Article is available in full to IFST members and subscribers.

Register on the FST Journal website for free

Click the button to register to FST Journal online for free and gain access to the latest news


If you are an IFST member, please login through the Members Area of the IFST website. 













Novel structuring routes with cellulose

Tim Foster of the University of Nottingham discusses the interactions of cellulose with food polymers and the beneficial properties of the cellulose/polymer blends for application in the food industry.

Research at the University of Nottingham over the past decade has been inspired by a knowledge of food industry needs combined with learning from parallel industries, such as paper/pulp and textiles. For some time the food industry has been calling for ‘natural’ ‘minimally processed’ ingredients. Cellulose is the most abundant polymer on planet earth and is the major structural material in the composite structures which comprise the cell walls of land and ocean plants; it is also exuded by some bacteria. The paper/pulp industry has functionalised cellulose from such sources through innovations in processing, e.g. steam explosion or homogenisation, to create fibrous material to enhance structuring properties. The textile industry has pioneered the ‘regeneration’ of cellulose into fibres and sheets through the use of solvent/anti-solvent technologies, which are clean and green, and create pharmaceutical grade materials. The question therefore is can the food industry benefit from such insights and enable the use of a material that is ordinarily insoluble and inert?

The obvious outcome would be to use cellulose to replace the structuring capabilities of highly refined hydrocolloids, starches and fats used in the industry in products ranging from high to low moisture content. However, additional benefits might also be acquired, such as the use of these materials for nutritional enhancement as dietary fibre, where they can be classified as non-fermentable polymers, therefore imparting structure with low or zero calorific loading allowing a reduction in the energy density of foods. A recent and important development in the food industry, and other fast moving consumer goods industries, is the need to decrease the use of plastic packaging. Therefore the use of ‘biomass’ for such applications is also becoming increasingly attractive.

Derivatised celluloses, (carboxymethylcellulose (CMC), methylcellulose (MC) and hydroxypropylmethylcellulose (HPMC)), microcrystalline cellulose (MCC), labelled as E466, E461, E464 and E460(i), cellulose gum and cellulose gel, have traditionally been used in the food industry as, for example, thickening agents and fillers1. Work has been carried out to investigate the functionality of cellulose derivatives, which are known to be ‘blocky’ in structure, imparting varying degrees of solubility and hydrophobicity2. Highly hydrophobic derivatives are insoluble in water, but soluble in more non-polar solvents, and the soluble and food grade versions are able to gel upon heating. The more hydrophobic (and blocky) derivatives can be induced to gel at body temperature.

Natural’ ingredients

More recently, there has been a focus on the creation of alternative cellulose-based ingredients as ‘natural’ particles. Citrus fibres (in a number of different forms – from different suppliers, e.g. Fiberstar’s Citri-fi or Herbafood’s Herbacel AQ+) and powdered cellulose (E460(ii)) fibres (e.g. Solka-Floc from International Fiber Corporation and Borregaard’s SenseFi) are examples of currently available materials. They can be used as they are, or modified further, if required. SenseFi products are microfibrillated, which makes them particulate and interactive with a high water binding capacity3,4, whereas Solka-Floc materials come with different aspect ratios providing the opportunity to build food microstructures with different textural attributes. Herbacel AQ+ swells when hydrated and its particles can be functionalised further if the aqueous dispersion is homogenised. All such materials can also be modified in a ball mill, creating varying ratios of amorphous to crystalline morphologies. The advantage of creating amorphous cellulose, is that it then recrystallises upon hydration5,6,7.

In order to make use of cellulose, either in its isolated form, or as a part of natural hierarchical assemblies, a detailed and fundamental understanding of its molecular and supramolecular structures, how they are influenced by typical or novel processing steps and how they interact with other hydrocolloids typically used in industry is required8. Aspects of thermodynamics and kinetics are therefore worth consideration, in line with the more conventional thinking behind product formulation.

Controlling cellulosic interactions with other polymers

It is now well known that mixtures of food polymers can ‘interact’ in different ways during formulation. They may specifically bind to one another or be repulsive to one another forming separate phases. Such concepts can be utilised to explain the functionality of cellulosic materials and to assist in formulation. Specific interactions, sometimes known as ‘synergistic interactions’, often have their origin in the molecular structures related to polysaccharides with β-1,4-glycosidic linkages, where cellulose is a β-1,4-glucan. Indeed it is known that the hemicelluloses (e.g. xyloglucans, galactoglucomannans and arabinoxylans) influence the properties of the plant cell wall through direct binding to the cellulose micro- and macrofibrils.

