Abstract. Introduction

Transcription

1 Evaluation of different methods for characterisation of physical properties of cosmetic emulsions Maria Grundén, Dept of Chemical Engineering, LTH, Lund University, Lund, Sweden Abstract Three special mixer systems (helical ribbon, small paddle and large paddle), have been investigated regarding the application on cosmetic emulsions. Measurements with the special systems are compared to a conventional cylinder system. The emulsion s particle size and distribution have also been studied, since the rheological properties of the products appear to be influenced by these properties. In the comparison between the different measuring systems it could be seen that all the systems often gave similar viscosity results. For the high viscosity products the conventional cylinder system could not be used, hence only the three special systems were compared to each other, often with good agreement. The viscosity of the various products changed differently with varying temperature. The particle size and distribution of an emulsion decrease when mixed at a higher rate. The distributions for the model samples were rather similar, with median sizes around 0,2 μm. For the real products, the distributions were more deviating. A connection between particle size and viscosity, when all the products are compared to each other, could not be seen, because the various products have different composition, with varying types and amount of emulsifiers and viscosity adjusters. For products created at the same place with the same equipment, agreement with the literature, stating that both a decrease in particle size and a decrease in distribution increases the viscosity, could be seen. Introduction In recent times, rheological measurements have become increasingly important to be able to characterize the consistency of cosmetic emulsions. Rheological measurements are now required in various cosmetic industries, e.g. quality control, storage stability, effects of formulation on consistency and prediction of flow behaviour in production (for pumping, mixing and packaging). In this study, the three mixer systems helical ribbon, small paddle and large paddle, previously evaluated by Roos [], are being investigated to see if they can be used for cosmetic emulsions. Measurements with the special systems using the parameters established by Roos are compared to a conventional cylinder system. The emulsion s particle size and distribution have also been studied, since the rheological properties of the products appear to be connected to these properties. It is possible to affect the particle size distribution of emulsions by changing either the amount or the kind of mixing. The particle size and distribution were measured through laser diffraction, which is one of the most widely used technologies for particle size measurements. The analysis is based on the principle of light scatter when a sample is irradiated with a light source. The resultant scatter pattern is then interpreted mathematically and the nature of the scatter pattern is based on the size of the sample.

2 Theoretical background Creams and lotions are semisolid emulsions, i.e. dispersions of immiscible or partially miscible liquids. A two-phase emulsion is a mixture of two immiscible liquids, where one in the form of microscopic or sub microscopic droplets is dispersed in the other. The majority of cosmetic products are oil-in-water (O/W) emulsions, i.e. oil droplets are dispersed in an external water phase. To stabilize the emulsion, an emulsifier is added to extend the inevitable separation of the phases. The emulsifier s effect on the viscosity of the continuous phase is the major contributor to the characteristics of the entire system, rather than its effects caused by changes in particle size or concentration of the internal phase. [2] [3] Viscosity is the measure of a fluid s resistance to flow, illustrated in Fig.. Fig : Velocity profile between parallel plates. A thin layer of fluid is enclosed between two plates with the area A (m 2 ). The upper plate is forced to move by a force F (N), with the velocity du x (m/s), while the bottom plate is fixed. [2] [4] [5] [6] Shear stress is defined as: F τ = (N/m 2 ) (Eq. ) A Shear rate is defined as: du γ& = x (s - ) (Eq. 2) dy The apparent viscosity of a fluid is shear stress divided by shear rate: τ μ app = (Pa s) (Eq. 3) & γ Fluids can be divided in two groups: Newtonian and non-newtonian. The viscosity of a Newtonian fluid is only dependant on temperature and pressure, which results in that shear stress is linearly proportional to the shear rate with zero intercept. Fluids in which the viscosity changes with the applied force are called non-newtonian. These can be divided into three groups: time-independent, timedependent and viscoelastic. The viscosity of commercial liquid products such as creams, lotions, dispersions and emulsions are time-independent and usually the viscosity decreases with increasing shear rate or shear stress. The behaviour of a non-newtonian fluid is often well described by the Herschel-Bulkley model: τ & n = τ 0 + K γ (Eq. 4) where K is the consistency coefficient, n is the flow behaviour index and τ 0 is the yield stress. [4] [6] A products viscosity is determined by its structure and the greatest impact on it has the continuous phase, e.g. water. The effect on the viscosity is dependent on the concentration, size and shape of the suspended particles and how they interact with the continuous phase. [7] It is possible to affect the particle size distribution of emulsions by changing either the amount or the kind of mixing. A large-particle-size emulsion usually has a broad particle-size distribution. If the average droplet-size decreases, so will the width of the distribution. Both the decrease in droplet-size and the decrease in the width of the distribution increase the viscosity. [4] A collection of particles never consist of particles with exactly the same size. Information on both average particle sizes and distributions about the average is required. The mode is the most frequently occurring value in a set of data, i.e. the value that passes through the peak of the relative frequency curve. The median is the particle diameter at which half the distribution is larger and half is smaller. 2

