See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225564834

Influence of formulation and processing variables on ball mill refining of milk

chocolate

Article  in  European Food Research and Technology · January 2006

DOI: 10.1007/s00217-006-0272-6

CITATIONS READS

43 3,853

3 authors:

Mara Lucisano Ernestina Casiraghi

University of Milan University of Milan
69 PUBLICATIONS   2,095 CITATIONS    100 PUBLICATIONS   3,110 CITATIONS   

SEE PROFILE SEE PROFILE

Manuela Mariotti

49 PUBLICATIONS   1,833 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Mara Lucisano on 04 June 2014.

The user has requested enhancement of the downloaded file.

Mara Lucisano · Ernestina Casiraghi ·
Manuela Mariotti

Influence of formulation and processing variables on ball mill

refining of milk chocolate

Received: 20 October 2005 / Revised: 10 January 2006 / Accepted: 19 January 2006 / Published online: 14 March 2006

Abstract The influence of both formulation and process- fully develop and turning the heterogeneous, flaky, dry re-
ing variables on the physical characteristics of milk choco- fined paste into a free-flowing suspension of solid particles
late has been studied during milling with a pilot ball mill. dispersed in a fat phase [1].
Nine experimental millings were carried out varying the Milk chocolate usually contains about 20% milk powder
type of dried milk used (roller, high free fat and spray- which affects the sensory characteristics of the final prod-
dried milk) and the time of lecithin and dried milk addition uct, the processing behaviour and the rheological proper-
during the refining process. For the samples collected dur- ties of fluid chocolate mass. A milk powder able to assure a
ing 2 h of milling, particle size, moisture, Casson plastic low viscosity and a low yield value at a given total fat con-
viscosity and yield value were evaluated. In all the experi- tent is particularly suitable for the production of chocolate
ments, the solid particle size was effectively reduced during when using the traditional technology [2]. The main milk
the first 120 min of refining; after this time the particle size products used for chocolate production are traditionally
reduction assumed an asymptotic trend. The addition of roller-dried and spray-dried milk products. More recently,
lecithin at the beginning of the refining kept the viscosity processes for enhancing the milk free fat content have been
low and constant throughout the process, thus allowing the proposed [3–5], moreover spray-dried milk with high free
addition of milk from the start of the process, regardless of fat content has been available in the market. The increase
the technology adopted for milk drying. of the free fat content makes the milk fat more functional
in milk chocolate production by reducing the amount of
Keywords Ball mill . Refining . Milk chocolate . cocoa butter required to obtain a low viscosity [5].
Chocolate production . Milk powder Traditional refining and conching could be replaced by
a process of ‘liquid’ refining carried out using a ball mill
which could also be combined with a liquefier and a mois-
Introduction ture removing device [1]. In this type of mill the chocolate
ingredients to be ground are placed in a stationary tank with
Chocolate traditional technology includes a phase during the grinding media. Carbon steel, stainless steel, chrome
which solid particle size is reduced. This phase, called refin- steel or ceramic balls are the commonly used media. The
ing, is carried out in a five roll mill with four of the cylinders mass and the balls are then agitated by a shaft with arms,
placed one above the other and a feed cylinder placed on rotating at a variable speed. During refining, the mass can
the side. The main aim of this operation is to reduce the also be recycled through the ball mill a number of times,
sugar crystals and the cocoa solids to such a size as to make thus going through a thick layer of balls, which are kept in
them small enough not to be detected on the tongue. This continuous movement and forced to bounce against each
operation is followed by a phase of chocolate conching, other. The solid particles are crushed into smaller pieces
during which the undesirable flavours are removed and the and, in the meantime, are partially covered by cocoa butter.
pleasant ones are produced, causing the chocolate aroma to This plant is currently proposed for cocoa nib grinding after
impact mill crushing for the production of cacao liquor,
M. Lucisano (
) · E. Casiraghi · M. Mariotti chocolate surrogates, creams for spreading, ice-creams
Dipartimento di Scienze e Tecnologie Alimentari e and biscuit coatings and compounds. In medium–small
Microbiologiche, University of Milan, size companies this technology can be considered as a
Via G. Celoria 2,
