Understanding the Crystallization Process in Detergent Formulations

Applied Chemistry is published by the Understanding American Chemical Society. 1155 Sixteenth the crystallization Street N.W., Washington, 20036 process inDC detergent Published by Subscriber access provided byAmerican Kaohsiung Chemical Society. Medical University Copyright © American Chemical Society. However, no copyright

formulations in the absence and is published by the American Chemical Society. 1155 Sixteenth presence of agitation Street N.W., Washington,

Emily Summerton, Jeanluc DC 20036 Published Subscriber access provided by byAmerican Kaohsiung Bettiol, Christopher Chemical Society. Medical University

Copyright © American Chemical Society. However, no copyright

Jones, Melanie M. Britton, and Serafim Bakalis is published by the

Ind. Eng. Chem. Res.,Chemical Just American Society. 1155 Sixteenth Accepted Manuscript • DOI: Street N.W., Washington, 10.1021/acs.iecr.8b02975 • DC 20036 Publication Date (Web): 02 Nov 2018 Published by American Subscriber access provided by Kaohsiung Chemical Society. Medical University Copyright © American Chemical Society. However, no copyright

Downloaded from http:// pubs.acs.org on November 4, 2018 is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, 20036 Just Accepted DC Subscriber access Published provided by byAmerican Kaohsiung Chemical Society. Medical University Copyright © American Chemical Society. However, no copyright

“Just Accepted” manuscripts have been pee online prior to technical editing, is published by the formatting fo American Chemical Society provides “Just Accepted” as a servic Society. 1155 Sixteenth of scientific material as soon as possible a Street N.W., Washington, full in PDF formatDCaccompanied by an HTM 20036 Published by Subscriber access provided byAmerican Kaohsiung peer reviewed, but should not be considere Chemical Society. Medical University Copyright © American Chemical Society. However, no copyright

Digital Object Identifier (DOI®). “Just Accep the “Just Accepted” Web site may not inclu is published by the a manuscript is technically edited and form American Chemical site and published as an ASAP Society. 1155 Sixteenth article. N N.W., Washington, to the manuscriptStreet text and/or graphics wh DC 20036

Subscriber access Published provided by byAmerican Kaohsiung Chemical Society. Medical University Copyright © American Chemical Society. However, no copyright

ethical guidelines that apply to the journal consequences arising from the use of infor is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Subscriber access Published provided by byAmerican Kaohsiung Chemical Society. Medical University Copyright © American Chemical Society. However, no copyright

PageIndustrial 1 of 35 & Engineering Chemistry Research 1 2 3 4 5 6 7

0 °C No agitation

time t1

0 °C Agitation

time t2 Simulate supply chain

t1 >> t2 Variability (tACS 1) >> Variability Paragon(t2)Plus

Environment

Industrial & Engineering Chemistry Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Understanding the crystallization process in detergent formulations in the absence and presence of agitation Emily Summerton a,*, Jeanluc Bettiolb, Christopher Jonesb, Melanie M. Brittonc, Serafim Bakalisa ,d aSchool

of Chemical Engineering, University of Birmingham, Edgbaston, B15 2TT, United

Kingdom bProcter

and Gamble Brussels Innovation Center, Temselaan 100, 1853, Strombeek Bever,

Belgium cSchool

of Chemistry, University of Birmingham, Edgbaston, B15 2TT, United Kingdom

dDepartment

of Chemical and Environmental Engineering, University of Nottingham,

Nottingham, NG7 2RD, United Kingdom

ABSTRACT

Detergents are used on a global scale and, therefore, are exposed to a wide range of temperatures and shear rates during distribution. At low and sub-zero temperatures crystallization may occur in the product. Manufacturers utilize a variety of methods to detect these crystal failures, typically involving the storage of formulations between 3 °C and 10 °C for up to 28 days. This paper describes the application of agitation as a route to reduce the timescale of the stability tests, thus increasing productivity. Agitation simulates vibrations that the formulations experience during distribution. In the absence of mixing, crystallization was found to originate from the air-liquid interface, whereas in the presence of mixing, the crystallization began around the mixing blade. 1 ACS Paragon Plus Environment

Page 2 of 35

Page 3 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Industrial & Engineering Chemistry Research

The times to failure detected by the current and proposed methods correlated, indicating the potential for this novel approach to be used in stability testing. 1. INTRODUCTION The soap and detergent industry manufactures a wide range of products, such as shampoos, fabric conditioners and dish liquids, with an estimated net-worth of $97.26 billion in 2016.1 Surfactants are the main components in these products, typically present at concentrations between 15 % and 40 %.2 The detergent industry is a major user of anionic surfactants.3 The anionic surfactant sodium dodecylsulfate (SDS) and the amphoteric surfactant N,N-dimethyldodecylamine N-oxide (DDAO) are examples of two surfactants present in detergent formulations.4-6 Under the alkaline conditions of dish liquid, amine oxides exhibit non-ionic behaviour.7 Since cleaning products are used worldwide, they must demonstrate chemical and physical stability across a wide range of environments. At high temperatures there is a risk of discolouration, due to degradation of dyes contained in the formulation,8 and at low and sub-zero temperatures surfactant crystallization may occur. Both of these are detrimental to physical stability. Despite the industrial significance of surfactant crystallization in detergent products, there are limited studies into the crystallization of mixed surfactant systems.9-14 The tendency of an anionic surfactant system to crystallize is reduced on addition of a non-ionic surfactant, as has been reported for SDS + nonphenol polyethoxylate9, 12 and SDS + DDAO mixtures.14-15 In the presence of a non-ionic surfactant, the tendency for micelle formation increases, reducing the concentration of anionic surfactant monomers from which crystals form.

