Ultrasonics Sonochemistry 11 (2004) 251–255

www.elsevier.com/locate/ultsonch

The effect of ultrasonic intensity on the crystal structure of palm oil

Maria Patrick a,*, Renoo Blindt a, Jo Janssen b

a Unilever Research and Development, Colworth House, Sharnbrook MK44 1LQ, UKb Unilever Research and Development, Olivier van Noortland 120, 3133 AT Vlaardingen, UK

Abstract

It has been known for a long time that both the crystal structure and kinetics of crystallisation can be affected by ultrasound. In

the past systems used have relied on high power ultrasonic probes to produce crystals. The majority of these probes produce

cavitation in the system and it has been difficult to differentiate between effects caused by the ultrasound alone or by the cavitation

produced by ultrasound on the crystal structure. Some materials, such as fats, are very susceptible to the production of free radicals

that lead to ‘‘off-flavours’’ being obtained. These ‘‘off-flavours’’ are easily produced when the standard high power probes are used.

This has meant that, although the crystal structure of the final product might be improved, the presence of ‘off’ flavours has

prevented ultrasound being considered as a commercial technique for the crystallisation of edible fats.

At Unilever R&D a system has been developed which can investigate the effect of ultrasound on the crystallisation of fats under

controlled conditions covering a range of intensities and cooling rates. The intensity levels used were both below and above the

cavitational threshold. By keeping the cooling regime constant it has been possible to show that the structure of the final product can

vary from a material looking similar to cottage cheese through to a fine cream simply by varying the ultrasonic intensity. This paper

describes the effect of ultrasound on both the crystal structure and kinetics of palm oil crystallisation at intensities below and above

the cavitational threshold.

� 2004 Elsevier B.V. All rights reserved.

Keyword: Sonocrystallisation

1. Introduction

Ultrasound enhanced crystallisation (sonocrystalli-

sation) was first observed in 1927 when ultrasound was

applied to a supersaturated thiosulphate solution. Sincethen many other studies have taken place with different

systems such as sugar solutions, waxes, water, fats and

supercooled melts, resulting in several large scale

applications in areas such as metallurgy and sugar

crystallisation. Although ultrasound has been widely

used in this area there have been differing views as to the

basic physical mechanisms. Many authors have differing

views on the reasons for sonocrystallisation but themajority agree that cavitation plays an important part in

the process [1,2]. The presence of cavitation has limited

the use of use of ultrasound with materials like fats due

to adverse sonochemical reactions that can occur.

In many instances ultrasound technology has been

treated as a ‘black box’ and as a result attempts to ex-

*Corresponding author.

1350-4177/$ - see front matter � 2004 Elsevier B.V. All rights reserved.

doi:10.1016/j.ultsonch.2004.01.017

ploit the technology have failed due to a lack of

understanding of the highly coupled nature of the

transducer/product/field interactions. Unilever R&D,

working in conjunction with the ISVR Group at

Southampton University, led by Professor Tim Leigh-ton, has designed an ultrasonic cell in which the field is

both well defined and measurable. This cell has enabled

us to develop a unique experimental capability that can

study the effects of ultrasound at levels both below and

above the cavitational threshold. The presentation that

follows will review the results of fat crystallisation

studies on palm oil carried out using this cell.

2. Cell description and experimental methodology

All the experiments were performed in a thermo-

statically controlled cylindrical batch cell with ring

transducers and stirrer (vertical blades moving along the

wall over the full height of the cell) shown in Fig. 1. The

ring transducers produce a well defined, radial field

Fig. 2. Typical crystals obtained from control sample showing a mix of

large and small dense crystals.

Fig. 1. The ultrasonic ring-transducer cell (inner diameter 10 cm,

volume of liquid about 700 ml, width of the blades 5 mm).

252 M. Patrick et al. / Ultrasonics Sonochemistry 11 (2004) 251–255

down the axis of the cell. The field is measured and

calibrated using a B&K 8105 Hydrophone. The cell is

resonant and its resonant frequency is dependent upon

both the geometry of the cell and the sample contained

within it.The cooling of the samples was controlled by circu-

lating a low viscosity oil from a thermostatically con-

trolled bath equipped with a Haake C25 programmable

temperature controller. Oil rather than water had to be

used as the coolant because of its direct contact with the

transducers, which are driven by a high AC voltage. The

intensity of the transducers was varied using a B&K

amplifier, the control knobs being calibrated in dB val-ues.

