Steam pressure treatment can improve the paste stability, but it is accompanied by a corresponding reduction in peak viscosity. However, surfactants were found to have a more desirable effect.

       In view of the presence of a large number of hydroxyl groups in starch, cross linking has been used to stabilise starch viscosity. Various di- and tri-functional chemicals have been used to crosslink the starch. The most effective chemicals have been epichlorhydrin, phosphorus oxychloride and sodium metaphosphate. These chemicals bind the starch molecules in the granules, increasing the associative forces, rendering the starch granules stronger and preventing breakdown during heating and stirring. The textural properties of cassava flour could be improved by treatment with phosphoric acid and suuccinic acid. Many salts can influence starch properties depending on the type of salt used.

       Whereas a large amount of work has been done with cassava starch, only very little work has been carried out on the other tuber starches. Sweet potato starch behaves almost similar to cassava starch in its viscosity characters, viz., peak viscosity, viscosity breakdown and setback viscosity.

       Data on the viscosity characteristics of the other starches are also widely variable. The viscosity of Colocasia starch extracted from ten cultivars was less than that of cassava starch and close to cereal starches. Considerable difference in peak viscosity values among the accessions was observed. At 5% concentration, the viscosity ranged from 130 to 350 BU. The lowest and highest values at 6% and 7% concentrations were 200 and 600 (6%) and 320 and 880 BU (7%) respectively.

       It was also interesting to observe that C-9 starch having the highest granule size and amylose content had the highest peak viscosity. As earlier stated, there was only nominal breakdown in viscosity even at the highest concentrations pointing out the possibility of using this starch in various applications which require paste stability. The viscosity values were nearly same for the corm and cormel starches.

       Variability in the viscosity properties of Amorphophallus paeoniifolius starch extracted from ten accessions was quite minor. The values ranged from 280-380 BU at 5% concentration, 450-620BU at 6% concentration and 770-920 BU at 7% concentration. The viscosity breakdown for the starch samples was very low, similar to other aroid starches and in the same range as the cereal starches Starch extracted from seven varieties of Xanthosoma sagittifolium has been compared for its viscosity. There was only a slight difference in the viscosity values among the varieties. The highest value was 360 BU and the lowest 280 BU at 6% concentration. The breakdown varied from 0 to 30 BU at 7% concentration, showing the stability of the starch paste.

       All the yam starches showed a characteristic pattern of slow rise in viscosity and even after 95°C an increase in viscosity were noticed. This implies that all the granules do not gelatinise at 95°C and some of them gelatinise only during the holding period. In this respect the yam starches resemble potato starch. The viscosities of starch of four varieties of D. esculenta extracted using water and ammonia solution were compared. There was only minor difference in the values between the varieties (800-950 BU), but the viscosities of starch extracted using ammonia solution were much higher than those obtained by water extraction. When the viscosity of D.alata starch from five varieties extracted with water and ammonia was compared, the results were not so conclusive. Mostly, there were no peak viscosity values. As observed for the D.esculenta starch, there was no observable breakdown in viscosity on heating and stirring.

       The 2% viscosity of six varieties of D. rotundata starch varied from 37.5 to 46 seconds . As far as the paste viscosity is concerned, the values ranged from 325 to 550 BU at 5% concentration and 680 to 920 at 6% concentration. The viscosity of yam starches during growth period has also been examined (Tab.18). The viscosity breakdown was quite low in spite of the high viscosity levels.

       The yam starches contain three to four times as much phosphorus as found in cassava and aroid starches. Steam pressure treatment of the D.alata and D.rotundata starch has been found to reduce the viscosity. The reduction in viscosity was found to be directly related to the pressure and time of treatment (Fig. 16). The peak viscosity came down to nil value at 15 psi for 60 minutes for both the starches. This method can be tried as a method for modification of starch, since the process is simple and does not involve use of any chemicals and so workup is simple. The Redwood viscosity of Coleus starch was found to be 37 seconds at 75°C for a 2% solution which increased to 56 seconds on cooling to room temperature.

