Previous PageTable Of ContentsNext Page

Detection of floating gluten susceptibility in wheat flour for starch/gluten manufacture

N.J. May1,2, I.L. Batey1,3, D. Miskelly1,2, J. Smit4

1 Value Added Wheat CRC Ltd, North Ryde, NSW, 1670, Australia
Allied Mills Pty Ltd, PO Box 1, Summer Hill, NSW, 2130, Australia
Food Science Australia, North Ryde, NSW, 1670, Australia
Penford Australia Ltd, Tamworth, NSW, 2340, Australia


Approximately 22% of wheat flour milled in Australia is used in the commercial manufacture of starch and gluten. Wheat flour dough may be separated into an insoluble gluten fraction and aqueous starch fraction by washing. One of the most widely used starch and gluten extraction processes from a flour starting material is the Martin process. This process involves mixing flour and water to a low viscosity dough, separating gluten from starch by washing with dilute sodium chloride solution in a screw conveyor then ring drying gluten to ~80% protein at ~60C and grinding to 212m particle size. Starch leaves the gluten extractors and is purified through a series of screens and centrifuges, filtered, dried and sieved. Floating gluten, as it arises, leaves the gluten extractors with the aqueous starch fraction and, if in excess, blocks starch purification screens resulting in screen blinding and starch and gluten overflow loss. It is believed to be a GxE effect and not necessarily a processing parameter. The development of a lab scale test to characterise flour for floating gluten susceptibility will potentially lead to better selection of wheat varieties for starch/gluten manufacture with minimal processing difficulty and yield loss.

Materials and methods

Sample Composition and Preparation

Flour grist samples from 2003/04 were commercially milled by Allied Mills, Tamworth. Samples comprised hard wheat from 17 northern NSW grain sites, in 3 grades, which were combined into a mixture of 4 to 6 varieties per sample. Of the sample set, 5 were positive and 20 negative for floating gluten susceptibility as determined in commercial starch/gluten modified Martin separation at Penford, Tamworth. Flour protein of the samples was 11.7-13.3% (mean 12.7%) and water absorption 65-68%.

For the wholemeal rapid test, single variety northern NSW hard wheat was ground on a Perten 3100 Laboratory mill with a 0.8mm screen. 450g ground samples were subsequently sieved with a Simon Laboratory plan sifter for 8 minutes through a 150m nylon screen. Throughs were retained as samples and overs, which physically interrupt the gluten matrix, were discarded.

Instrumentation and Methodology

For measurement of floating gluten susceptibility, a Brabender Farinograph with 300g Z-arm bowl attachment was used to give an indication of the likely degree of gluten break-up in commercial separation from starch. Various modifications made to the standard Farinograph method (AACC 54-21) are described.

The mixing bowl temperature was set at 37C and commercially milled flour grist (300g) was transferred to a 300g mixing bowl. The dispensing burette was filled with 37C (+/- 0.5C) water. Flour was then dry mixed on speed 1 (63 rpm) for 1 minute and water added to achieve a water absorption of 400BU (+/- 20BU). After dough development time was attained, the dough was mixed for a further 1.8 minutes (= 3 squares) to stabilise the dough. Mixing was then stopped and water (500mL, 37C) added to maximally fill the bowl without causing spillage. Mixing was resumed for 0.9 minutes (= 1.5 squares) in which time a glass funnel attached to a 0.8mm PVC hose was affixed to the centre bowl opening. At the same time a plastic hose hand pump was inserted into the left side of bowl lid, just immersed in the starch liquor. While mixing, water (2L, 37C) was steadily poured through the centre funnel while pumping out starch liquor with a hand pump into a 2-litre beaker. The flow rate of wash water was 500mL/min. Pumped out starch liquor including floating gluten was then poured through a 180m nylon screen attached to a beaker. Wet gluten overs are weighed as floating gluten from 300g flour. The % floating gluten overs can then be calculated thus: (Wet gluten overs (g) / 300g flour) x 100.