2 ACS Paragon Plus Environment


In addition to the types of surfactants present in the mixture, physical factors, such as agitation, presence of dust particles, temperature, humidity and air-liquid surface area are also expected to influence the mechanism of crystallization in detergent products and associated stability tests. In this paper, the effect of agitation on the failure mechanism is discussed and compared to the behaviour observed in the absence of any agitation. The mechanism of crystallization within detergents, with or without agitation, is yet to be addressed in the literature. Understanding the effect of agitation on detergent crystallization will provide insight into product stability during the supply chain. Detergents are exposed to higher levels of agitation during road transport and the unloading and loading of goods, compared with distribution via ferry or plane.16 Depending on the condition of the road surface, the background vibration during vehicle transportation has accelerations between 0.5 and 1 g, where g is the gravitational constant. In contrast, loading products into aircraft can result in shocks up to 10 times greater. When products are transported across the globe, they are also subjected to a variety of temperatures, typically ranging between 3 °C and 50 °C. The temperatures are prone to change, and a rate of 0.1 °C/min is typical of environmental fluctuations.17 When products are transferred between different environments, such as from inside to outside storage, the rate of thermal change ranges from 0.5 to 50 °C/min.17 In warehouses or production plants such changes are unavoidable. Products are typically held in warehouses for limited periods, between 2 and 4 h, before onward transportation.18 Home and personal care products must remain stable for a 2 year shelf life period.19 Manufacturers have methods in place to ensure product stability through the preparation, packaging, transportation and subsequent shelf life stages, Typically, formulations are stored in 3 ACS Paragon Plus Environment

Page 4 of 35

Page 5 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


temperature chambers (50 °C to 3 °C) and their chemical properties, including viscosity, appearance and odour, are monitored at regular intervals.20 In addition, the effect of continuous heating and cooling on sample stability must also be tested, to account for temperature fluctuations experienced during the supply chain. The current tests for detergent products solely monitor the stability of quiescent formulations, which is not necessarily representative of the supply chain where background vibrations exist. Therefore there is a requirement for a method which incorporates agitation and more closely simulates the distribution process. The application of agitation may also improve the efficiency of the stability tests at low temperatures. The timescale of traditional stability tests can be long, requiring up to 28 days of storage at a variety of temperatures, and the time to crystallization also varies between samples of the same formulation. This is attributed to factors that are difficult to control such as the existence of dust particles, temperature fluctuations or pressure differences. Dust or other foreign objects can provide nucleation sites for crystallization.21-22 There is a requirement to identify an alternative test capable of determining whether a formulation remains physically stable. An ideal test would generate reliable results within a shorter timescale. Using agitation to reduce the timescale of crystallization in detergent products is an area of scientific research that has not previously been reported in the literature. However, agitation influences the time to crystal formation in other systems as a result of both primary and secondary nucleation effects.23 A study into the kinetics of alpha-lactose monohydrate crystallization revealed an increase in the rate of primary nucleation upon agitation.23 From classical nucleation theory, the primary nucleation rate 𝐽total is a sum of the homogenous (𝐽hom) and heterogenous ( 𝐽het) nucleation contributions: 4 ACS Paragon Plus Environment


𝐽total = 𝐽hom + 𝐽het

Page 6 of 35

(1)

Both the homogenous and heterogeneous components are given by: 𝛥𝐺

𝐽 = 𝐴𝑒

― 𝑘𝑇

(2)

where A is a pre-exponential factor, ΔG is the energy of nucleation, k is the Boltzmann constant and T is the temperature.24 Upon mixing, the increase in the rate of primary nucleation is attributed to an increase in the pre-exponential factor A of both the heterogenous and homogeneous components and a decrease in the energy barrier for heterogenous nucleation. The increase in A is due to an increase in the Brownian motion of the molecules resulting in a higher probability of crystal formation.25-26 This has also been demonstrated for silica particle systems, in which only a small amount of shear was applied.27 The rough surface of the stirrer provides additional nucleation sites for crystallization, reducing the energy barrier to heterogenous nucleation.23, 28 In addition to primary nucleation, the application of mixing also influences secondary nucleation. In a study focused on the crystallization of glutamic acid,29 it was found that mixing increased the rate of secondary nucleation in the bulk. The high shear at the stirrer blade diminishes the boundary layer surrounding the crystals and mixing blade thus promoting further crystal growth. Furthermore, crystals are broken apart by the mixer to yield additional nucleation sites throughout the system.30 The effect of flow on crystallization has also been reported for various systems.31-32 The flow rate and rheological properties of the fluid impact on the crystal morphology and growth rate.33 In these continuous systems, increasing the flow rate increases the rate of the crystallization causing fouling in pipes. It is expected the similar principles may apply with mixing, since both methods rely on an input of energy to the system. 5 ACS Paragon Plus Environment

Page 7 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


This paper explores the potential of mixing to reduce the time to crystallization across a range of dish liquid products, with varying degrees of stability. The mechanism of the failure, both in the presence and absence of mixing, is discussed. This knowledge will enable industry to develop more robust formulations, improve product packaging and introduce more efficient stability tests. In particular, an insight into the origin of the crystallization is crucial for improving product stability. 2. MATERIALS AND METHODS 2.1 Materials 2.1.1 Model SDS + DDAO studies Binary surfactant solutions were prepared with 20 wt. % SDS and 3 wt. % DDAO in distilled water and used for conductivity measurements. SDS (Fisher Scientific, 97.5 %) and DDAO (Sigma Aldrich, 30 wt. % in water) were used without further purification. The solutions were made up 24 h in advance to minimize hydrolysis. The solution was mixed for 15 min at 25 °C and then left for 12 h to release any entrapped air. 2.1.2 Commercially relevant dish liquid formulations When crystals appear in complex commercial products, the resulting appearance is unacceptable for consumers. By understanding the mechanism of the crystallization process, it is expected that product stability at low temperatures, between 10 °C and 3 °C, and the efficiency of the stability tests can be improved. Therefore, to ensure the industrial relevance of this work, commercially relevant dish liquid products were used throughout the study. A typical dish liquid formulation consists of anionic surfactants (linear, branched and ethoxylated alkyl sulfates), amphoteric 6 ACS Paragon Plus Environment