Initial experiments were carried out to determine the

cavitational threshold and calibration runs were carried

out to determine the resonant frequency for the system.

The experiments were performed with a PO103

(Summa). This palm oil is a commercially available

product supplied by Loders Croklaan. This blend was

heated to a temperature of 65 �C and held at this tem-perature for a minimum of 2 h to remove any memory

effect of crystallisation. The liquid was then cooled in a

controlled manner from 65 to 23 �C over a period of 6 h.

During this time the palm oil was stirred to maintain a

uniform temperature throughout the cell but once the

final temperature was reached the stirring was stopped

and the palm oil allowed to crystallise overnight. The

crystals were then examined under the microscope. Thecontrolled cooling regime was repeated for fresh samples

of the test blend with ultrasound being applied at dif-

ferent intensities, the ultrasound being turned on when a

temperature of 45 �C was reached and stopped when the

blend reached a temperature of 23 �C. The temperature

of the palm oil was maintained at 23 �C overnight. The

ultrasound that was applied was at intensities of 30, 35,

40 and 45 dB. These intensities were both below and

above the cavitational threshold of 41 dB for this

product. Samples of the crystals formed were placed on

a microscope slide and examined using a Leica DMRB

microscope with the polarised light and a k plate con-figuration. The initial images were obtained using ·10magnification. The crystals formed obtained are shown

in the following figures.

3. Results

The crystals formed with the control when no ultra-sound was applied contained a mixture of large and

smaller spherulitic crystals. A typical example of the

crystals is shown in Fig. 2. This product was pourable

but opaque and resembled a thick batter in appearance.

Although there was some free liquid, the product was

not easy to filter and it did not separate easily.

Applying ultrasound to the sample had a dramatic

effect, even at the lowest level of 30 dB. The samplebecame very viscous, and had to be spooned out of the

cell. There was very little free liquid and the sample had

a clotted cream type of texture. Fig. 3 shows the

microscope image obtained. The crystals were of a more

uniform size than the control and were denser. Exami-

nation of the matrix surrounding the crystals showed

that it had the appearance of being filled by a network of

hairs or needles.Increasing the intensity to 35 dB (again below the

cavitational threshold) resulted in a different product

altogether. The ultrasonic cell contained a clear liquid

with the majority of the crystals in a layer at the base,

with some crystals floating at the top of the cell and

some suspended in the liquid. Fig. 4 shows typical

crystals obtained form the top (a), middle (b) and base

(c) of the cell.

Fig. 3. Crystals obtained at an ultrasonic intensity of 30 dB showing

the dense uniform crystals surrounded by a matrix consisting of long

hair or needle-like structures.

Fig. 5. Crystals obtained at an ultrasonic intensity of 40 dB showing

typical crystal structure.

Fig. 6. Crystals obtained at an ultrasonic intensity of 45 dB showing

typical ‘cotton wool’ area.

M. Patrick et al. / Ultrasonics Sonochemistry 11 (2004) 251–255 253

As can be seen from Fig. 4(a) the crystals floating on

the surface are very small. The sample taken from thesuspension contained a mixture of clumps of small

crystals as well as crystal structures similar to those seen

on the surface. The sample taken from the bottom of the

cell consisted totally of crystals that had clumped to-

gether. If a quantity of this mix was placed in a small

container and then shaken the crystals could be visibly

seen to fall, suggesting that this intensity could be used

for processes such as fractionation of oils.At an intensity of 40 dB the structure was again very

different to that seen at lower intensities. The sample

was very smooth and of a consistency similar to face

cream. The crystals were very uniform and small, with

no free liquid present as can be seen from Fig. 5. The

figure below shows the structure at a ·10 magnification

but on increasing the magnification to ·40 it was pos-

sible to show that the ‘crystals’ at ·10 magnificationwere made up of even smaller crystals.

