Coarse-grained materials properties for fiber-based materials from computer simulations

Transcription

1 Coarse-grained materials properties for fiber-based materials from computer simulations Mikko Alava Aalto University, Department of Applied Physics, P.O. Box 14100, FIN-00076, FINLAND The understanding of the most important physical qualities of paper and other similar end-products has had some interesting successes based on simulations and modeling over the last 20 years. The recent trends of using nanotechnology to expand product property space, to use novel raw materials such as Nano Fibrillar Cellulose (NFC), and to look for other applications beyond the traditional ones create new challenges. Clearly, the prospects of quantitative modeling are much greater for certain physical quantities than for some others. Transport properties from optical opacity - to barrier ones for gases and liquids demonstrate this clearly, as e.g. the success of simulations of the diffusion of gases and water vapor in porous fiber networks illustrates, compared to similar attempts to understand the permeability to liquids. The structure of paper can be caught well with deposition models starting from the famous KCL-PAKKA. The structure of the solid phase and the pore space are important for all such physical properties, and in particular for the mechanical ones such as strength and elasticity. The prediction of material strength is a fundamental problem in the mechanics of disordered media so the difficulties are not a surprise. However, structural simulations together with physical arguments allow a description of the elasticity starting from fiber-level properties, and also to have a well-founded theory of strength and fracture toughness on the continuum level. The next generation of fiber-material based products will depend on the fine-tuned properties and novel openings that derive from the applications of NFC. This can take various forms over a wide range of possibilities as regards the detailed application of NFC in various contexts and the process details and conditions that apply. The big challenge that this presents is due to a number of reasons: the unknown detailed properties of NFC in particular as compared to usual fibers in papermaking and the reduced dimensions of the NFC particles. Moreover, one is no longer concerned with standard paper-making so the processes and their conditions may vary in a large fashion. At the Complex Systems and Materials group (Aalto, Department of Applied Physics and COMP Center of Excellence in Computational Nanoscience) we have started a firstprinciples research activity about the wet-end properties of NFC suspensions. There are (at least) three fundamental issues: i) how to combine NFC with other suspension constituents added for various reasons ranging from fibers to filler particles to additives as polyelectrolytes and other flocculation and de-watering agents, ii) how to control the formation of flocs and meso-scale structures out of the primary particles and iii) how in general to understand the physical properties of multiphase flows in transient conditions.

2 The understanding of these aspects is necessary for instance because of the consequences for upscaling of laboratory experiments to pilot scale and beyond, and a combination of modeling and experiment becomes thus necessary. The aggregation of particle suspensions can be simulated by various means, including discrete (suspension) particle techniques. We have resorted to using the so called Population Balance Equation (PBE) method. The method is based on the grounds of the idea of writing down an equation, which describes the mass balance in the aggregating system. On a practical level, one breaks up the complicated equation into a set of first order differential equations, which describe the birth rate of aggregates of certain mass due to the collisions of other compatible massed aggregates. To that one adds a description of the flow geometry (e.g. pipe flow) by applying aggregation and breakage rates that derive from the situation at hand. In a shear flow, it can be found, on the basis of elementary geometry and physics, that the aggregation rates depend on the velocity gradient of the flow field. Figure 1 (A. Puisto et al.). illustrates the usage of the PBE formalism to the computation of the viscosity of a NFC/water suspension. The thixotropic/shear-thinning behavior of these suspensions can be quantitatively caught by choosing the parameters needed partly by using measurement and partly by applying recipes from the theory of particulate suspension flows. In the case at hand, the solids content is 0.25 %, and initial NFC particles have a dimension of 43 nm (radius of gyration). Meanwhile, estimates are plugged in for the cohesive force (8.5 un), aggregate fractal dimension (2.6), and the gel point in dry solids content (20 %). The behavior of NFC suspension viscosity vs. shear rate.

3 The currently pursued research directions include a number of different scenarios. It is of interest to study the transient flows of NFC suspensions, so we are developing a multiscale-modeling approach using fluid flow solvers coupled to the PBE technique. This will allow to follow the viscosity profile of a flowing suspension as a function of the microscopic NFC parameters, while the wet end chemistry can also be tuned. Another approach is to consider the physics of multiscale aggregation processes, which will be of relevance in the case of pulp fiber-nfc suspensions.

4 Materials properties from computer simulations Presented by: Mikko Alava Aalto Univ., Applied Physics Actual work done by: Antti Puisto, Xavier Illa, and Mikael Mohtaschemi, Mikko Niemi

5 Why use a computer? Nanotechnology in action: from the nano to the meters understand the various scales and their interactions. Avoid the laboratory, with simulations and a well-thought model gain more insight. Predictive modelling (elasticity, other transport properties). Advances in software and hardware.