surfactants (linear amine oxides), water, preservatives, dyes, perfume, sodium chloride (NaCl), polypropylene glycol and ethanol. 2.1.2.1 Formulations A and B Mechanistic studies, in the absence and presence of mixing, were performed using a dish liquid product, Formulation A, specifically formulated to be unstable at low temperatures. Previous stability tests revealed that Formulation A develops crystallites at 0 °C within 28 days. Formulation B was formulated to be stable under the same conditions. In Formulation A, the degree of ethoxylation m of the alkyl sulfate component CxHy(OCH2CH2)m Na+ averaged at 0.5 moles and the wt. % ratio of alkyl sulfates to amine oxides was 4.4:1. The degree of ethoxylation m of the alkyl sulfate component CxHy(OCH2CH2)m Na+ in Formulation B averaged at 0.6 moles and the wt. % ratio of alkyl sulfates to amine oxides was set as 3:1. Commercial grade sunflower oil was also used to explore the role of the air-liquid interface in the crystallization process. The effect of mixing speed on the crystallization process was investigated using Formulation A. 2.1.2.2 Formulations X1-X6 Whilst Formulations A and B were tailored to be stable and unstable at low temperatures, respectively, Formulation X1-X6 spanned a range of product stabilities as determined from standard stability tests at 0 °C. These were used to compare the times to crystallization in the absence and presence of mixing to determine any correlation between the current and proposed methods. The six formulations varied in the molar amounts of the linear alcohol (LA), branched alcohol (BrA), linear alcohol ethoxylate (LAE) and branched alcohol ethoxylate (BrAE) contributions. 7 ACS Paragon Plus Environment

Page 8 of 35

Page 9 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The blends were subsequently sulfated and the other ingredients added, while maintaining the same alkyl sulfate to amine oxide ratio (3.0:1), the same total surfactant concentration (35.8 wt. %) and a comparable completeness level between the two products. Completeness is related to the amount of alcohol remaining in the system (equation 3): 𝑀𝑊𝑎𝑙

Completeness (%) =

(𝐴𝑐𝑡 𝑀𝑊 ) 𝑎𝑠

(𝐴𝑐𝑡 ) + 𝑈𝑛𝑅 𝑀𝑊𝑎𝑙 𝑀𝑊𝑎𝑠

𝑥 100 %

(3)

where Act is the active level (wt. %), MWal and MWas are the molecular weights of the alcohol and alkyl sulfate respectively and UnR is the amount of remnant alcohol in wt. %. 2.2 Methods 2.2.1 Conductivity 30 g of the surfactant solution was placed in a 50 mL glass beaker and inserted into a jacketed vessel. The temperature of the 50:50 water:ethylene glycol mixture contained in the jacket was controlled by a Grant LTD 6G cooling bath (Grant Instruments, UK). Upon cooling from 25 °C to 0 °C, the conductivity of the 20 wt. % SDS + 3 wt. % DDAO solution was measured every 30 seconds via an Orion star A212 conductivity meter (ThermoFisher Scientific, USA). A mini-stirrer was used to induce mixing in the system for the duration of the experiment. The stirrer paddle was supplied as an attachment to the Orion star A212 conductivity meter (ThermoFisher, UK), consisting of a double blade propeller with a 15 mm diameter. The length of the probe was 210 mm with the corresponding width and height both measuring 20 mm. A schematic of the setup is provided in Figure 1.



Figure 1. Schematic diagram of the setup for conductivity measurements. The rate of cooling was approximately 0.5 °C/min, although this could not be controlled. Each measurement was performed in triplicate at five different speeds (indicated in Table 1) and in the absence of any stirring. Table 1. Speed levels and corresponding rpm values. Stirrer speed level 1 2 3 4 5

Average rpm 914 1385 1842 2310 3333

2.2.2 Time lapse photography 2.2.2.1 Cooling chamber experiments Crystallization of dish liquid formulations was visually recorded using a Canon EOS 5D Mark II. 80 g of each formulation was added into a 100 mL test bottle composed of polyethylene terephthalate (PET). Samples were placed in a Memmert ICP 450 chamber (Memmert, Germany), set to 0 °C. Images were automatically captured every 10 min for a period of 7 days for each formulation. A backlight panel was used to minimize any light reflections from the surroundings. 9 ACS Paragon Plus Environment

Page 10 of 35

Page 11 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


For each test, three replicate measurements were performed. Image analysis was performed using a processing software developed by P&G. This software functions by converting an image set to greyscale and assigning each individual pixel a value from 0 to 255, depending on the greyscale intensity. For each test, the maximum area of the bulk absent from light reflections was used in the analysis. Each image can be envisaged as a 2D-matrix filled with pixel values between 0 and 255. A greyscale-co-occurrence matrix (GCOM) was produced for the positional relation x = 0, y = 1, where each pixel was compared to the one vertically above.34 The matrix depends on how many times a certain difference in greyscale occurs between pixels with this chosen spatial relationship. The matrix was subsequently normalized and the homogeneity, a statistical property of the matrix,35 was calculated (equation 4). 𝑝(𝑖,𝑗)

Homogeneity = ∑𝑖,𝑗1 + |𝑖 ― 𝑗|

(4)

where i and j refer to the greyscale intensity of the ith and jth pixels; p(i,j) is the cooccurrence probability of pixels i and j. The area of analysis was selected such that it incorporated most of the bulk within the bottle. Graphical plots showing the change in homogeneity across the different formulations were produced in OriginPro (v.9.0). From these plots, the time to failure for each quiescent formulation was estimated as the point at which the gradient changes indicating a decrease in homogeneity within the system. In addition, 60 mm diameter and 100 mm diameter petri dishes containing 20 mL of Formulation A were placed in the chamber. 2D images of the petri dishes were captured at 24 and 48 h time points using a Canon 100D camera positioned vertically above the sample. Thresholding in ImageJ was used to determine the area of the formulation which contained crystals. 10 ACS Paragon Plus Environment