       Rapid Visco Analyser has become more popular in studying viscosity in view of the fact that it is faster and requires smaller quantity of sample. The RVA profiles starch of some varieties of D. alata, D.esculenta and D. rotundata harvested at different maturity have revealed that maturity did not affect the rheological properties to any major extent. Pacchyrrhizus starch had similar viscosity as arrowroot and there was only very little difference among 10 varieties examined. The peak viscosity values for the starches from three accessions of Canna varied from 3887 to 4187 cPs. The RVA patterns of the starches from the three accessions do not coincide exactly with the Brabender data. The difference may be attributed to the higher concentration of starch and rate oh heating used for RVA. The results indicate some breakdown for all the three starches but is not widely different among them . Viscosity studies on Curcuma starch showed that variation exists between the two species studied, and removal of curcumin resulted in an increase in peak viscosity of C. zedoariastarch almost to the level of C. malabarica starch. The breakdown in viscosity was quite low showing that the granules are quite strong and resist breakdown under shear and heat. In this respect also, it resembles yam starches rather than cassava.

   12. Theological studies on tuber starches.

       A comparative study of rheological properties of different tuber starches was made using the Bohlin rheometrer. The experimental conditions were:

       a. Torque element 1.542 gcm.
       b. Measurement interval – 60s.
       c. Thermal eqlbm. Time – 10s.
       d. Sensitivity – 1%
       e. Amplitude – 3%
       f. Heating rate – 1.5 °C/min
       g. Auto strain off.

       The cup in which the starch slurry had been poured into was a C25 measuring system. Frequency, phase angle, viscosity, storage modules(G’) loss modules (G”), range , strain and correlation were determined at each one minute time interval. The ratio G’/G” was calculated at 95°C 75 °C, 60°C and 35°C for different time intervals.

       Three concentrations were examined viz. 3,4 and 5%. From the data, three factors were compared including viscosity, elastic modulus and phase angle. The results indicated wide variability in these properties among the different starches.

       1. Viscosity

          The viscosity data obtained from the rheograms did not provide clear trends except that increase in concentration led to increase in viscosity, but the increase was not quantitative. However, the difference in the viscosity among the starches was obvious. Canna edulis and D rotundata starches had higher levels of viscosity at all the three concentrations while Colocasia and D. esculenta had low values. The aroid starches appear to provide more reliable results compared to the yam starches There is good agreement with the Brabender results that Canna edulis , D. rotundata and D. alata starches possess high viscosity while Colocasia and Xanthosoma have low viscosity levels. Breakdown in the viscosity and setback observed with the Brabender viscographs are not very evident (Fig.18).

       2. Storage modulus

          The storage modules (G’) for the different starches at the three concentration are given in Figs 18. Wide variation between the different root starches could be observed.

          The range of values for Xanthosoma starch at 3% concentration was 0.1 to 0.2 Pa during the heating-cooling cycle; which increased to 5-7 Pa at 4% and 12-33 Pa for 5% starch concentration. During cooling, a slight decrease was observed for 3% concentration. The results indicate low gel strength at this concentration. At 4%, the increase in G’ during cooling was more apparent and this became more significant at 5%, showing high gel strength at 5% concentration.

          For D. alata starch , the G’ values were quite high especially at higher concentrations. The range at 3% was 6-12.6 Pa . There was a small increase during holding at constant temperature (95°) and cooling to 35°C. At 4% concentration, the values were much higher (36-53 Pa) with a small increase during cooling and holding at 35°C but was not very significant. But at 5% , the increase during later stage was very evident. Earlier experience on this starch had shown that the starch has a good gel strength and is confirmed by the present study especially at higher concentration.

          For Pacchyrrhizus starch, the G’ was quite low at 3% during heating and till last stages of cooling, when there was a noticeable increase. At 4% concentration, a discrepancy was evident during initial phase, but the increase in cooling was obvious. This was further confirmed by the large increase in 5%, The results indicate strong elastic nature and gel strength for the starch at high concentration.

          Colocasia starch had very low G’ at 3% and hence very low elastic character. When concentration was increased to 4% the values were much more perceptible and there was a sharp increase during cooling . For 5% concentration the G’ was nearly 30 times that at 3% and increase on cooling was very evident. Thus elastic character becomes clear at 5% and though the gel strength was not high , tendency to gel was obvious. These characteristics may be related to the small granule size of this starch.