Using this approach in formulation could provide new and additional functional properties, such as enhanced thickening or gelation within food product microstructures. To this end it has been shown that locust bean gum (a β-1,4-mannan) can both protect cellulose crystallinity during ball milling and subsequently enhance the rate of recrystallisation upon rehydration6. This concept has recently been exploited in the creation of cellulose based 3D structures7 utilising the capability of the binder jetting process to control temperature and moisture content, such that the rates of recrystallisation are sufficient to withstand the pressures exerted on the structure as layer-by-layer the macrostructure is formed5,7. CAD-designed macrostructures containing xanthan gum (a β-1,4-glucan), enabled by bespoke design of the food grade ‘ink’, can therefore be built (Figure 1).  A benefit of this approach, all-be-it still a niche technology, is the potential to build complex micro- and macrostructures using this low calorific food material. In a world striving for low energy density foods, to begin to combat the obesity epidemic, such enabling technologies are attractive and just need the enthusiasm and commitment of industry to make them happen.

Figure 1 Star shapes created using binder jetting technology using a 9:1 cellulose/Konjac Glucomannan co-ball milled powder and an ink formulation containing 1%wt milled Xanthan Gum.

Specific interactions, however, are still hotly debated in academic circles and therefore development work using advanced technologies, such as NMR (nuclear magnetic resonance), DSC (differential scanning calorimetry) and rheological testing, is underway to provide evidence for such interactions. There has been development in the fluorescent tagging of polysaccharides and, when visualised using confocal laser scanning microscopy or fluorescence microscopy, their location relative to the cellulose fibres can be confirmed. Figure 2 shows the location of xyloglucan and psyllium, both covalently tagged with fluorescein isothiocyanate (FTIC), interacting with Solka-Floc 900. Xyloglucan coats the surface of Solka-Floc 900, and the particulate nature of psyllium attaching to fibrillated Solka-Floc 900 (non-covalently stained with methyl blue) is apparent. The results from the work with xyloglucan, galactomannans and glucomannan suggest that changes in polysaccharide structure, influenced by solvent type, affect their interaction with cellulose9. The nature of such interactions may provide structural elements in e.g. replacing gluten networks in baked products, or high energy density starch or fat.

Figure 2 Fluorescein sothiocyanate tagged xyloglucan (top) and psyllium (bottom) interacting with Solka-Floc
900 (fibrillated; bottom), using confocal laser scanning microscopy (top) and
fluorescence microscopy (bottom).
Scales bars are 75µm (top) and 400µm (bottom).

The inert nature of cellulose suggests that it would act as a dispersed phase in a continuous polymer matrix and the properties of the mixture would be dependent upon the dispersed phase volume of the cellulose. Figure 3 shows the efficiency of different types of cellulose in providing effective volume occupancy, with the low aspect ratios of MCC and Solka-Floc 300 requiring much higher concentrations (by weight) than fibrillated cellulose.

Figure 3 Comparative concentrations of microcrystalline cellulose (MCC), ball milled cellulose (BM), Solka-Floc 300 (S300) and Solka-Floc 900 (S900) and microfibrillated cellulose (MFC) required to raise the dispersed phase volume of a cellulose dispersion.

The laws of colloid science apply to systems in which it is known that polymers may cause depletion flocculation of a dispersed phase. In a recent publication, mixtures of bacterial cellulose and maize starch were shown to occupy separate phases10. However, there are some unexplained effects of cellulose on starch gelatinisation, which also tends to suggest a ‘specific interaction’ between the two glucans (starch being an α-1,4-glucan). Such a specific interaction is not necessarily expected from a molecular dynamics perspective, but does begin to raise questions regarding helical polymers (e.g. starch) interacting with β-1,4- extended ribbon-like molecular structures.