A magnetic stirrer plate was also inserted in the chamber. To investigate the effect of mixing, a magnetic flea of 35 mm length and 6 mm diameter was added to a formulation contained in a 100 mL PET bottle and placed on the plate, with the flea rotating at approximately 10 rpm. However, the effect of mixing was investigated in detail with the jacketed vessel setup outlined below. 2.2.2.2 Jacketed vessel experiments The vacuum jacketed glass vessel depicted in Figure 2 was used to observe the effect of mixing on dish liquid crystallization at low temperatures. 400 mL of Formulation A was initially poured into the hemispherical vessel, with an internal diameter of 75 mm and a height of 180 mm. A Grant LTC4 cooling bath (Grant Instruments, UK), filled with 50 % ethylene glycol and 50 % water by volume, was connected to the jacketed chamber. The bath was set to 0 °C and the vessel left to cool for 2.5 h. This was the time required for the centre of the vessel to reach 0 °C. A temperature of 0 °C was selected for these investigations in order to be consistent with stability tests within P&G. A Hei-TORQUE Precision 100 overhead stirrer (Heidolph Instruments, Germany) was set to the chosen speed. A PTFE screw propeller impellor was used with an 8 mm shaft diameter and a 400 mm shaft length. The impellor measured 40 mm in diameter and 15 mm in height. The blades were positioned at a 45 ° rotor angle with the length of each blade measuring 21 mm. The stirrer was positioned approximately 30 mm above the base of the vessel. Images were captured every 10 seconds for a total period of 60 min, or until crystallization was complete, using a Cannon 100D mounted on a Velbon EF-61tripod. A backlight panel was used to attain optimal images. Stirrer speeds of 20 rpm, 50 rpm, 150 rpm and 300 rpm were investigated, with three repeats performed at each speed. 11 ACS Paragon Plus Environment

Page 12 of 35

Page 13 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The resulting time lapse images were analysed in MATLAB 2016b. Images were converted to 8bit greyscale images with each pixel assigned a number from 0 to 255, corresponding to the intensity. Values of 255 and 0 relate to white and black pixels, respectively, with numbers between relating to differing amounts of greyscale. The change in average greyscale intensity for the pixels in each boxed area, A-D (Figure 2(b)), was determined. The results were plotted as percentages of greyscale intensity over time, with a value of 255 corresponding to 100 %. The time relating to a reduction in greyscale intensity is related to the onset of crystallization. This method was also used with Formulations X1-X6 to investigate the effect of mixing over a wider range of formulations. A stirrer speed of 20 rpm was used for these tests.

Figure 2. (a) Schematic diagram of vacuum jacketed vessel set up. Adapted from a figure supplied by Asynt Ltd, UK. (b) Photograph of the vessel set up indicating the areas used in analysis. 2.2.3 Transmissivity The time to crystallization for Formulation A was also detected using a Crystalline unit (Technobis, Netherlands), which can simultaneously record the temperature and turbidity of multiple samples. Crystal formation was detected by an increase in the turbidity of the solution and a corresponding reduction in transmissivity. The instrument was connected to a Lauda Proline RP 845 cooling bath (Lauda-Brinkmann USA) to attain the necessary low temperatures. For each measurement, 4 mL of Formulation A was pipetted into a vial and placed into the furnace of the 12 ACS Paragon Plus Environment


instrument. To apply mixing, a mini overhead stirrer replaced the vial cap and the speed was set to 1000 rpm. The samples were exposed to temperature cycles from 20 °C to 0 °C at a maximum cooling rate of 15 °C/min and with a hold time of 1.5 h at 0 °C and 30 min at 20 °C. Each experiment consisted of 13 consecutive temperature cycles, with every test performed in duplicate. The temperature and transmittivity were reported for the duration of each experiment. This method employs smaller sample volumes than experiments performed in test bottles or the jacketed vessel. 3. RESULTS AND DISCUSSION 3.1 Mechanism without agitation Time lapse images of Formulation A at 0 °C show crystal formation occurs at the air-liquid interface (Figure 3). In the absence of any physical forces or seeding, crystallization is commonly observed to commence from the surface. The crystals subsequently sink to the bottom of the product over a time due to the higher density of crystals, with respect to the bulk.

Figure 3. Time lapse images acquired during the crystallization of Formulation A at 0 °C.


Page 14 of 35

Page 15 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


This observation is hypothesized to be a result of segregation of surfactants at the air-liquid interface.36 Figure 4 shows a simplified schematic diagram of the different surfactant environments, before and after crystallization. Surfactant molecules in the monolayer exist in a close packed arrangement and, therefore, bear a close structural resemblance to the crystals.37 The crystals are known to be alkyl sulfate hydrates, consisting of layers of alkyl sulfates.14 Because of this similarity, crystal formation from the interface is proposed to be the lowest energy pathway.38 In addition, solvent evaporation at the surface may result in high local supersaturation , promoting crystallization from this region.39 However, at 0 °C, it is unlikely there will be significant evaporation from the formulation. The rate of evaporation depends on the air-liquid surface area and the vapour pressure of the formulation There is also evidence that the surfactants are organized in a bilayer just below the interface, rather than a monolayer at the surface.40-41 If this is the case, it is expected that crystallization would originate from the close-packed bilayer.

Figure 4. Simplified schematic diagram illustrating the different surfactant environments residing in a dish liquid sample at 25 °C. 14 ACS Paragon Plus Environment


A further test was performed in which an oil layer was added to Formulation A at 0 °C. On addition of the oil, no crystals were observed in Formulation A after 7 days at 0 °C (Figure 5). The presence of an oil layer on the surface was hypothesized to supress crystallization at the interface. The hydrophobic oil is predicted to stabilize the surfactant hydrocarbon chains of the monolayer, effectively ‘freezing’ the interface and reducing the drive for crystallization. It is likely that the presence of the oil layer also reduces the degree of evaporation from the surface, minimizing local supersaturation at the interface. Since supersaturation dictates whether a sample will crystallize, this effect may also contribute to a reduced tendency for crystallization. In the absence of an oil layer crystalline entities were observed after 2 days, as shown in Figure 5. Crystals sink to the bottom of the formulation, as indicated by the areas labelled 1, 2 and 3. These findings corroborate the proposed mechanism, which is that crystal formation occurs from the close-packed surfactant monolayer. There is a possibility of the surfactants partitioning into the oil layer, in the form of inverse micelles, and disrupting the equilibrium.42 This could contribute to the increased stability of the system. Dish liquid comprises a mixture of surfactants, where some have larger hydrophobic contributions than others. The more hydrophobic surfactants will tend to reside in the oil phase, thus improving the overall stability. A monolayer will still form at the interface regardless of whether the surfactants partition into the oil. On addition of an oil layer no crystals are observed within 7 days, indicating surface phenomena have a significant role.