          Arrowroot starch at 3% concentration had a detectable G’ at 3% concentration, which increased with concentration, two fold at 4% and three fold at 5% concentration. The storage modulus followed the same pattern as the viscosity, a fall during holding at 95°C , than a slight increase on cooling and then remaining steady. This was true for all the three concentrations and the increase was nearly same at these concentrations. For cassava starch, the trends were clear. At 3%, the values were insignificant, but at 4% , they became evident while at 5% they were much more significant. . However at 5%, there was a steep fall during holding period and then a small rise, very similar to arrowroot starch. In this respect G’ resembled viscosity pattern at 5% concentration. The rapid fall at 5% concentration during heating and holding reflects the breakdown of the starch which is also evident in the Brabender Viscograph. Though the starch exhibits good elasticity , the gel strength doesn’t appear to be very high on cooling.

          D. esculenta starch had very low G’ at 3%. Even at 4%, the G’ was low, but during cooling some increase was evident. At 5%, the G’ was quite detectable and showed a rapid raise during cooling. No further change was observed. The viscosity results also had shown a similar trend. D. esculenta starch has very small granules and thus similar to Colocasia starch. For these starches swelling is also not very high and hence the observed low values for viscosity and G’

          Canna starch has been found to possess very high gel strength both in the Brabender Viscographic measurements and also during preparation of various food products This was confirmed by the G’ data in the present study. At 3% concentration itself, the values were highest during the starches examined (10-22 Pa). At 4%, the increase was two fold which further rose 10 fold at 5% concentration. Though there was a tendency to increase during cooling, it was not very high. The starch thus possess high elasticity and gel strength and can be useful in many food applications. Canna starch is known to possess high phosphorus content and perhaps this helps in its possessing high G’ and gel strength.

          D. rotundata starch, was similar to canna starch in having high G’ . At 3%, the value was 10-35 Pa increasing to 20-35 for 4% and further to 45-65 Pa for 5% concentration . D. rotundata starch has also been found to be possess high gel strength and elasticity and is confirmed by the present results.

          For Amorphophallus starch trends were more evident . The G’ values were 1.4-3.5 Pa for 3%, increasing to 8-10 for 4% and further to 14-23 Pa at 5% concentration. There was also an increase during latter phase of cooling. This indicates that the starch has higher elasticity compared to the –other aroid starches.

          The results on storage modulus indicate that Canna edulis and D.rotundata starches possess high elasticity and gel strength. The aroid starches have lower G’ , but definite trends could be observed for these starches.

       3. Phase angle

          The change in phase angle due to shear gives an indication of the strength of the starch gel. It reflects the ratio of elasticity to viscosity of the samples.

          The phase angle values for Xanthosoma starch were in agreement with the G’ values. At 3%, the phase angle values were in the range 27-65 which dropped to 12-20 at 4% and 2-24 for 5% concentration. There was a steep fall at the initial stage of heating at 3%, which may be due to very low values for G’ at this stage. For 4% concentration also, there was a small increase in phase angle during cooling, which is not explainable. However at 5%, the behaviour was as expected and there was a steady decrease during cooling and then remained steady

          For D. alata starch, the data was rather irregular and though the trend indicated a decline with increase in concentration, the data was not in consistency with G’ values. G”, though low, may also be influencing the deformability of the starch paste.

          Pacchyrrhyzus starch exhibited definite trend of a reduction with increase in concentration. Lowest phase angle of 13-28 was observed for 5% concentration as expected. Compared to other starches, the values are quite high and in conformity with G’ values. Though the starch does not possess high elasticity, the phase angles were on the higher side.

          The data for Colocasia starch was quite inconsistent at 3%, but for 4% and 5% the values were more reliable and as expected there was a reduction with increase in concentration. However there was no decrease in the values on cooling, -indicating that gel strength does not increase with cooling.

          The values for phase angles were nearly same at 3, 4 and 5% concentrations for arrowroot starch. Though G’ increased steadily with concentration, the phase angles were nearly equal . The range was 16-70 for 3%, decreasing to 14-48 at 5% and there was no decrease in the phase angle values during cooling showing poor gelling tendency.

          Cassava starch had a phase angle value of 38 at 3% which fell to 10-25 at 5% concentration . The high value at 3% is due to the low G’ at this concentration. The values were nearly same at 4 and 5% concentration.

          Cassava starch thus appears to achieve high elasticity at 4 and 5%. The gel strength cannot be considered very high but still large.