The examples provided so far relate to the use of isolated wood cellulose. However, in the case of a more natural hierarchical assembly, citrus fibres are a good example. They too are dispersions and when they are mixed with xanthan gum, some interesting phenomena begin to emerge. In mixtures which have the same final concentration of each component, but are attained by mixing different ratios of given concentrations of stock solutions and dispersions, enhancements in viscosity beyond expectation are witnessed. The zero-shear viscosities of the xanthan and citrus fibre AQ+ (CFAQ+; Herbafoods) mixtures are plotted in Figure 4. The value of 0% xanthan in the mixture corresponds to 100% CFAQ+ at the final mixture concentration of 1.1%, while the 100% xanthan corresponds to the final mixture concentration of xanthan of 0.4% for low pyruvate xanthan (LP) and 0.1% for high pyruvate xanthan (HP). All the mixtures show a higher zero-shear viscosity than the xanthan or CFAQ+ alone, with the mixtures using 75% of the xanthan stock solution exhibiting significantly (p<0.05) higher viscosities. These mixtures are, by design, attained using higher CFAQ+ concentrations in the starting stock dispersions, which seem to affect the competition for water; separate phase generation could be the root cause of the viscosity increase. This appears to be impacted by the molecular structure of the xanthan gum, which is also known to affect the efficacy of ‘specific interactions’ with other β-1,4-linked glycans.

To investigate this further (Table 1) other dispersions were used, at matched disperse phase volume, and generated at the same 75:25 mixing ratio of stock xanthan solution and particulate dispersion. What is evident is that describing this phenomenon is not straightforward: the hierarchical structures of citrus and apple fibre are effective at enhancing mix viscosities compared with xanthan alone; ‘specific interactions’ might be at play given the enhancement seen by the ‘inert’ celluloses of Solka-Floc and MCC; and the swellable apple and citrus fibres are more effective than cold water swellable starch in impacting competition for water. This therefore requires more in-depth, fundamental understanding, but shows that these types of mixtures can have synergistic effects in high moisture food products and may go some way to increasing the textural control of low moisture products, where competition for water in formulation and processing (water removal at high temperatures or during freezing) will have significant impact.

Table 1 Zero-shear viscosity of mixtures of stock solutions / dispersions of low pyruvate (LP) and high pyruvate (HP) xanthan with citrus Fibre AQ+
(CFAQ+), apple fibre (AFC) (both Herbafoods), Solka-Floc 900 (SF 900;
International Fiber Corporation), microcrystalline cellulose (MCC; Sigma)
and cold water swelling starch (UT2; UltraTex 2, Ingredion).

Surpluses of the agricultural and food industries offer potential as sources of materials for food structuring applications.

Valorisation: utilising natural hierarchical structures

The work highlighted above has used commercially available cellulosic materials, but there is increasing interest in utilising materials, which are often termed ‘waste’ (or co-products / by-products) from either the field or factories. Surpluses of the agricultural and food industries offer potential as sources of materials for food structuring applications (Table 2).


A project involving Nottingham has recently been completed, which investigated waste flow modelling and environmental analysis through LCA (life cycle analysis) to enable socio-techno-economic evaluation of these surpluses in terms of cost-benefit analysis, risk analysis, systemic barriers to exploiting the materials and the gap between concept and final implementation. This work is bound by confidentiality at present, but will be the subject of future publications. Progress, however, can be reviewed in conference presentations and the project’s annual reports11 and will soon be summarised in the Routledge Handbook of Food Waste in a chapter entitled ‘Upcycling and valorisation of food waste’.

Rheological measurements indicated that all cellulosic suspensions showed viscoelastic gel-like behaviour, however the non-fibrillated particulate structure of citrus fibres released the tastant more effectively and faster.

Potential applications

Taste perception

Two different citrus fibres and fibrillated spruce fibres have been compared in high moisture dispersions to assess the impact of their structures on taste perception (saltiness)5. As mentioned previously, the fibrillated samples perform as ‘interactive’ particles, producing a highly entangled network structure, which lowers the taste perception compared to particulate suspensions of the citrus fibres. Rheological measurements indicated that all cellulosic suspensions showed viscoelastic gel-like behaviour, however the non-fibrillated particulate structure of citrus fibres released the tastant more effectively and faster. Therefore a highly entangled network microstructure of cellulosic fibres, responsible for higher water retention capacity may be beneficial for certain structural and nutritional aspects of food products, but if taste release is of importance for increased sensory perception, then a more particulate non-interacting material would be preferred.

Biodegradable packaging

An alternative application of cellulose fibres is their use in biodegradable packaging. One way in which this might be realised is to borrow technology used in the textile industry. Mixtures of food polymers and cellulose have been studied to trial such possibilities using the Lyocell technology based on NMMO (N-Methylmorpholine N-oxide) dissolution of cellulose, and regeneration into aqueous solvents. Fibres and film production was possible and Figure 5 shows the incorporation of a fluorescently tagged konjac glucomannan into a regenerated cellulose fibre. Knowledge gained on dissolution of starch in NMMO was applied to make solutions of cellulose and starch together in NMMO in different cellulose to starch ratios using maize starch with varying amylose and amylopectin content. Compared to cellulose films, the blend films had better water retention behaviour. Important fibre properties, like tenacity, elongation and elastic module value, were higher for cellulose fibres than fibres made from blend solutions; fibres made from high amylose starch and cellulose blends showed the most similar properties to cellulose only films.