Page 16 of 35

Page 17 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 5. Time lapse images of an unstable dish liquid formulation held isothermally at 0 °C over 7 days (a) in the absence and (b) in the presence of a layer of sunflower oil. The importance of the interface was evident when observing crystallization of Formulation A in petri dishes of two different diameters. After 24 h, there was only a minor difference in the measured crystalline areas. However, after 48 h, the area containing crystals was larger for the 100 mm diameter dish (see Figure S1, supplementary information). 3.2 Effect of agitation 3.2.1 Application to the model SDS + DDAO system The effect of mixing on the crystallization kinetics was investigated, using conductivity measurements, on a model SDS + DDAO system. When the SDS + DDAO system crystallized the conductivity of the solution decreased significantly. The counter-ions in the solution bind to the crystal phase and become less available to transport charge.43 A typical conductivity profile is provided in Figure 6(a) which shows that crystallization has occurred when the conductivity of the 16 ACS Paragon Plus Environment


SDS + DDAO solution reaches 8 mS/cm. The time to crystallization is indicated by the time for the conductivity to reach 8 mS/cm and is plotted across the different stirrer speeds in Figure 6(b). The average time for the quiescent sample to reach this value is 41 min longer than the time measured when stirring the solution at maximum speed. These results imply that mixing promotes the onset of SDS crystallization.

Figure 6. (a) Example of a conductivity curve upon cooling a 20 wt. % SDS + 3 wt.% DDAO system at speed level 1 (949 rpm) (b) Plot of the time to crystallization across the range of speed levels where the time for crystallization is taken as the time for the conductivity of the surfactant solution to reduce to 8 mS cm1. 3.2.2 Application to dish liquid products A representative time series of photos are shown in Figure 7 for Formulation A at 0 °C, in the presence of mixing. Crystallization was observed to begin around the stirrer bar rather than the airliquid interface. From studies on similar surfactant systems, the crystals are likely to be composed of non-ethoxylated alkyl sulfate hydrates since the system is below the Krafft temperature of these ionic surfactants.14-15 Although these studies were done on quiescent samples, it is expected that 17 ACS Paragon Plus Environment

Page 18 of 35

Page 19 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the presence of mixing will not affect which component crystallizes first. These initial crystals are particularly important since their formation determines the time of product failure. However, the input of energy may cause further components to crystallize over longer timescales, which is a subject of further investigation. Formulation B did not develop crystals even after mixing the formulation for 4 days at 0 °C. This was expected since Formulation B did not fail current stability tests. The plot in Figure 8 shows the relative times to crystallization for Formulation A under the influence and absence of mixing, as determined by visual inspection. As expected from prior work on the SDS + DDAO model system (see Section 3.2.1), the time to crystallization reduces when the solution is stirred. Furthermore, the box plot in Figure 8 shows that the variability of failure times also reduces when mixing is present. The effect of mixing on dish liquid crystallization was observed through a change in light transmission. A snapshot of the measured light transmission for the Formulation A across fluctuating temperature cycles, both in the absence and presence of mixing is provided in the supplementary information (Figure S2).



Figure 7. Time lapse images of the crystallization of Formulation A in the presence of mixing. The initial colour differences in the solution are due to unavoidable light reflections from the surroundings. Circles 1, 2 and 3 indicate the growing crystalline area.

Figure 8. Box plot demonstrating the reduction in the timescale to crystallization and statistical variability across 8 replicates of Formulation A held at 0 °C. Mixing is observed to not only reduce the time to crystallization, and its variability, but appears to also affect the mechanism of crystal formation. When Formulation A was mixed at 0 °C, crystallization originated from the stirrer blade (Figure 7) but, in the absence of mixing, crystals 19 ACS Paragon Plus Environment

Page 20 of 35

Page 21 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


were initially observed at the air-liquid interface (see Section 3.1). This change in mechanism, and reduced timescale, are believed to be explained by secondary and primary nucleation effects.23, 30 Secondary nucleation occurs when existing crystals are used to create further nucleation sites throughout the system. Possible mechanisms for secondary nucleation, when in the presence of mixing, include crystal breakage and dispersion, crystal-crystal and crystal-stirrer collisions or the replacement of crystal boundary layers with further solute molecules. The rate of secondary nucleation B0 can be expressed by equation 5: (5)

𝐵0 = 𝑘𝑆𝑀𝑇𝑊

where k is a constant, S is supersaturation, MT is crystal density and W is the agitation rate. Fluid movement is greatest for areas of the formulation closest to the mixing blade.44-45 Hence, the rate of secondary nucleation is also higher in these areas, which may explain why crystallization is observed close to the stirrer. The rate of primary nucleation is also affected by the presence of mixing.23,

46

The increased

movement of the bulk, especially near the mixing blade, results in a higher collision frequency between the monomers and hence an increased probability of crystal nuclei formation.47 This effect is more pronounced at the mixing blade, where the velocity of the fluid is highest. The rough stirrer surface provides additional nucleation sites for crystallization. The energy barrier to heterogeneous nucleation reduces and crystallization originates at the stirrer blade.48



3.3 Effect of mixing speed Images were acquired during the crystallization of Formulation A at 0 °C at various mixing speeds. Typical images are provided in Figure 10, at a speed of 20 rpm. These images provide further evidence for crystal formation starting around the stirrer, rather than the air-liquid interface.

Figure 10. Time series of images, acquired at 10 min intervals, during the crystallization of Formulation A at a stirrer speed of 20 rpm, with the solution held at 0 °C. Plots of the average greyscale intensity over time for the areas contained within boxes A-D (Figure 2(b)) are provided in Figure 11 across the different mixing speeds. Crystallization is detected by a reduction in the greyscale intensity. Following an initial plateau, which corresponds to the induction time of the phase transition, crystallization is observed by a reduction in the greyscale intensity, typically following a sigmoidal curve. On completion of crystallization, a plateau at a lower greyscale intensity is visible in the plots.


