Elsevier

Powder Technology

Volume 350, 15 May 2019, Pages 36-42
Powder Technology

Development and application of a novel cake strength tester

https://doi.org/10.1016/j.powtec.2019.03.024Get rights and content

Highlights

  • A new caking tester has been developed to measure powder plastic caking.

  • Tester could distinguish caking strength between experimental conditions.

  • The tester overcomes the limitation of traditional uniaxial caking tester.

  • A statistical model has been developed.

  • Proposed approach could be used to determine the suitable storage condition.

Abstract

Caking can cause many problems in industries during processing or storage of particulate materials. Caking magnitude depends on several factors, for instance temperature, consolidation stress and storage time. In this research paper, a novel force displacement and easy-to-use caking tester for measuring quantitatively cake strength as a result of elevated temperature, consolidation stress and storage time is introduced. The developed tester outweighed the conventional uniaxial unconfined failure caking tester due to the defined location of the failure plane to maximise repeatability, the necessity for a lower quantity of powder, maximised exposed surface and lower wall friction as well as production costs. The experimental design has been conducted by changing the temperature, consolidation stress and storage duration. The results showed that the tester could distinguish cake strength between different experimental conditions. A statistical model has been successfully developed to study the effect of each variable on the cake strength.

Introduction

Sometimes, powder producer companies will face new challenges when storing powders at elevated temperature, humidity and/or consolidation stress. Particulate materials such as detergents, pharmaceuticals and food powders have the potential to gain strength to the extent that caked lumps can form during handling and storage. The magnitude of caking may range from fairly weak lumps to the permanent fusion of particles, whereby these materials may no longer be acceptable to the process or end user. Caking could also decrease or even stop production rates at industrial facilities, hence, often, extra equipment is needed to de-agglomerate the caked powder. Particulate materials should ideally retain their desired flow properties during processing, handling and storage as well as through the distribution chain to the final consumer [1]. These types of reduction in the handling behaviour of particulate materials can lead to huge costs being imposed on industries [2]. Such costs justify the need for developing a tester for measuring as well as means for predicting caking at industrial scales.

There are several mechanisms responsible for powder caking. In general, these mechanisms are divided into four groups, 1. chemical caking, 2. electrical caking, 3. mechanical caking and 4. plastic flow caking (temperature dependent caking). Many different inter-particle forces are responsible for each of the caking mechanism. For instance, caking could happen due to different inter-particle forces such as liquid and solid bridges, van der Waals forces, magnetic force and electrostatic force [3]. The major adhesive force in a dry, consolidated and fine-graded powder bed is the van der Waals forces [4]. These interactions mainly depend on electric dipoles of atoms and molecules and the intensity of these interactions are affected by particle surface roughness, particle sizes and inter-particles distance (packing structure); these forces cause cohesiveness in the bulk solids [5] but no caking. Where moisture is present, liquid bridges between particles can be formed at the presence of moisture, and, hence, give rise to interparticle capillary forces. The capillary forces are formed as a result of surface tension, subject to the bridge surface curvature; hence this force contributes to the cohesive strength of the particulate solids. The formed capillary forces between particles do not themselves constitute caking forces, but are probably the initial mechanism which can cause subsequent cake formation [6] as described below. In contrast to liquid and van der Waals interactions, solid bridges make a continuous strong solid interaction between particles leading to a strong caked powder bulk [7]. Several mechanisms, i.e. dissolution-crystallization [8], sintering [9,10] and supersaturations (Ostwald caking [11]) are responsible for solid bridges. In sintering [9], particles coalesce, by bridging between two or more particles. The coalescence of particles is driven by atomic diffusion, the system minimizing the surface free energy by reducing the surface area. Other smaller particles are transported between the adjoining bigger particles in order to minimize the free surface energy of the new particle. The driving force for this process, sintering, is the difference between the capillary pressure and Laplace pressure in primary particles. Sintering is often associated with plastic deformation of the particles under pressure, increasing the interparticle surface contact; a high level of plasticity is associated with ease of relative movement of the atoms in the structure, facilitating at the same time the atomic diffusion that is thought to lead to bonding across these interparticle surfaces. This is therefore often referred to as plastic-flow caking. As particles coalesce it leads to irreversible densification and hardening of the bulk powder. In solvent evaporation [8], when the relative humidity of the surrounding environment is higher than a critical relative humidity, particles start to dissolve into adsorbed water forming liquid bridges between particles. If adsorbed water is then evaporated from these capillary bridges, dissolved solids then recrystallizes, thereby forming solid bridges between the particles. Furthermore, the liquid bridges can be supersaturated by decreasing the surrounding temperature until crystallization occurs. This caking mechanism can be seen in powders stored in closed containers where water evaporation is rare [12]. Ostwald caking [11] can also be responsible for the formation of solid bridges. In the Ostwald Caking mechanism, particles are dissolved in liquid bridges until saturated. However, particles with a small radius of curvature can dissolve in a saturated solution and make supersaturated solution. This supersaturated solution enables small particles to join, making large and harder particles in the powder, and thereby increasing the particle size and tensile strength of the powder bed. A further type of caking is due to chemical reaction either between particles, or between particles and the interstitial gas. For example, when humid air is present, cement will react with the moisture and the particles bond together due to the formation of products of reaction across the contact points [1].