3 The most common value for representing the central tendency of a set of measurements is the mean, which is defined as the sum of any group of values divided by the number of values in the group. The most common way to describe the distribution is the standard deviation, defined as: N 2 σ = ( x i x) (Eq. 5) N i= where N is the number of data points, the particle size and x is the mean size. The standard deviation will be low if the data points are all similar and high if the data points are more spread out. [8] [9] [0] x i is Fig. 2: Concentric cylinder. Fig. 3: Helical ribbon. Materials and methods Samples Experiments were performed on both simple models and on real products. The prepared models were a simple emulsion (model sample ) and a lotion (model sample 2) both mixed at two different speeds (2000 rpm and 5000 rpm) and a cream (model sample 3) mixed at the higher speed. The real products were two different hair treatment creams (Wella Creambath Emulsion, Hair Care Cream), three different body lotions (Mildness Body Milk Honey, H&M Body Lotion, Eucerin ph5 Lotion) and one skin cream (Eucerin ph5 Cream). Viscosity All viscosity measurements were carried out on a Physica Rheolab MC viscometer. The conventional cylinder system used was a Z2 DIN (D=45mm, gap=,90mm). The special mixer systems were a helical ribbon (D=36mm, L=36mm), a small paddle (D=20mm, L=40mm) and a large paddle (D=40mm, L=60mm), shown in Fig For all the systems a cup with D=48,8mm was used. Fig. 4: Small paddle. Fig. 5: Large paddle. Each sample was measured at 5 C, 30 C and 45 C, with 20s integration time and 30 measuring points. The shear rate interval used ranged from s - to the maximum shear rate for each system respectively (032s -, 93s -, 60s - and 390s - ). Particle size The particle size measurements were carried out on a Coulter LS 230, a laser diffraction instrument. The measurements were carried out at 5-20 C. A small amount of the sample was diluted before being added to the instrument. Information about the sample was typed in and most important, the correct optical model ( oil in water ) was chosen. Analysis time was set to 90 seconds. Results and discussion Viscosity The different products acted very differently when heated. When Eucerin ph5 Lotion and Eucerin ph5 Cream was heated the viscosity decreased and became very low at 45 C. These follow the theory 3

4 that viscosity decreases with temperature (Fig. 6). 000 Viscosity for Wella Creambath and Hair Care Viscosity [Pa s] , Viscosity for Eucerin Lotion and Eucerin Cream 0, Shear rate [/s] EucerinLotion, 5C, small paddle EucerinCream, 5C, small paddle EucerinLotion, 30C, helix EucerinCream, 30C, helix EucerinLotion, 45C, large paddle EucerinCream, 45C, large paddle Fig. 6: Changes in viscosity with temperature for Eucerin ph5 Lotion and Eucerin ph5 Cream. Mildness Body Milk Honey and H&M Body Lotion, on the other hand, showed the same viscosity at 5 C and 30 C, but at 45 C they got a lot higher viscosity (Fig. 7). 00 Viscosity [Pa s] 0 Viscosity for Mildness Honey and H&M Lotion Viscosity [Pa s] , Shear rate [/s] Wella Creambath, 5C, small paddle Hair Care Cream, 5C, small paddle Wella Creambath, 30C, helix Hair Care Cream, 30C, helix Wella Creambath, 45C, large paddle Hair Care Cream, 45C, large paddle Fig. 8: Changes in viscosity with temperature for Wella Creambath Emulsion and Hair Care Cream. For all the samples, except for the model samples and 2, comparisons between the three special systems and the conventional cylinder system were made. Good agreement between the systems was often obtained. Two examples are shown in Fig. 9 and 0. In the graphs, the shear rate (/s) is on the x-axis, the shear stress (Pa) is on the right y-axis and the viscosity (Pa s) is on the left y-axis. 0, Shear rate [/s] MildnessHoney, 5C, small paddle H&M, 5C, small paddle MildnessHoney, 30C, helix H&M, 30C, helix MildnessHoney, 45C, large paddle H&M, 45C, large paddle Fig. 7: Changes in viscosity with temperature for Mildness Body Milk Honey and H&M Body Lotion. The viscosity changes with temperature were not as prominent for Wella Creambath Emulsion and Hair Care Cream as for the others. For Wella Creambath Emulsion the viscosity was almost the same at 5 C as at 30 C and decreased only slightly at 45 C, whereas for the Hair Care Cream the viscosity decreased a bit more and followed the temperature (Fig. 8). Fig. 9: Eucerin ph5 Lotion, measured at 5 C with helix, small paddle, large paddle and Z2. Fig. 0: Hair Care Cream, measured at 45 C with helix, small paddle, large paddle and Z2. 4