Page 22 of 35

Page 23 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 11. Time resolved plot of greyscale intensity for the 4 boxed areas, A-D, when mixed at (a) 20 rpm (b) 50 rpm (c) 150 rpm and (d) 300 rpm. The black line for each profile corresponds to the fit of a 4 parameter sigmoidal curve to the data. At the lower mixing speeds, there are notable differences between the profiles for regions A, B, C and D. The region closest to the mixing blade (A) is the first to exhibit a reduction in greyscale intensity, as crystallization starts around the stirrer. However, no differences are observed at the higher rpm, 300 rpm, when the crystal growth spreads quickly through the system. The plots are fitted to a 4-parameter sigmoidal curve:



𝑓 = 𝑦0 +

Page 24 of 35

𝑎

(

―(𝑥 ― 𝑥0)

1+ 𝑒

𝑏

(6)

)

where x0 is the midpoint, y0 is the minimum value, a is the maximum value and

1 𝑏

relates to the

slope of the curve which indicates the rate of crystallization. In all cases, there is an acceptable fit with R2 values greater than 0.98. 1

The change in 𝑏 and x0 across the regions A-D for the different mixing speeds are plotted in Figure 12, along with the change in induction time. The induction time is a sum of the relaxation time, collision time and time for the crystal to grow into a visible failure.24 The induction times for the different speeds and areas were estimated, across all replicates, by determining the crossover points between linear trendlines fitted to the induction period and those fitted to the slope of crystallization. The induction time relates to the point at which there is a reduction in greyscale intensity.


Page 25 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 12. Variation of (a)

1 𝑏

(steepness), (b) x0 (midpoint) and (c) induction time across the

different mixing speeds and boxed areas (A-D). The slope, related to the rate of crystallization, does not vary significantly with mixing speed, except for the area closest to the mixing blade (area A). The induction time and midpoint x0, are influenced by the mixing speed with both displaying a similar trend. Both parameters decrease as the mixing speed is increased. This can be attributed to a higher monomer collision frequency at higher mixing speeds. Since crystals are formed from the surfactant monomers,10 an increased number of collisions results in a higher probability of nucleation. The time to crystallization, and the midpoint, decrease as a result.



3.4 Comparison with the current test method For a selection of formulations (X1-X6) the failures times in the absence and presence of mixing were compared. In the absence of mixing, the time to crystallization was estimated from plots of homogeneity over time following the method described in Section 2.2.2.1, The induction time is related to the point at which the homogeneity begins to decrease. An example plot, for X6, is provided in Figure 13. After an average induction time of 0.63 days it is possible to observe a decrease in homogeneity, which relates to the onset of crystallization.

Figure 13. Change in homogeneity for the three X6 samples across a 7-day period at 0 °C. The green line indicates the point at which the homogeneity begins to decrease. In the presence of mixing, the time to crystallization was estimated by the change in greyscale intensity, as explained in Section 2.2.2.2. A relatively low mixing speed of 20 rpm was used. For each formulation, time-resolved greyscale plots for the boxed areas (A-D) were attained, as previously described for Formulation A. In line with mechanistic findings, crystals consistently formed around the mixer initially. Therefore, the boxed area closest to the mixer blade (area A)


Page 26 of 35

Page 27 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


was used to determine the time to crystallization, measured as the point at which the greyscale intensity starts to sharply reduce. The time to crystallization for Formulations X2-X6, in the absence and presence of mixing at 20 rpm, have been plotted against each other in Figure 14. Without agitation, X1 did not fail within the 7 day experimental run time, so this sample was omitted from the analysis. Despite the high degree of variability with the quiescent test, the two methods were found to be positively correlated, as shown in in Figure 14. Even though the two methods appear to exhibit different mechanisms, the same kinetic trend can be observed, which leaves scope for further investigation. This correlation is important for determining new specification limits if the new proposed method is to be implemented.

Figure 14. Correlation of failure times between the current and proposed low temperature stability tests across a range of dish liquid formulations. In the presence of mixing, the speed is 20 rpm. The application of mixing appears to be suitable for dish liquid formulations, as there is positive correlation between the current and proposed tests. The key aim of this study was to develop a more efficient test and was focused on dish liquids for commercial reasons. The dish liquid 26 ACS Paragon Plus Environment


products used within this study have similar rheology, since the viscosity is tailored to specification limits. Dish liquid is shear thinning at the range of temperatures investigated, with a typical rheology profile supplied in the supplementary information (Figure S3). The relevance of this work to other fluids is a subject of further study. In the case of shear thickening fluids or fluids with a high yield stress, the method may not be appropriate. 4. CONCLUSIONS Agitation is known to initiate crystallization in some systems, particularly in some consumer products and foods. In this study it has been shown for the first time that mixing can reduce the time to dish liquid crystallization, due to primary and secondary nucleation effects. This insight is not only important for developing accelerated stability tests, but also for exploring the effect of transportation on stability. Detergent formulations are likely to experience movement and vibration during the supply chain. In the absence of any mixing, crystallization was found to originate from the air-liquid interface due to the close-packed nature of the monolayer. In the presence of mixing, the crystal growth began from the mixing blade, the point of highest shear. An increase in mixing speed resulted in a decrease in both the induction time and the time to the midpoint of the crystallization process. A higher mixing speed also resulted in faster dispersion of crystalline entities through the bulk. This was attributed to an increase in the degree of crystal breakage, dispersion and boundary layer removal. The application of mixing also reduced the variability of the failure times. This was because in the presence of mixing the crystal growth was accelerated and spread in a cloud-like formation through the bulk whereas, in the absence of mixing, it was difficult to observe the individual crystal entities 27 ACS Paragon Plus Environment

