Changes in starch molecular structure during pasting

J. Doutch (1), E.P. Gilbert (1) and M.L. Bason (2)

(1) Bragg Institute, Australian Nuclear Science and Technology Organisation, Locked Bag 2001, Kirrawee DC NSW 2232, Australia
(2) Perten Instruments Australia, PO Box 70, North Ryde BC, NSW 1670, Australia

Starch pasting rheology plays a critical role in determining the processing and end-use properties of many foods and industrial products like paper and biofuels. However the corresponding changes in starch structure at the nanometre scale are not well understood. Small-angle neutron scattering (SANS) is well suited for studying the structure of polymer systems like starch at this scale (Figure 1a) because of its high penetration through dense media or sample environments, limited radiation damage and real time analysis during modification. Rapid visco analysis is a well-established technique for studying starch pasting properties (Figure 1b). In this study we report results of simultaneous SANS and RVA relating structural to functional changes in starch during cooking. 

Materials and Methods

An RVA and paddle were modified to provide a suitable neutron path through the sample and cooking chamber, and fitted to the Quokka SANS instrument at ANSTO Australia (Figure 2) running at wavelength (λ) = 5.078Å. 3.00g of six commercial starches (Table 1) were each added to 47.8g of D2O and pasted in the RVA using the AACC method 76-21 ‘standard 1’ profile modified to start and end at 25oC. Neutron scattering (q = 0.018 – 0.200 Å-1) and viscosity data were collected simultaneously during the 13-minute starch pasting tests. The scattering vector q = (4π/λ) sin θ where 2θ is the scattering angle.

Scattering data from each test was fitted with power law (0–4 min.) and fractal (5–13 min.) models

where I(q) is the scattering intensity, A is the power law prefactor, δ is the power law decay, I0 is the height of the peak, B is the half width at half maximum and β is the background for the power law model; and Df is the fractal dimension, ξ is the correlation length, R0 is the average radius of the building block particles, and P(q) is the form factor scattering for the fractal model (Doutch et al. 2012).

Results and Discussion

Combined SANS and viscosity data showed a marked shift in the q intensity spectrum before and after viscosity development (Figure 3a). Prior to viscosity onset, the well-known lamellar scattering pattern of raw starch was observed (Figure 3b) with fitted power-law values around -2 and lamellar repeat spacing of 90–100Å.

Subsequent pasting led to a sudden loss of lamellar structure and a dramatic increase in scattering at low q values indicating progressive formation of a network of larger aggregates (Figure 3c). Modelling of this phase (Figure 3d) indicated network building blocks of ca. 1nm forming fractal structures with dimensions consistent with distorted chain structures and occupying a sample volume fraction of approx. 4.4%.

Potato, tapioca and waxy maize, which gave the highest viscosities, appeared to form more extensive but less complex structures (Df values nearer to 2, Table 1). By contrast, regular cereal starches apparently formed more complex and polydisperse aggregates but smaller networks with possible phase discontinuities due to granule ghosts restricting viscosity development.


Simultaneous SANS/RVA indicate loss of starch lamellar structure with gelatinization abruptly followed by the progressive formation of complex fractal structures during starch pasting and retrogradation. Significant botanic variation was evident with less complex but more extensive networks correlating with higher viscosity starches.

Table 1. Fractal dimensions of starch pastes during cooking.

Time 5 6 7 8 9 10 11 12 13
Waxy 2.28 2.29 2.24 2.32 2.27 2.40 2.31 2.28 2.31
Maize 2.82 2.80 2.68 2.81 2.88 2.88 2.85 2.78 2.71
Wheat 2.78 2.81 2.73 2.69 2.86 2.90 2.76 2.78 2.79
Potato (a) 2.11 2.15 2.15 2.12 2.18 2.19 2.21 2.21 2.26
Tapioca (a) 2.04 2.19 2.12 2.15 2.10 2.14 2.19 2.21 2.26
Acid 2.65 2.68 2.59 2.89 2.93 2.92 2.95 2.80 2.78


Figure 1. Granule structure (a) and pasting curve (b) of starch.

Figure 2. ANSTO Quokka SANS facility showing the full beam-line instrument (a), RVA (b), detector (c), sample scatter (d) and principle of measurement (e).

Figure 3. SANS data of waxy maize starch during RVA showing viscosity overlay (a), progressive time scans (b), 2-minute scan fitted with power-law model (c), and 6-minute scan fitted with fractal model (d).


Doutch, J., Bason, M.L., Franceschini, F., James, K., Clowes, D, and Gilbert, E.P. 2012. Carbohydrate Polymers 88: 1061–71.


The authors with to thank Ferdi Franceschini, Kevin James and Douglas Clowes for technical assistance and NSW Industry & Investment for partially funding this work.