          For D. esculenta starch, the values were very low at 3% concentration but at 4 and 5% levels, the values were in line with G’ values. At 5% concentration, a high gel strength is indicated by low value for phase angle. Though G’ had risen very rapidly at later stages, this effect was not visible in the phase angle values.

          Canna edulis starch had exhibited high G’ at 4 and 5% concentration, but the phase angle values do not reflect the same effect. Though there was a progressive decrease with increase in concentration, it was not following a definite relation. Even at 3% concentration, the values were quite low indicating high elastic nature of the starch paste.

          For D rotundata starch, the values were quite low even at 3% concentration showing high elasticity , but at 5% concentration, the values did not reflect the high G’ observed. Generally high gel strength was noticed for this starch.

          For amorphophallus starch, the results do not show any special trend, 5% values being lower than 3 and 4% values.

          There was no decreasing trend on cooling. In contrast to G’ which showed high values, the effect was less prominent.

       4. Strain sweep tests

          The results of the strain sweep experiments are presented in Fig 19. The following trends emerged from the studies. Higher the G’, the greater is the tendency to breakdown under shear. This was found to be true for all high G’ starches, viz., D. rotundata, D. alata and Canna edulis.

          For medium G’ starches the thinning was evident at the highest concentration, while it was quite negligible for the low G’ starches. Amorphophallus starch however did not show such tendency. Even among the different

          concentrations, the same effect is very clear. At higher concentrations, the intermolecular collisions will be more and this can lead to breakdown in the G’. In addition, for Colocasia and D. esculenta starches, their smaller granular size may also be contributing to the stability.

          The summary of results is presented in Tab. 18. The results clearly point out that viscosity measured by the Bohlin rheometer is not very reliable at low concentrations and also for starches which gelatinise slowly beyond 95°C . It may not be correct to compare the Brabender values with the Bohlin data. The G’ values are consistent with the gelatinization profile for most of the starches and most of the starches show elastic characteristics especially at higher concentration. The phase angles do not follow the same trend as G’ indicating that the effect of variation in loss modulus (G”) may be important in many cases.

       5. Rheological properties of starch from different pretreatments

          The rheological properties of starch from aroids and yams pretreated with different concentrations of Sodium hexametaphosphate, Glyceryl monostearate , Sodium chloride, Potassium metabisulphite and Ammonia were also determined following the same procedure. The results obtained are summarized below (Tab.19).

          For D. rotundata , GMS treated starch had grater G’ value and G’/G” ratio compared to other treatments indicating a greater elastic behavior for this treatment. For all the treatments and control G’ and G” are maximum at 95° C.

          Thereafter, they decrease and on cooling beyond 35°C, G’ and G’/G” values increase. The result shows that elastic behavior gradually decreases upto 35° C and then an increase in elastic property takes place.

          For control D. alata starch, the G’ value ranges from 105 Pa to 90.4 Pa and G’/G” value from 11.4 to 6.41. SHMP, NaCl, KMS and NH4OH treated starches have lower G’/G” and G” values compared to that of control starches. D. alata control starches possessed lesser G’ values than D.rotundata starches at higher temperature. Hence these starches possess more elastic properties than D .rotundata. at higher temperature. But at 35°C after 45 minutes, the values were higher for D.alata starch. The G’ values were lowest for NH₄OH treated starches and the highest for SHMP treated starches. It was concluded that viscous (liquid) characteristics were more significant at higher temperature while at lower temperature the elastic (solid) characteristics become more prominent.

          G’ values of the D. esculenta control starch ranged from 19 to 25.3 Pa. The values decrease upto 60°C and thereafter increased and on cooling to 35°C the values remain constant. All chemically pretreated starches behaved in a similar manner as that of the control. SHMP treated starches had the highest G’ value but the ratio of G’/G” value was highest for GMS treated starches.

          Amorphophallus tuber starches possess almost same G’/G”values at different temperatures. A small decrease in the G’/G” ratio was observed upto 35°C and thereafter an increase was noticed.. The G’/G” value were found to be highest for GMS treated starches.

          Colocasia tuber starch had higher G’/G” values in the range of 1.4 to 6.8. G’ values increased when the temperature decreased from 95°C to 35°C. Hence the elastic behavior was more prominent at lower temperature for Colocasia starch.

          NaCl ,GMS and KMS treated starches behaved in a similar manner as that of the control samples but cooling after 35 °C did not make any significant change in G’/G”. The G”/G” were higher for GMS treatment than control.