5 For almost all the samples, turbulence was achieved caused by the formation of eddies, giving the curves an odd appearance. For the model samples and 2, the viscosity seems to increase with shear rate and for the real products the viscosity at the higher shear rates seem to level off or even increase (as in Fig. 6). Particle size In all the figures presented the distributions are log-normal, because it gives a better general idea of the distribution. However, the calculated means and standard deviations are arithmetic, since one of the properties of interest is the Sauter mean diameter, i.e. the arithmetic surface area mean. The model samples and 2 were mixed at two different speeds: 2000 rpm and 5000 rpm. When mixed at the higher speed, the particle size got smaller and the distribution s width decreased. This is due to the larger particles being ruptured into smaller ones. (Fig. ) more from one sample to another (Fig. 2, Table ). Fig. 2: Particle size distributions for model sample, 2 and 3. Table : Properties of the particle size distributions for model sample, 2 and 3. Model sample Mean Median Mode S.D., 2000 rpm 0,59 0,24 0,6,05, 5000 rpm 0,30 0,23 0,27 0,28 2, 2000 rpm 0,85 0,26 0,4 3,37 2, 5000 rpm,07 0,7 0,3 5,52 3, 5000 rpm 0,37 0,23 0,23 0,77 Since the real products come from different manufacturers, they are produced in different ways and with different ingredients. Various types and amount of emulsifiers and viscosity adjusters are used in the products, giving them a range of particle sizes and very varying viscosity. Because of this, no clear connection between particle size and viscosity can be found when comparing all of them. One example of this is that the high viscosity Hair Care Cream has a similar particle size distribution as the much looser Mildness Body Milk Honey and H&M Body Lotion (Fig. 3). Fig. : Particle size distribution for model sample mixed at 2000 rpm and 5000 rpm. Because of the tails, the mean value differ from the median and the mode, and the distribution (e.g. standard deviation) becomes wider. All the model samples lie approximately in the same size range, hence the viscosity adjusters added have none or little effect on the particle size. All the medians and the modes are close to each other but the mean values differs Figure 3: Particle size distribution for Hair Care Cream, Mildness Body Milk Honey and H&M Body Lotion However, for products created at the same place with the same equipment, like model sample 3, Wella Creambath Emulsion and Hair Care Cream, a connection between 5

6 particle size distribution and viscosity could be seen. According to the literature both a decrease in particle size and a decrease in distribution increases the viscosity. Agreement with this can be seen for model sample 3, which has the highest viscosity (Fig. 4), the smallest particle mean size and smallest particle size distribution (S.D.) (Table 2). Viscosity [Pa] C, Small Paddle Shear rate [/s] Sample 3, 5000rpm Wella Creambath Hair Care Cream Figure 4: Viscosity results for model sample 3, Wella Creambath Emulsion and Hair Care Cream at 5 C. Table 2: Properties of the particle size distributions for model sample 3, Wella Creambath Emulsion and Hair Care Cream. Product Mean (μm) S.D. (μm) Model sample 3, 0,37 0, rpm Wella Creambath 2,04 7,73 Emulsion Hair Care Cream 0,76 3,96 One interesting observation is that many of the real products have a peak at 4 μm. Eucerin ph5 Lotion has its main peak at this size and for Wella Creambath Emulsion, Mildness Body Milk Honey and H&M Body Lotion there is a smaller peak (Fig. 5, Table 3). Fig. 5: Particle size distributions for the real products. Table 3: Properties of the particle size distributions for the real products. Product Mean Median Mode S.D. Wella 2,04 0,25 0,23 7,73 Creambath Emulsion Hair Care 0,76 0,4 0, 3,96 Cream Mildness, 0,4 0, 5,82 Body Milk Honey H&M Body,0 0,4 0, 6,0 Lotion Eucerin ph5 8,6 8,70 4,05 2,63 Lotion Eucerin ph5 Cream 24,49 8,20 8,00 20,37 The distribution for Eucerin ph5 Lotion had an unusual shape, hence four different bottles were analysed (Fig. 6). Figure 6: Particle size distributions for Eucerin ph5 Lotion from four different bottles. All the curves have a peak at 4 μm. The old measurement and a new from the same bottle (bottle number ) gave approximately the same shape and bottle number 2 gave a similar curve. All the bottles were analysed a number of times and usually the results were reproduced, but not for bottle number 3. From all the measurements from that bottle, two different curves (3-a and 3-b) appeared. Curve 3-a have completely different shape than all the other curves, with the only thing in common is the peak at 4 μm. On the other hand, Curve 3-b and the result from bottle number 4 showed matching curves. Conclusions Viscosity Measurements for emulsions with low viscosity are difficult to do with all the measuring systems. Even with the Z2, 6