Figure 1. Continuous water addition/removal (left) and typical Farinograph trace (right) showing peak development at 400BU, dough stabilisation and gluten washing stages

Results and discussion

Method Development

A novel, small-scale test to detect the presence of, and quantify floating gluten in commercial starch/gluten separation was developed on a Farinograph and optimised to give best differentiation between main gluten and floating gluten fractions. The test is based on ranking samples for degree of gluten break-up and is not quantitatively related to commercial gluten losses. The test was thus developed by predicting gluten break-up on the Farinograph with commercial flours positive and negative for floating gluten development. Of the five samples positive for floating gluten commercially, all had a % of floating gluten overs >3.1% per 300g flour. Of the 20 samples negative for floating gluten, all had a % of floating gluten <2.0% per 300g flour. This was determined to be the tolerance of the test. Mean results indicated that there was an even greater average difference between the two groups of samples. Floating gluten samples contained 5.3% mean gluten overs whereas non-floating gluten samples contained 1.2%.

Method Optimisation

The test as described was developed after optimisation of the process variables. The water to flour ratio (0.62-0.65) was selected to attain peak development at 400BU which is a lower viscosity dough than typically used in the standard Farinograph test but more closely represents commercial practice. Mixing speed 1 was chosen by default, as speed 2 gave more spillage problems and less floating gluten break-up differentiation as most of the main gluten piece would disintegrate. A shorter test duration (8 minutes) was selected as longer time sampling did not show any greater differentiation between floating gluten and main gluten break-up. The bowl and water temperature of 37C was selected after trialing various temperatures. At 20C gluten break-up was less evident and starch separation was inhibited whereas at 40C water evaporation was too rapid during water addition and the dough excessively softened. A large wash water volume (2 litres) was necessary to initiate starch washing, which effectively allows gluten to break-up and allow enough time to sample all floating gluten. Sodium chloride wash concentrations (0-2%) were also examined, however this hindered floating gluten development possibly due to an improvement of the extensibility and cohesiveness of partially washed gluten.

Method Validation

The method was validated for repeatability and reproducibility in order to estimate the uncertainty of measurements resulting from random errors. Repeat analyses (7) of a single variety (Sunstate) commercially milled flour, were conducted at a single time (Figure 2). The % relative standard deviation (% standard deviation/mean of replicate analyses) was 15.5. This error estimate is reasonably large indicating strict adherence to test conditions is required. Reproducibility of the test over 3 days was similar to repeatability indicating little change in error can be expected over time.

Figure 2. Floating gluten Farinograph method - repeatability and reproducibility

Rapid Wholemeal Test

Following the development of a functional test for flour, a similar test for wholemeal was devised that would offer advantages as a screening test before affected grain reaches milling. The wholemeal test was developed on 4 hard wheat varieties and results compared to matching Quadrumat milled flours (65% extraction). Wholemeal samples (450g) were plan sifted (150m nylon) for 10 minutes and overs discarded to remove large bran particles which physically interrupt the gluten matrix. Changes to the flour method were made by lowering the bowl and water temperature to 20C and shortening mixing and washing times to limit dough break-up.

Results for the wholemeal samples compared to flour demonstrate that the wholemeal test ranks samples similarly to the flour test, however in greater magnitude (Figure 3). Also, the wholemeal slope provides greater differentiation between samples than the test for flour.

Figure 3. Comparison of wholemeal and flour tests – % floating gluten overs

Rheological Properties and Microstructure

The rheological properties of floating gluten were also studied to determine characteristics of gluten that contribute to floating gluten susceptibility. Wet gluten was produced from floating and non-floating gluten grist flours using a Glutomatic with 0.5% NaCl solution and TA-XT2 Keiffer Rig setup. Floating gluten flours were consistently associated with higher Rmax values (r=0.67**) and moderate extensibility (r=0.59**). Similar results were obtained from an Extensograph. Floating gluten separated using the Farinograph method also had higher Rmax and extensibility than the main gluten mass. This greater resistance may be related to dough mixing characteristics such that stronger dough/gluten is too tough and breaks up during washing.

Floating and main piece gluten were stained with Ponceau 2R to observe protein structure and starch granule distribution that may affect rheological results. The floating gluten sample showed a less continuous protein structure and larger voids than main gluten (Figure 4). This is indicative of its broken up nature, showing interruptions within gluten pieces.

Figure 4. Light micrographs (10x magnification) of floating gluten (left) and main gluten (right)


The proposed test is capable of ranking samples for degree of gluten break-up in commercial separation. It is also adaptable to wholemeal samples, simple and quick to perform (test duration is 8 minutes), accurate (results were compared to 25 reference samples), repeatable, reproducible and economical as it utilises existing instrumentation.


American Association of Cereal Chemists, Approved Methods, (2000) 10th Ed. Vol 2.

Previous PageTop Of PageNext Page