Figure 5 Incorporation of fluorescently tagged konjac glucomannan into regenerated cellulose fibres

Tim Foster, Professor of Food Structure

Division of Food, Nutrition and Dietetics, University of Nottingham, Sutton Bonington Campus, Loughborough, Leicestershire, LE12 5RD.

Email Tim.foster@nottingham.

At the University of Nottingham Tim continues to progress the area of polysaccharide research. He is indebted to the researchers he has had the privilege to work with: Antonio Sullo, Fuad Hajji, Amir Abbaszadheh, Nagamani Koganti, Mita Lad, Marie Janin, Charles Winkworth-Smith, Deepa Agarwal, Sabrina Paes, Shaomin Sun, Sonia Holland, Jade Phillips, Yi Ren, Avinash Manian, Barbora Siroka, Roger Ibbett, Bill MacNaughtan, Paulo Diaz-Calderon and Zainuddin Umar.

This research has been undertaken at University of Nottingham and includes work sponsored and supported by EPSRC (Centre for Innovative Manufacturing in Food, CDT in Additive Manufacturing, Circular Economy: ‘Whole systems understanding of unavoidable food supply chain wastes for re-nutrition’ and Case Awards), EU (EPNOE – the European Polysaccharide Network of Excellence and the ITN Shaping and Transformation in the Engineering of Polysaccharides (STEP ITN)), The Newton Fund (Newton-Bhabha ‘NewTrition: ReNEWable, sustainable nuTRITION’), Innovate UK (Food Processing and Manufacturing Efficiency ‘Transforming wet perishable food waste streams for high value human consumption’), Oslofjordfond grant scholarship, Industry funding and kind donations from University of Nottingham (Vice Chancellor’s Scholarships and University funded PhDs).


1 Wustenberg T. “Cellulose and Cellulose Derivatives in the Food Industry: Fundamentals and Applications”, Wiley-VCH, (2015).

2 Sullo A., Wang Y., Koschella A., Heinze T and Foster T. “Self-association of novel mixed 3-mono-O-alkyl cellulose: Effect of the hydrophobic moieties ratio”, Carbohydrate Polymers, 93(2), 574-581, (2013).

3 Agarwal D., MacNaughtan W. and Foster T. “Interactions between microfibrillar cellulose and carboxymethyl cellulose in an aqueous suspension”, Carbohydrate Polymers, 185, 112-119, (2018).

4 Agarwal D., Hewson L. and Foster T. “A comparison of the sensory and rheological properties of different cellulosic fibres for food”, Food and Function, 9, 1144-1151, (2018).

5 Paes S., Sun S., MacNaughtan W., Ibbett R., Ganster J., Foster T. and Mitchell J. “The glass transition and crystallization of ball milled cellulose”, Cellulose, 17(4), 693-709, (2010).

6 Abbaszadeh A., MacNaughtan W. and Foster T. “The effect of ball milling and rehydration on powdered mixtures of hydrocolloids”, Carbohydrate Polymers, 102, 978-985, (2014).

7 Holland S., Tuck C. and Foster T. “Selective recrystallization of cellulose composite powders and microstructure creation through 3D binder jetting”, Carbohydrate Polymers, 200, 229-238, (2018).

8 Foster T. “Natural structuring with cell wall materials”, Food Hydrocolloids, 25(8), 1828-1832, (2011).

9 Winkworth-Smith C., MacNaughtan W. and Foster t. “Polysaccharide structures and interactions in a lithium chloride/urea/water solvent”, Carbohydrate Polymers, 149, 231-241, (2016).

10 Diaz-Calderon P., Macnaughtan W., Hill S., Foster T., Enrione J. and Mitchell J. “Changes in gelatinisation and pasting properties of various starches (wheat, maize and waxy maize) by the addition of bacterial cellulose fibrils”, Food Hydrocolloids, 80, 274-280, (2018).




View the latest digital issue of FS&T or browse the archive


Click here

Become a member of the Institute of Food Science and Technology