Apart from the inter-particle forces, intrinsic properties of material, such as particle size and shape, cohesion, elasticity and level of hygroscopy [13], as well as extrinsic factors, i.e. level of consolidation and temperature have great influence on cake formation and needs to be addressed properly in order to avoid caking during storage and handling of the materials [13,14]. The plastic flow caking is the main focus of this research project.

The prerequisite for powder plastic caking is an increase in adhesion forces which could be related to an increase in Van der Waals forces between particles [15] augmented by likely sintering. These adhesive forces between particles determine the tensile strength of the caked powder which could be calculated through the Rumpf approach [16] modified by Molerus [17]. This model was initially developed to relate material strength, in terms of isostatic tensile stress, to the van der Waals forces, however it was also successfully applied to uniaxial state of stresses [18,19]. This approach is based on several assumptions, i.e. particles are spherical and monodispersed which is not valid for most of the powders used in industries. The other obstacles to use Rumpf approach for predicting caking strength are firstly unknown composition of the tested powder, hence it was not possible to establish the Hamaker constant, secondly measurements of the mean curvature radius at contact points as well as the plastic compressive yield strength of the material at the contact point need special technique and apparatus. In light of Rumpf approach limitation, statistical modelling for predicting and interpreting the caking strength results could be a useful approach.

A strict definition of caking is difficult to formulate, because its progression involves different stages. At any given stage, lumps may be few or numerous, of different sizes and of varying degrees of hardness. In practice, a quantitative caking measurement in each of the above stages is highly desirable [20]. There are number of different instruments and techniques for evaluating powder caking at these stages [21]. The capability of a caking tester to correctly determine the level of cake strength during storage would be the most desirable feature of such a tester. These methods are divided into several categories, for instance.

  • a)

    Tester based on stirred impeller (rheometer): The tester measures the vertical stress necessary to turn an impeller moving downwards through the caked sample and report it as the cake strength [22]. In another study, cake strength was reported as a torque necessary to rotate an impeller through the bed of caked powder, however this method is not suitable for powder that has undergone strong cake formation [23].

  • b)

    Fluidized bed testers: This method investigates the effect of relative humidity and temperature on powder stickiness. The relative humidity of the fluidized air is gradually increased until the powder becomes stickier at a given fixed air flow rate. Extra fluidization above the sticky point of the powder is not possible due to formation of channels in the fluidized bed column. This method is useful for determining ‘sticky’ curve of the powder as a function of temperature and/or humidity [21]. This test determined at what humidity the powder transitions from free flowing to cohesive due to moisture effects, but does not give an indication of cake strength, or even whether caking will occur when the moisture dries out.

  • c)

    Blow tester: an imposing stream of air is introduced to the bed of caked powder. The velocity of compressed air gradually increases till the point where the flow rate is sufficient to dislodge particles from the powder bed. The air velocity at which particle dislodging occurs is correlated to the cake strength of the sample [24].