6 Computer Modeling - Multiscale Modeling Methods for Cellulose Structure and Aggregation Five talks (myself and ) Stoyanov: Solvation Structure and Thermodynamics of Chemically Modified NC Vattulainen: Perspective to Cellulose Nanofibrils through Atomistic Simulations Vidal: Predicting Self-Assembled Nano-Cellulose Fibre Structures Bergenstråhle: Cellulose Crystal Structure and Forcefields

7 Coarse-grained properties? Add nanomaterials to fiber products: I. Influence and improve the current products/materials (paper, board) II. New scenarios (no fibers, or only NFC) Important quantities to be understood: structure, its formation, mechanical (elasticity, strength) properties, barrier properties (gases/liquids, electrical ones), interactions with fluids Here: rheology and first steps of structure formation: aggregation/flocculation, retention.

8 Challenges in rheology Practical question: what is the solids content that can be used in the presence of nano-sized (NFC) components? This relates to fundamental issues such as: gelation transiton, shear-thinning (or thickening) behavior, thixotropy. Paper-making suspensions exhibit flocculation and aggregation and can be controlled by wetend agents (polymers). But: at the end, retention and water-removal

9 NFC rheology NFC and a solvent/liquid: thixotropic or shearthinning suspension. In other words, the viscosity can be affected with shearing (thixotrophy?) What is the role of NFC chemistry and size? M. Bercea, P. Navard, Macromolecules 33, 6011 (2000).

10 NFC (an example) What is NFC? How does it depend on the preparation route? How to characterize NFC, the particle geometry, and chemistry? What is the relevance of such quantities?

11 Rheology for NFC + water Viscosity vs. shear rate vs. concentration. This kind of data related to process engineering: solids content, NFC used. Here, no Newtonian regime (flat). Reliability of rheological measurements? (Shear band formation ) M. Pääkkö et al, Biomacromolecules 8, 1934 (2007)

12 How to model suspensions? Particle-based (fibers in water), multiphase-flow methods. In this work: we take a mesoscopic approach: use rate equations or population balance equations (Smoluchowski) to follow the solid phase dynamics. Aggregate size distribution couples to physics: fluid viscosity changed ( immobile liquid phase ); structural parameter.

13 PBE modelling Write down balance equations that include the elementary processes: aggregation, fragmentation. Basic physics: particle attraction, floc destruction by shear forces.

14 Latex particles Test system: weakly attractive latex particles in water. Can be used to test the various assumptions that go into PBE models. Numerical treatment of PBE systems?

15 Rheology modelling Aggregtion modeling of NFC + water % solids content. No other additives (another problem).

16 Interactions w/wet end agents Various mechanisms: briging by high molecular weight polymers, changing the charge/surface potential. Summed in PBE kernels.

17 Thixotropic suspension models An example of the coupling of aggregate size distribution and shear force (velocity) history. Transients appear to the suspension viscosity. Thixotropy! NFC

18 Multiscale models Schematics for pipe flow Study the consequences for suspensions. One example: pipe flow.

19 Dynamic flow in a pipe Interesting physics already for non-attractive suspensions. Plug formation at a rate which depends on the aggregation dynamics.

20 Multi-species modelling Schematics of 2-species aggregation. Various rates ( kernels ) for the species and intraspecies. Central idea: NFC (+ chemicals) interacts with fibers, fillers. Size disparity: difficult to model otherwise at all. The results will influence flocculation, and retention.

21 Two examples Aggregation by mostly small (NFC) particles. Volumefractiondictated by aggregation and by the swelling due to fractality. Effect of enhanced aggregation: composite aggregates (NFC and fillers) dominate.

22 Challenges to experiment Properties of the constituent materials in wet-end conditions. NFC: a nanomaterial. Rheology measurements in transient conditions. Theoretically, there are issues that are by far not resolved.

23 Summary The control and prediction of materials properties from using novel nanomaterials needs a combination of models. Several basic physics/chemistry problems need to be tackled, due to the complexity of the scenarios and the wide range of spatiotemporal scales. Experiment is often lacking!

24 Acknowledgements This research has been partly funded by Reengineering Paper project belonging to the Intelligent and Resource Efficient Production Technologies (EffTech) program of the Forestcluster Ltd and also partly the EU FP7 consortium Sunpap The Academy of Finland Center of Excellence funding is thanked for. A number of collaborators are gratefully acknowledged (in particular VTT, and TUT).