          In case of Xanthosoma control starches, G’/G” value decreased after 45°C and increased on further cooling. But all the chemically treated starches showed decrease in G’/G” upto 60°C and thereafter increased with decreasing temperature. At higher temperatures, GMS treatment provided the highest G’/G” value. But on cooling at 35°C for 60 minutes, G’/G” values were more or less similar for GMS and KMS treated starches.

          Thus it is clear that different pretreatments bring about different effects on rheological properties. However the results indicate that the treatments do not bring about any major changes in the elastic characteristics of the starches and therefore no deleterious effect on the starch properties occurs.

   13. Swelling power and solubility

       Starch swells on heating in water and the extent of swelling depends on the origin of the starch. Swelling power and solubility provide evidence of non covalent bonding between starch molecules. Factors like amylose-amylopectin ratio, chain length and molecular weight distribution , degree / length of branching and conformation decide the swelling and solubility. Cassava starch has swelling power which is in between those of potato and cereal starches- a property in conformity with its observed viscosity. The swelling volume of different varieties of cassava varied from 25.5 to 41. 8 ml g^-1 of starch (Tab. 2). During the growth period , starch of two varieties H-2304 and M4 maintained their swelling volumes within small ranges, while that for some varieties such as H-165 expressed wide variations which indicate that these varieties are very much susceptible to environmental influences, Fig 4. Swelling volumes also depend on the presence of various chemicals and any treatments carried out on starch. A high amylose content and presence of stronger or greater numbers of intermolecular bonds can reduce swelling. Formation of lipid-starch complex can also affect the swelling volumes as also presence of naturally occurring carbohydrates and noncarbohydrates along with starch. This has been amply illustrated in the effect of fiber on the swelling volumes of different varieties. The fiber acts as barrier to free swelling of starch. It was also found that extraction with ethanol or defatting of flour did not change the swelling volumes showing that the suppressive effect is more due to the fibrous material rather than lipid or sugars present in the flour. The starchy flour extracted from fermented tubers also exhibited the same trend . Sodium sulphite was found to have noticeable effect in suppressing the swelling volume of cassava starch. The swelling volume dropped to very low values at particular concentrations The values dropped nearly to zero at 0.05 and 0.1% concentration of the salt (Fig 20). At higher concentrations, the swelling volumes increased to nearly the same level as the native starch. The effect has been attributed to the oxidative-reductive depolymerisation brought about by the sulphite ions. The effect was nullified at higher concentrations due to destruction of the sulphite ions. Similarly the effect of the sulphite was neutralised by addition of propyl gallate which is an oxygen scavenger.

       The effect was not unique to cassava starch and swelling volume of Dioscorea starches was lowered by sulphite at similar concentrations also. Thus the effect can be attributed to the salt. In order to check whether is effect is only with sodium sulphite, other salts were examined including potassium sulphite, sodium chloride, sodium sulphate and sodium phosphate. Potassium sulphite gave similar results showing that the suppressive effect is due to sulphite ions while other salts failed to produce such a trend. Another salt that brought about a similar effect was sodium thiosulphate at similar concentrations, Fig. 21. The swelling volumes were lowered at 0.01% concentration and then increased. The solubility on the other hand experienced a drastic increase at the same concentration and this salt may also be acting similar to sodium sulphite.

       Surfactants also affect the swelling volume of starches. The swelling volume was reduced to half by potassium palmitate and potassium stearate even at lowest concentrations, while glyceryl monostearate affected only to a small extent.

       In contrast, sodium lauryl sulphate and cetyl trimethyl ammonium bromide enhanced the swelling volume considerably. Steam pressure treatment also lowered the swelling volume by compressing the starch molecules and thus restricting the free swelling of starch.

       Considerable variation in swelling volume of different varieties of Colocasia was noticed. The values ranged from 26.5 to 60 ml g^-1 – which indicates a high degree of variability. For C-9 starch, having highest granule size, the swelling volume was the least. Inverse relationship was noticed between the granule size and swelling volume of ten accessions of taro.