  • d)

    Penetration method: this method is based on probe penetration through the caked samples. Billings et al. [24] and Ozkan et al. [25] used a penetrometer technique to measure the stress needed for the penetrometer to pass through the caked sample. In another study, ball indentation technique was developed to measure the hardness of the compacted powder. This method is very useful as an indication for the onset of caking at the surface of the powder bed [26].

  • e)

    Force induced uniaxial compression method: some research works have been conducted to develop a method based on the principle of a uniaxial compression test in order to break a caked sample. A modified (porous cylinder base in order to let the moisture migrate to the sample) uniaxial compression tester used to evaluate cake propensity of different powders at elevated temperature, humidity and powder consolidation. However, authors did not confirm if the powder bed reached environmental equilibrium [27,28]. Fitzpatrick et al. [29] developed force-displacement caking tester to measure the caking strength of the amorphous maltodextrin, skim milk powder and crystalline common salt at elevated humidity and storage duration. They measured caking strength by using a texture analyser as the maximum force for the rod to pass through the caked sample.

  • f)

    Shear tester: this technique has been used to measure caking of particulate materials at room conditions or elevated temperature, humidity and consolidation stress [[30], [31], [32]]. The details of shear tester principles have been illustrated elsewhere [33]. This method is time consuming, since powder will be consolidated for a desired period of time before a single point in yield locus line can be measured.

  • g)

    Tensile tester: Leaper et al. [34] reported the stress necessary to pull apart the two halves of a split cylinder containing caked material as the tensile strength of the caked specimen. Wang et al. [35] used a centrifuge technique to create a tensile stress inside the bed of caked powder. The tensile stress was determined when the powder bed start to fracture at a constant rotational speed.

The aims of this study were to develop a simple and easy to use caking tester in order to quantitatively measure the temperature dependent caking strength of a detergent powder at different temperatures, storage times and levels of consolidation stress. The main focus of this work is to determine if the developed tester could successfully measure caking strength and distinguish it between different experimental conditions. The second aim is to determine the most influential experimental variables on cake strength of the material through developing a statistical model.

Section snippets

Possible plastic flow caking mechanism

In this work, it is assumed that the cake strength increase occurs because of an increase in the cross-sectional area at each inter-particle contact throughout the bulk solid, through plastic deformation of the particles (creep) as a function of temperature and or pressure over a long period of time. This contact area increase will result in an increase in the bulk density (compaction), reduction in void fraction, and an increase in the tensile strength of material.

Materials

The powder used was a commercial household detergent which is sensitive to plastic flow caking. The instantaneous flow properties of the powder were measured with the Brookfield Powder Flow Tester (PFT). Flow function, i.e. unconfined yield strength values, fc, reported as a function of the major principal stress, σ1, during consolidation, is reported in Fig. 1. The unconfined yield strength in powder shear tester is a property derived from the static yield locus through the determination of

Greenwich Caking Tester (GCT)

A novel force displacement and easy-to-use caking tester, Greenwich Caking Tester, (GCT) for quantitatively measuring cake strength of powder was developed with some similarity to the Johanson hang- up Indicizer [37] and to the tester developed by Fitzpatrick et al. [15]. The sketch and dimension of the developed tester are illustrated in the Fig. 2. It excels the uniaxial unconfined failure caking (UUFC) strength tester since it has a low height to diameter ratio to minimize the wall friction

Experimental design

Full factorial experimental designs have been conducted. The variables were temperature, Te, consolidation stress, Stre, and consolidation duration, Dur. Each variable considered at 3 levels that were chosen to be representative of commonplace conditions of temperature, stress and duration experienced in the shipping of these powders. This gave us 3 × 3 × 3 = 27 experimental conditions, in order to improve the precision of the statistical model as well as examining the suitability of the tester

Results and discussion

The cake strengths of all experimental conditions are presented in Fig. 4a for 2 days storage time, Fig. 4b for 4 days storage time and Fig. 4c for 7 days storage time. In general, the shift in caking strength is observed, as storage time, consolidation stress and temperature increase. Furthermore, the developed tester and the approach for caking strength measurement showed its potential for correctly evaluating the caking propensity of the powder at the different experimental conditions.