Page 28 of 35

Page 29 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


forming. Two techniques were demonstrated as viable options to follow the crystallization process whilst mixing. The first involved using an experimental setup consisting of a transparent jacketed vessel and an overhead stirrer. Time lapse photography, in combination with greyscale analysis, was used to determine the point of crystallization. Alternatively, crystallization was detected through a reduction in light transmission passing through the sample. The sample vial cap was replaced with a mini-overhead stirrer to induce mixing in the sample. Extending the investigation to other formulations revealed that the time to crystallization under the current and the proposed test methods were positively correlated, despite proceeding via different mechanisms. These insights have a significant application in industry, with the potential for the acquired knowledge to be applied to the stability test methods at P&G. By using agitation, the tests will have shorter timescales and be more representative of the supply chain. Further work within P&G will include applying this new method to a larger number of formulations to determine the commercial suitability of this proposed route. AUTHOR INFORMATION Corresponding Author *Email: [email protected] Funding Sources The Engineering and Physical Sciences Council are gratefully acknowledged for their financial support (Grant Reference EP/G036713/1). ACKNOWLEDGMENT



Thanks go to the facilities at the P&G Brussels Innovation Center, at which some of the experiments were performed. Dr. Thomas Moxon is acknowledged for producing the MATLAB code for the greyscale analysis. REFERENCES (1) GrandViewResearch Soap and Detergent Market Size, Share & Trends Analysis Report By Product (Household Detergents, Industrial Soaps & Detergents, Household Soaps), Competitive Landscape, And Segment Forecasts, 2018 – 2025. https://www.grandviewresearch.com/industryanalysis/soap-detergent-market (accessed 21/05/2018). (2) Lai, K. Y. Liquid Detergents; CRC Press: New York, 1996. (3) Scheibel, J. J. The evolution of anionic surfactant technology to meet the requirements of the laundry detergent industry. J. Surfactants Deterg. 2004, 7, 319-328. (4) Bettiol, J. P., Morabet, S., Scialla, S. Alkaline liquid hand dish washing detergent composition. U.S Patent: 8343906. 2013. (5) Ofosu-asante, K. C., Clarke, J. M., Owens, R., Castro, A., Embleton, G. K. Hand washing detergent compositions. U.S Patent: 6521577. 2003. (6) Lamb, C. N., Lawson, J. R., Pearson, C. Liquid detergent compositions. U.S Patent: 4671894. 1987. (7) Singh, S. K.; Bajpai, M.; Tyagi, V., Amine oxides: a review. J. Oleo Sci. 2006, 55, 99-119. (8) Bechtold, T.; Mussak, R. Handbook of Natural Colorants. Wiley: Chichester, 2009. (9) Soontravanich, S.; Scamehorn, J. F. Use of a Nonionic Surfactant to Inhibit Precipitation of Anionic Surfactants by Calcium. J. Surfactants Deterg. 2009, 13, 13. (10) Stellner, K. L.; Scamehorn, J. F. Surfactant precipitation in aqueous-solutions containing mixtures of anionic and nonionic surfactants. J. Am. Oil Chem.' Soc. 1986, 63, 566-574. 29 ACS Paragon Plus Environment

Page 30 of 35

Page 31 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(11) Fan, X. J.; Stenius, P.; Kallay, N.; Matijevic, E. Precipitation of surfactant salts. 2. The effect of nonionic surfactants on precipitation of calcium dodecyl-sulfate. J. Colloid Interface Sci. 1988, 121, 571-578. (12) Shiau, B. J.; Harwell, J. H.; Scamehorn, J. F. Precipitation of mixtures of anionic and cationic surfactants. 3. Effect of added nonionic surfactant. J. Colloid Interface Sci. 1994, 167, 332-345. (13) Soontravanich, S.; Walsh, S.; Scamehorn, J. F.; Harwell, J. H.; Sabatini, D. A. Interaction between an anionic and an amphoteric surfactant. Part II: Precipitation. J. Surfactants Deterg. 2009, 12, 145-154. (14) Summerton, E.; Hollamby, M. J.; Zimbitas, G.; Snow, T.; Smith, A. J.; Sommertune, J.; Bettiol, J.; Jones, C.; Britton, M. M.; Bakalis, S. The impact of N,N-dimethyldodecylamine Noxide (DDAO) concentration on the crystallisation of sodium dodecyl sulfate (SDS) systems and the resulting changes to crystal structure, shape and the kinetics of crystal growth. J. Colloid Interface Sci. 2018, 527, 260-266. (15) Summerton, E.; Hollamby, M. J.; Le Duff, C. S.; Thompson, E. S.; Snow, T.; Smith, A. J.; Jones, C.; Bettiol, J.; Bakalis, S.; Britton, M. M. Nuclear magnetic resonance and small-angle Xray scattering studies of mixed sodium dodecyl sulfate and N,N-dimethyldodecylamine N-oxide aqueous systems performed at low temperatures. J. Colloid Interface Sci. 2019, 535, 1-7. (16) Saunders, D. Monitoring Shock and Vibration during the Transportation of Paintings. National Gallery Technical Bulletin 1998, 19, 64-73. (17) Miller, R. M.; Ces, O.; Brooks, N. J.; Robles, E. S. J.; Cabral, J. T. Crystallization of Sodium Dodecyl Sulfate-Water Micellar Solutions under Linear Cooling. Cryst. Growth Des. 2017, 17, 2428-2437. (18) Seiler, T. Operative Transportation Planning; Springer: Berlin Heidelberg, 2011