       Swelling volume of starch of six clonal selections of D. rotundata has also been examined at different concentrations and only slight differences were noticed among the selections. At 1% concentration the values ranged from 15-25 ml g^-1 and at 5% concentration, the range fell considerably due to deficiency of enough water to swell all granules . Starch from different varieties of D. esculenta, D. alata, Xanthosoma sagittifolium and Amorphophallus paeoniifolius had much lower ranges. The relatively lower swelling of Dioscorea starch compared to potato starch has been attributed to the higher lipid content in the starch and also higher inter-associative forces compared to potato starch. Swelling volume of Canna edulis starch was observed to be 11-14 ml g^-1 for three accessions of canna starch. In general, the aroid starches had rather low swelling volumes. Starch from ten accessions of A. peoniifolius had swelling volumes ranging from 11.7 –13.0 ml g^-1 at 0.5% concentration and 21.5 to 24.4 ml g^-1 at 1.0% concentration. For Coleus starch, the swelling volume was around 25 ml g^-1. For Curcuma starch, the value obtained was 19 ml g^-1 for C. zeodoria and 30 ml g^-1 for C. malabaricum starches. The swelling volume of starch from Amorphophallus extracted from pretreated tubers with different chemicals depended on the chemicals used. All the chemicals lowered the swelling volume, but highest reduction was with Glyceryl monostearate. For Xanthosoma starch, Glycerol monostearate (GMS) and ammonia enhanced the swelling volumes to a small extent.

       Solubility of starch also varies with origin of starch. Solubility depends on a number of factors like inter-associative forces, swelling power, presence of other components etc. Cassava starch has a higher solubility compared to the other tuber crops and the higher solubility can be attributed partly to the high swelling it undergoes during gelatinization. The solubility of starch of different cassava varieties varied from 17.2 to 27.2%. However no direct correlation between swelling power and solubility could be observed. Solubility data of starch from different varieties during the growth periods also showed that starch of varieties H.2304 and M4 had good stability in their solubility characteristics, whereas the others had medium or poor stability.

       Solubility of cassava starch in various non-aqueous solvents has been examined. It was found that maximum solubility was obtained in DMSO and formalin, while in glycerol it was moderate. Starch was insoluble in anisole and methyl cellosolve (Tab.20). The solubility data indicate that starch is more soluble in polar solvents or solvents with affinity towards water. The solubility of starch of yams and aroids pretreated with various chemicals was affected to different extent by the chemicals used and also the concentration. The values were between 18-32%. Solubility of the other tuber starches varied from 10-30% and the aroids had usually lower solubilities.

       The studies on solubility of starch in different solvents opens up the scope for using Phase Transfer Catalysis in derivatisation of starch like esters, ethers and silyl derivatives which can have important industrial applications.

   14. Clarity and Sol stability

       Transparency of starch paste varies tremendously among the different starches. The high clarity has much relevance in food and textile applications. Clarity depends on the associative bonds between the starch molecules in the granules. Cassava starch, having weaker associative forces compared to cereal starches has thus better clarity. When derivatives of cassava starch were compared, acetylated and propylated derivatives had better clarity , though only to a small extent. The clarity was best when pyridine-acetic anhydride was used for acetylation. This reagent gives highest clarity probably due to the high D.S achieved. Esterification tends to weaken the associative forces by reducing the available hydroxyl groups. In contrast, heat moisture treatment reduced the clarity by strengthening the associative forces.

       Dioscorea starches have almost equal clarity as cassava starch, indicating that their associative forces are similar. Variation was not observed among different varieties of Dioscorea alata, D. esculenta and D. rotundata starches. Steam pressure treatment was found to decrease the clarity of D. alata and D. rotundata starches and the reduction was directly proportional to the pressure used and time of treatment. The reduction in clarity on pressure treatment can be attributed to the strengthening of associative forces. The clarity of aroid starches is, as expected, poor and the values are closer to that of cereal starches. The clarity of starch pastes of ten accessions of Colocasia esculenta was nearly equal. Similarly no significant variation among starch of ten different accessions of Amorphophallus paeoniifolius was observed. The clarity of three accessions of Canna edulis was found to be much higher than the aroid starches. Generally the ‘B’ starches appear to have higher clarity compared to ‘A’ starches.