Statistical model development

Explanatory variables for statistical modelling of caking strength, Ca, are temperature, Te, consolidation stress, Stre, and storage time, Dur. In addition, their squared values and interactions are also considered in the proposed model. The influence of explanatory variables on the response variable were compiled, modelled and analysed by multiple linear regression (MLR) method using the MODDE 12.0 software package (Umetrics, Umeå, Sweden). All values were scaled and centred before evaluation.

Conclusion

A simple and easy to use caking tester has been developed for measuring caking strength when powders are exposed to various mechanical and environmental (elevated temperature) conditions. The caking tester has potential as a commercial tester due to its high repeatability, ability to measure low level of caking strength, low level of production costs, the ability to differentiate between different experimental conditions, low powder usage and high level of selectivity. The tester showed the

Nomenclature

    c

    cohesion, Pa

    Cai

    response variable of the ith observation

    Ĉai

    model fitted value

    C¯a

    response variable mean

    fc

    unconfined yield strength, Pa

    ff

    flow factor, −

    J

    number of explanatory variables in the model

    n

    number of experimental points, −

    R2

    coefficient of determination, −

    Q2

    coefficient of multiple determination, −

Acknowledgment

We thank British Engineering and Physical Sciences Research Council (EPSRC) for providing the grant for the Virtual Formulation Laboratory (VFL) for Prediction and Optimisation of Manufacturability of Advanced Solids Based Formulations (EPSRC project number EP/N025261/1). Authors are grateful for the contribution given by Dr. Hajar Hajmohammadi, Department of Civil Engineering at the University College London, for the statistical analysis.

References (38)

  • V. Ganesan et al.

    Flow properties of DDGS with varying soluble and moisture contents using jenike shear testing

    Powder Technol.

    (2008)
  • M. Röck et al.

    Investigations on the caking behaviour of bulk solids—macroscale experiments

    Powder Technol.

    (2005)
  • H. Salehi et al.

    A comparison between powder flow property testers

    Particuology.

    (2017)
  • The Wolfson Centre for Bulk Solids Handling Technology

    Overview of Particulate Handling Technology

    (2016)
  • H. Salehi et al.

    On the use of a powder rheometer to characterize the powder flowability at low consolidation with torque resistances

    AICHE J.

    (2017)
  • H. Salehi et al.

    Dust release from aggregative cohesive powders subjected to vibration

    Ind. Eng. Chem. Res.

    (2017)
  • N. Christakis et al.

    Aggregation and caking processes of granular materials: continuum model and numerical simulation with application to sugar

    Adv. Powder Technol.

    (2006)
  • S.J.R. Simons

    Modelling of agglomerating systems: from spheres to fractals

    Powder Technol.

    (1996)
  • Y. Listiohadi et al.

    An exploration of the caking of lactose in whey and skim milk powders

    Aust. J. Dairy

    (2005)
  • Cited by (5)

    • Predicting powder caking using cohesion energy density

      2021, Powder Technology
      Citation Excerpt :

      The change in stress during compression of caked powder made of 4 mm particles is presented in Fig. 6. An increase in UYS with an increase in consolidation pressure, time of consolidation, and humidity have been reported in previous studies [46,47]. A similar increase in UYS was observed with the increase in kced in this study.

    • An original method for measuring liquid surface tension from capillary bridges between two equal-sized spherical particles

      2020, Powder Technology
      Citation Excerpt :

      At liquid-air surfaces, the surface tension results from the greater attraction of liquid molecules to each other than to the molecules in the air. The surface tension of liquids is encountered in many problems of science and engineering, such as in selective laser sintering [5,6], caking and agglomeration [7–10], medical problems involving respiratory diseases [11], particle-stabilised foams and emulsions [12,13], capillary cohesion in partially saturated granular materials [14–23] and capillary evaporation and condensation in porous media [24,25]. Therefore, it is crucial to accurately measure the surface tensions of liquids.

    View full text