At intensities greater than 41 dB cavitation occurred

and could be observed both visually and physically

using an oscilloscope to show the presence of sub-

harmonics and noise. Increasing the intensity to 45 dB

Fig. 4. Crystals obtained at an ultrasonic intensity of 35 dB from

ensured that the cell was working well above the cavi-

tational threshold and microscopic examination of typ-

ical crystals formed (Fig. 6) were very similar to those

when no ultrasound was applied. There were, however,

large areas that had a cotton wool type of appearanceand no clear crystal structure could be seen, even at ·40magnification.

the top (a), middle (b) and base (c) of the ultrasonic cell.

254 M. Patrick et al. / Ultrasonics Sonochemistry 11 (2004) 251–255

4. Kinetics and ‘‘off-flavour’’ studies

4.1. Kinetics

Applying ultrasound during the cooling process alsoaffected the kinetics of crystallisation. The temperature

of the palm oil was monitored both during the cooling

and application of ultrasound by thermocouples in the

liquid. There were three thermocouples in the oil en-

abling the temperature of the material to be monitored

near to the surface, at the centre and near the base of the

cell. Table 1 gives both the temperature and time at

which crystallisation occurs during the cooling process.The application of ultrasound above 30 dB produced a

dramatic decrease in the onset of crystallisation with an

optimum time occurring at subcavitational levels just

below the cavitational threshold.

Although there are marked differences in the crystal

structure between the control and the crystals produced

by applying ultrasound at 30 dB there is very little dif-

2.000 4.000 6.000 8.000 10.000 12.0000

100

%

0

100

%

0

100

%

V0435

V0437

V0436

20kHz/5 min Branson probe

66kHz/5 min ring transducers

Control oil

Fig. 7. A mass spectrograph comparing sunflower oil (bottom) with sunflowe

probe (top).

Table 1

Table showing the effect of ultrasound on the time taken for crystal-

lisation to occur and the temperature at which crystallisation occurred

Ultrasound intensity

(dB)

Onset of

crystallisation

Temperature of

crystallisation (�C)

None 6.2 26

30 6.0 28

35 5 34

40 4.6 36

45 4.8 33

ference in both the time and temperature of crystallisa-

tion. This indicates that there appears to be a threshold

below which the kinetics are not altered. There also

appears to be a minimum time and maximum temper-

ature at which the onset of crystallisation is recorded.This occurs at the 40 dB level and suggests that the

system appears to be working at the most efficient level

at intensities just below the cavitational threshold.

4.2. Off-flavours

As part of this study some work has also been con-

ducted on the production of ultrasonically induced ‘‘off-flavours’’ in fats. It is widely believed in the literature

that sonochemistry is directly related to cavitation [2,3].

The collapse of the ultrasound-induced cavities can

produce very high local temperatures and pressures,

which have been suggested to cause free radical pro-

duction in water. With fats and oils an oxidation process

could occur resulting in the break-up of pre-existing

lipo(hydro)peroxides. Some oils contain more unsatu-rated C@C bonds than others and are therefore more

susceptible to off-flavours when sonicated. This has

limited the use of ultrasound as a commercial technique

for the production of edible fats. As sunflower oil oxi-

dises more readily than palm oil when sonicated it was

used as a test material for off-flavour studies in the

ultrasonic cell. The sunflower oil used was a commer-

cially available product and was purchased in 2 l bottles.The fixed volume of sunflower oil was sonicated in

the cell using either the ring transducers or a commer-

14.000 16.000 18.000 20.000 22.000 24.000rt

Scan EI+ 35_3006.00e6

Scan EI+ 35_3006.00e6

Scan EI+ 35_3006.00e6

r oil sonicated with the ring transducer (middle) and with the Branson

M. Patrick et al. / Ultrasonics Sonochemistry 11 (2004) 251–255 255

cially available ultrasonic (Branson) probe placed in the

cell. The temperature of the oil kept constant during this

time. All samples were then analysed for ‘‘off-flavours’’

using mass spectroscopy as attempts at using other

methods, apart from the nose, failed to differentiatebetween any of the samples. The mass spectroscopy

analysis was performed on volatile compounds present

in the control and sonicated sunflower oil extracted

using the Likens–Nickerson method.

Using the lowest power setting with the Branson

probe equipment the sunflower oil from a freshly opened

bottle was sonicated for a period of 5 min at a nominal

frequency of 20 kHz. Samples of the sonicated oil weretaken for analysis and the ultrasonic cell was refilled

with a fresh quantity of sunflower oil that was sonicated

using the ring transducers. The sample was sonicated at

66 kHz for a period of 15 min at the maximum possible

level ensuring that cavitation was present.