(19) Chantraine, F.; Viana, M.; Brielles, N.; Mondain-Monval, O.; Pouget, C.; Branlard, P.; Rubinstenn, G.; Chulia, D. Parametric study of surfactant effect on mechanical and dissolution properties of detergent tablets. J. Surfactants Deterg. 2006, 9, 267-277. (20) Falbe, J. Surfactants in Consumer Products: Theory, Technology and Application; Springer: Berlin Heidelberg, 2012. (21) Hartel, R. W.; Shastry, A. V. Sugar crystallization in food products. Crit. Rev. Food Sci. Nutr. 1991, 30, 49-112. (22) Hartel, R. W.; Ergun, R.; Vogel, S. Phase/State Transitions of Confectionery Sweeteners: Thermodynamic and Kinetic Aspects. Compr Rev Food Sci F. 2011, 10, 17-32. (23) McLeod, J. S.; Paterson, A. H. J.; Bronlund, J. E.; Jones, J. R. The effect of agitation on the nucleation of α-lactose monohydrate. Int. Dairy J. 2016, 61, 114-119. (24) Mullin, J. W. Crystallization; Butterworth-Heinemann: Oxford, 1997. (25) Mazzanti, G.; Guthrie, S. E.; Marangoni, A. G.; Idziak, S. H. J. A Conceptual Model for Shear-Induced Phase Behavior in Crystallizing Cocoa Butter. Cryst. Growth Des. 2007, 7, 12301241. (26) Rathee, V.; Krishnaswamy, R.; Pal, A.; Raghunathan, V. A.; Impéror-Clerc, M.; Pansu, B.; Sood, A. K. Reversible shear-induced crystallization above equilibrium freezing temperature in a lyotropic surfactant system. Proc. Natl. Acad. Sci. U.S.A. 2013, 110, 14849-14854. (27) Wu, Y. L.; Derks, D.; van Blaaderen, A.; Imhof, A. Melting and crystallization of colloidal hard-sphere suspensions under shear. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 10564-10569. (28) Maksimov, A. O.; Kaverin, A. M.; Baidakov, V. G. Heterogeneous Vapor Bubble Nucleation on a Rough Surface. Langmuir 2013, 29, 3924-3934.


Page 32 of 35

Page 33 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(29) Liang, K.; White, G.; Wilkinson, D.; Ford, L. J.; Roberts, K. J.; Wood, W. M. Examination of the process scale dependence of L-glutamic acid batch crystallized from supersaturated aqueous solutions in relation to reactor hydrodynamics. Ind. Eng. Chem. Res. 2004, 43, 1227-1234. (30) Callahan, C. J.; Ni, X.-W. An investigation into the effect of mixing on the secondary nucleation of sodium chlorate in a stirred tank and an oscillatory baffled crystallizer. CrystEngComm 2014, 16, 690-697. (31) Gudipaty, T.; Stamm, M. T.; Cheung, L. S. L.; Jiang, L.; Zohar, Y. Cluster formation and growth in microchannel flow of dilute particle suspensions. Microfluid. Nanofluid. 2011, 10, 661669. (32) Stamm, M. T.; Gudipaty, T.; Rush, C.; Jiang, L.; Zohar, Y. Particle aggregation rate in a microchannel due to a dilute suspension flow. Microfluid. Nanofluid. 2011, 11, 395. (33) Sicignano, L.; Tomaiuolo, G.; Perazzo, A.; Nolan, S. P.; Maffettone, P. L.; Guido, S. The effect of shear flow on microreactor clogging. Chem. Eng. J. 2018, 341, 639-647. (34) Sebastian, V.; Unnikrishnan, A.; Balakrishnan, K. Gray level co-occurrence matrices: generalisation and some new features. Int. J. Comput. Eng. Inf. Technol. 2012, 2, 151-157. (35) Rao, C. N.; Sastry, S. S.; Mallika, K.; Tiong, H. S.; Mahalakshmi, K. Co-occurrence matrix and its statistical features as an approach for identification of phase transitions of mesogens. Int. J. Innov. Res. Sci. Eng. Technol. 2013, 2, 4531-4538. (36) Goodwin, J. Colloids and Interfaces with Surfactants and Polymers: An Introduction; Wiley: Chichester, 2004. (37) Scamehorn, J. F. An Overview of Phenomena Involving Surfactant Mixtures. ACS Symp. Ser. 1986, 311, 1-27.



(38) Chen, Q.; Yao, J.; Hu, X.; Shen, J.; Sheng, Y.; Liu, H. Monolayer effect of a gemini surfactant with a rigid biphenyl spacer on its self-crystallization at the air/liquid interface. J. Appl. Crystallogr. 2015, 48, 728-735. (39) Hurle, D. T. J.; Jakeman, E. The Role of Convection and Fluid Flow in Solidification and Crystal Growth: Physicochemical Hydrodynamics; Pergamon Press: Oxford, 2014. (40) Shibata, O.; Nakahara, H.; Moroi, Y. New Adsorption Model—Theory, Phenomena and New Concept. J. Oleo Sci. 2015, 64, 1-8. (41) Moroi, Y.; Rusdi, M.; Kubo, I. Difference in Surface Properties between Insoluble Monolayer and Adsorbed Film from Kinetics of Water Evaporation and BAM Image. J. Phys. Chem. B 2004, 108, 6351-6358. (42) Graciaa, A.; Lachaise, J.; Sayous, J. G.; Grenier, P.; Yiv, S.; Schechter, R. S.; Wade, W. H. The partitioning of complex surfactant mixtures between oil/water/microemulsion phases at high surfactant concentrations. J. Colloid Interface Sci. 1983, 93, 474-486. (43) Sasaki, S. Metastable Crystalline Lamella of Cetylpyridinium Chloride in the Krafft Transition. J. Phys. Chem. B 2007, 111, 2473-2476. (44) Myerson, A. Handbook of Industrial Crystallization; Butterworth-Heinemann: Woburn, 2002. (45) Paul, E. L.; Atiemo-Obeng, V. A.; Kresta, S. M. Handbook of Industrial Mixing: Science and Practice; Wiley: New Jersey, 2004. (46) Liu, J.; Svärd, M.; Rasmuson, Å. C. Influence of Agitation on Primary Nucleation in Stirred Tank Crystallizers. Cryst. Growth Des. 2015, 15, 4177-4184. (47) Liu, J.; Rasmuson, Å. C. Influence of Agitation and Fluid Shear on Primary Nucleation in Solution. Cryst. Growth Des. 2013, 13, 4385-4394.


Page 34 of 35

Page 35 of 35 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(48) Liang, K.; White, G.; Wilkinson, D.; Ford, L. J.; Roberts, K. J.; Wood, W. M. L. An Examination into the Effect of Stirrer Material and Agitation Rate on the Nucleation of l-Glutamic Acid Batch Crystallized from Supersaturated Aqueous Solutions. Cryst. Growth Des. 2004, 4, 1039-1044.


Understanding the Crystallization Process in Detergent Formulations

Recommend Documents