7 measurements for model sample and 2 were problematic to do because of turbulence caused by eddies. The viscosity of the various products changed differently with varying temperature: the viscosity of Eucerin ph5 Lotion and Eucerin ph5 Cream decreased greatly with increasing temperature, for Mildness Body Milk Honey and H&M Body Lotion the viscosity was the same at 5 C and 30 C but at 45 C it increased and for Wella Creambath Emulsion and Hair Care Cream the viscosity slightly decreased with increasing temperature. In the comparison between the different measuring systems it could be seen that all the systems often gave similar results. For the high viscosity products Z2 could not be used, hence only the three special systems were compared to each other, often with good agreement. This means that the mixer systems can work as a complement to the conventional cylinder systems. However, there are limitations, especially for the large paddle, where formations of eddies resulted in turbulence at the higher shear rates. Particle size The particle size and distribution of an emulsion decrease when mixed at a higher rate. This occurrence is demonstrated by the model samples. The distributions for the model samples were rather similar, but for the real products, the distributions were more deviating. A connection between particle size and viscosity of the model samples could not be seen, and since the various real products have different composition, no connection between particle size and viscosity could be seen when comparing all of them. For products made at the same place with the same equipment, a connection between particle size distribution and viscosity appeared that agree with the literature, that the viscosity increases with both the decrease in particle size and the decrease in the width of the distribution. This could be seen for model sample 3, which has the highest viscosity and both the smallest particle mean size and smallest particle size distribution when compared to Wella Creambath Emulsion and Hair Care Cream. Acknowledgement I would like to thank my supervisor Ulf Bolmstedt, Tetra Pak Processing Components, for all the encouragement, guidance and support. I also want to thank my examiner Anders Axelsson, Dept. of Chemical Engineering, LTH, Lund University, for help and support. I would like to thank everyone who has contributed with material for my experiments: Danial Irfachsyad, Tetra Pak Stainless Equipment in Jakarta and Joachim Hallström at Forte Cosmetic AB, Höör. The work was carried out at Tetra Pak, and I would like to thank Lisbeth Frank, Tetra Pak Dairy and Beverage Systems, for all the help, support and optimism throughout the work. Nomenclature A area m 2 D diameter mm du x velocity m/s dy gap width mm F force N K consistency coefficient Pa s n L length m N number of data points - n flow behaviour index - x i particle size μm x mean particle size μm. γ shear rate s - μ app apparent viscosity Pa s σ standard deviation μm τ shear stress Pa τ yield stress Pa 0 7

8 References [] Roos H., Evaluation of new methods and measuring systems for characterisation of flow behaviour of complex foods, MSc Thesis, Department of Chemical Engineering, LTH, Lund University, Sweden, 2005 [2] Brummer R., Rheology Essentials of Cosmetic and Food Emulsions, Springer- Verlag, Berlin Heidelberg, Germany, 2006 [3] Miner P. E., Chapter 9 - Emulsion Rheology: Creams and Lotions; In Laba D. (Ed.) Rheological Properties of Cosmetics and Toiletries, Marcel Dekker, New York, 993 [4] Barnes H. A., Viscosity, Cambrian Printers, Aberystwyth, Wales, 2002 [5] Naé H. N., Chapter 2 Introduction to Rheology, In Laba D. (Ed.) Rheological Properties of Cosmetics and Toiletries, Marcel Dekker, New York, USA, 993 [6] Steffe J. F., Rheological Methods in Food Process Engineering, Freeman Press, Michigan, USA, 992 [7] Barnes H. A., A Handbook of Elementary Rheology, Cambrian Printers, Aberystwyth, Wales, 2000 [8] Allen T., Particle Size Measurement, Chapman and Hall Limited, London, Great Britain, 990, 4 th Edition [9] Cadle R. D., Particle Size, Theory and Industrial Applications, Reinhold Publishing Corporation, New York, USA, 965 [0] Coulter LS 230-manual, Appendix B, Statistics 8