       Sol stability or paste stability reflects the retrogradation tendency of starch pastes. Cassava and sweet potato starches have low retrogradation tendency and therefore high paste stability. The low retrogradation tendency of cassava may be due to the higher weight-average molecular weight of the amylose fraction in cassava. Dioscorea starches also have good stability, while the aroid starches have poor stability. The results also reflect the differences in associative forces among the different tuber starches. The sol stability of cassava starch in various non-aqueous solvents has been studied. The values varied from 3 hours in ethanediol to over 20 days in formalin. Formalin may be preventing parallel association of the starch molecules, especially the amylose chains, by forming complexes. Sol stability was also affected by added surfactants. Sol stability was enhanced by derivatisation also, though there was no correlation between degree of substitution and sol stability. The paste stability of starch of different varieties of D. alata , D. rotundata , D. esculenta, X. sagittifolium, and A. paeoniifolius was nearly same. Heat-moisture treatment (steam pressure treatment) of Dioscorea starches decreased the paste stability and the observed reduction at higher levels of treatment was so high that the starch gel started settling within 2-3 hours, indicating that the starch molecules come so close to each other by the compressive treatment that they associate themselves very easily leading to fast settling.

   15. Digestibility

       Digestibility of starch by enzymes is of importance for evaluating nutritive value and also in industrial applications. Cassava starch is one of the least resistant root starches. In vivo digestibility of different tuber starches was studied on albino rats. The results indicated that in case of raw starch, digestibility of cassava, sweet potato, Colocasia, Xanthosoma and Amorphophallus starches was quite high (65-75%), comparable to corn starch (76%) but that of the Dioscorea was low (15-25%) similar to potato (10%). On cooking, the digestibility was substantially increased for all the starches including yam and potato starches. The large increase in digestibility on cooking can be attributed to the change in starch structure on gelatinization. This was confirmed by the observation that the XRD patterns of all the cooked starches were similar, unlike the uncooked starch in which aroid starches have ‘A’ pattern, and yams posses ‘B’ pattern. A comparison of the digestibility was also carried out using pancreatic a-amylase under in vitro conditions. Under in vitro conditions, maximum activity was obtained on cassava starch while Colocasia suffered breakdown to much lower extent indicating that in vivo and in vitro digestibility cannot be directly compared. The digestibility could not be related either to the amylose content or the soluble amylose contents in these starches. The studies on digestibility of aroid and yam starches extracted from pretreated tubers showed that pre-treatment influenced a-amylase digestibility to a small extent.

       It was desired to find out if added lipases modify the gelatinization characteristics of starches. It is well known that these lipids affect the starch gelatinization to a large extent. The removal of lipids may also be helpful in reducing the lipid flavour associated with cereal starches. Lipase at three different concentrations (viz 1,3 and 5%) was added to a starch suspension in water (1:1) and the pans incubated at 37°C for, 15, 30, 45 and 60 min and DSC run at the following conditions. First heating 15-150 at 5°C/min, second heating upto 130° at the same rate. The Gelatinization temperatures and gelatinization enthalpy starch-lipid complex melting temperatures and enthalpy were calculated using inbuilt software. The results do not indicate any major change in the starch gelatinization temperatures or enthalpy for any of the concentrations examined showing that lipase is unable to break the starch-lipid complex.

   The studies on the different tuber crops reveal the vast variability available among them. Such differences are not observed in case of cereal starches. The high viscosity of cassava and Canna starches makes these starches very useful in many food and industrial applications especially where high thickening power is desired. The low viscosity of aroid starches can be exploited in paper industries, which prefer lower viscosity and good film forming capacity. The small granular size of Colocasia and D. esculenta starches make this ideal as filler in biodegradable plastics, and in aerosols and talcum powders. The clarity of cassava, Canna and yam starches can be very useful in many food applications. Similarly the good gel strength of these starches, especially Canna starch can be utilized in a wide array of food products. The easy gelatinization of cassava and sweet potato starches can make them suitable in manufacture of hydrolytic products derived from starch. The range of characters observed makes the tuber starches amenable to different applications based on their properties in place of chemically modified starches which may have harmful effects. An awareness of their potential uses can help in large-scale cultivation of these crops and extraction of starch from them. It is also possible to modify the starch properties by simple physical methods like hydrothermal or steam pressure treatments. Latest developments in biotechnology can also be tried to modify the starches. These include fermentation of starch by use of selective organisms or enzymatic modifications, which can bring about specific substitutions. Use of enzymes in starch derivatisation has not been exploited and offers very good scope for value addition.