The mass spectroscopy results obtained (Fig. 7)

showed that although no ‘‘off-flavours’’ were produced

using the ring transducers, a peak occurred showing thepresence of methyl methacrylate or Perspex. As the cell

is made from Perspex it is evident that the ring trans-

ducers are working but that, even at cavitational levels,

no detectable ‘‘off-flavours’’ were produced. The sample

sonicated with the Branson probe, in the cell, on the

other hand, yielded a number of oxidation products one

of them being identified as benzene, although in a very

small quantities. This result was unexpected but possi-ble. Randall and Wiltshire [4] have pointed out that

adverse sonochemistry in oils and fats caused by intense

sonication could lead to the formation of (possibly

carcinogenic) ring structures from fatty acids, although

the amounts produced are small.

5. Discussion and conclusion

A cell has been developed which can be used to apply

ultrasound to liquids, including emulsions, at levels both

below and above the cavitational threshold. Using this

cell we have shown that it is possible to change both the

crystal structure and kinetics of the palm oil by changing

the intensity of the ultrasound applied to the cell.

The kinetics of crystallisation appears to be verystrongly influenced by the ultrasound above a minimum

threshold level. The results obtained also show that at a

level just below the cavitational threshold crystallisation

occurs in the minimum time implying that the system is

working at its most efficient level at this point.

The crystal structure appears to be highly dependent

upon the intensity of the ultrasound with the resultant

samples having different properties dependent upon theintensity of ultrasound applied. The same material (in

this case palm oil) can be used to yield a variety of

textures from something that resembles clotted cream to

a smooth cream, similar to a face cream simply by

altering the intensity at which the ultrasound is applied,

provided it is below the cavitational threshold. When the

ultrasound was applied at the 30 dB level this had verylittle effect on the kinetics but it appeared as if the

ultrasound had prevented the large spherulitic crystals,

observed in the control, from forming. The network

surrounding the crystals, when ultrasound was applied

at the 30 dB level, appeared to be made up of needle like

structures and it has been hypothesised that the ultra-

sound has caused an increase in the number of nucle-

ation sites [5] limiting the size to which the crystals cangrow. The matrix of long needle like structures are

thought to be due to lack of space for either the crystals

to grow or needles that have been knocked off larger

crystals during the crystallisation process. Increasing the

ultrasonic intensity to 35 dB has produced smaller more

uniform crystals that tend to clump together and fall to

the bottom of cell. A further increase in intensity to 40

dB, below the cavitational threshold, produced a uni-form product of very small crystals suggesting that there

were numerous nucleation sites that produced crystals

simultaneously. The increase in the applied ultrasound

to above the cavitational threshold gave areas where no

clear crystal structure could be seen and it is hypothes-

ised that these are areas where cavitation has its maxi-

mum effect.

Through this work we have been able to illustratethat, once the range of crystal structures has been

determined it is possible to select a particular texture

simply by choosing the correct ultrasonic intensity.

This cell has also shown that it is possible to use

ultrasound with fats without producing the adverse so-

nochemical reactions that typically occur with very high

power probes [4]. This gives us the ability to structure

materials simply by choosing the correct power levels forultrasonic treatment.

References

[1] O.V. Abramov, High intensity ultrasonics, Theory and Industrial

Applications, Gordon and Breach Science Publishers, Overseas

Publishers Association.

[2] T.J. Mason, in: M. Povey, T.J. Mason (Eds.), Power Ultrasound in

Food Processing––the Way Forward, Blackie Academic & Profes-

sional, London, 1998.

[3] T.J. Mason, Practical Sonochemistry. Users Guide to Applications

in Chemistry and Chemical Engineering, Ellis Horwood, Chiches-

ter, 1991.

[4] N. Randall, M.P. Wiltshire, Food sonochemistry and sonoprocess-

ing. Use of high intensity ultrasound in the food industry, Scientific

and Technical Survey No. 169, Leatherhead Food Research

Association, November 1989.

[5] P. Walstra, Fat crystallisation, in: J.M.V. Blanshard, P.J. Lill-

ford (Eds.), Food Structure and Behaviour, Academic Press,

1987.