20133 Milan, Italy
valid alternative to the traditional lines which are more
e-mail: [email protected] cumbersome and expensive with regards to investment,
Tel.: +39-02-50316628 conduction and energy consumption. Lehmann Maschi-
Fax: +39-02-50316632 nenfabrik GmbH (Aalen, Germany) has coupled ball mills
with in-line liquefiers to set up a continuous chocolate and samples collected during refining. Chocolate at 45 ◦ C
production process in which the ball mill breaks up the (approximately 2 g) was dispersed in 2 g of white mineral
solid particles and the in-line liquefier coats the newly oil before being placed on the jaws of the micrometer.
broken surfaces with fat and reduces the viscosity [1]. Measurements taken at various times during refining, were
Nevertheless, very few articles have been published dealing replicated 10 times.
with chocolate refining by means of ball mills [6–8].
The aim of this study was to evaluate the influence of the
addition of various types of dried milk on the rheological Experimental tests
characteristics of chocolate and to investigate the effect
of adding lecithin and milk at different times during the All the experimental tests were performed on 8 kg batches
refining process. of chocolate with a WA-FA SFERE 20-ball mill (Mazzetti
Renato s.r.l., Milan, Italy) containing 6.5 mm diameter
stainless steel balls. Refining was carried out at 45 ◦ C at an
Materials and methods agitator shaft speed of 64 rpm, recycling the mass through
the ball bed at a medium flow of 3–4 kg/min of the recy-
Raw materials cling pump, for 180 min. At various time intervals during
the process, samples were taken out to evaluate the char-
In this research, milk chocolate was produced using a acteristics of the chocolate being refined. The first particle
pilot plant of an Italian chocolate manufacturer, making size measurement was taken after 20 min. During the first
use of a formulation containing cocoa powder instead of 60 min of the process it was not possible to determine the
cocoa liquor, the more common ingredient for chocolate viscosity due to the excessive size of the particles.
production. The following raw materials were considered: Table 1 reports the formulation and the timing of the
alkalised cocoa powder 10/12, with 0.5% of the particles three series of experiments performed with the three types
having a size greater than 75 µm (Barry Callebaut, of powdered milk: roller-dried milk (tests 1, 4 and 7), HFF
Lebbeke-Wieze, Belgium), deodorized cocoa butter spray-dried milk (tests 2, 5 and 8) and spray-dried milk
(Barry Callebaut, Lebbeke-Wieze, Belgium), powdered (tests 3, 6 and 9). All formulations contained 31% cacao
milk, medium-grain sugar (Südzucker AG, Mannheim, butter, introduced at the beginning of the refining as a high
Germany), soy lecithin with a minimum insoluble con- fat content during grinding is necessary as the ball mills
tent of 62% in acetone and a phospholipid content of require a liquid feed.
approximately 40% (Central Soya European Lecithins The milk powder was added either after 85 min (tests
Italia Srl, Italy) and vanillin (Eurovanillin Supreme, 1, 2 and 3) or together with the other ingredients at the
Borregaard Synthesis, Norway). Three different types of beginning of the refining process (tests 4, 5, 6, 7, 8 and 9).
powdered milk were used, all containing 26% of fat and In the refining tests from 1 to 6, lecithin (0.4 g) was divided
26% of protein: roller-dried milk (Hofmeister Kasewerk, into two equal amounts and added after 75 and 120 min;
Germany), spray-dried milk with a high free fat content in tests 7, 8 and 9 the lecithin was added all at once at the
(HFF) (PROCHOC S 26 + ) produced by Ingredia (Arras beginning of the refining.
Cedex, France) and spray-dried milk (BHA Belgium S.A.
Hombourg, Belgium).
Results and discussion
Analytical methods
Particle size reduction
The moisture content in the chocolate samples was de-
termined in accordance with the official O.I.C.C. method A micrometer was adopted to evaluate the chocolate parti-
E/1952 [9]. The moisture content of dried milk was deter- cle size. According to Ziegler and Hogg [12], the microme-
mined according to the official standard method FIL-IDF ter measurements approximate the d90 value, i.e. 90% finer
26A:193 [10]. than this size, as measured by laser light scattering. Tests 1,
The flow properties of the melted chocolate were evalu- 2 and 3 were carried out to evaluate the influence of the type
ated using a Böhlin VOR Rheometer (Böhlin Reologi AB of powdered milk—roller milk, HFF and spray-dried—on
Corporate Headquarters, Lund, Sweden). The tests were the particle size reduction of the chocolate during the refin-
carried out at 40 ◦ C using a coaxial cylinder system (C25) ing process when lecithin was added in two equal amounts
and applying a shear rate varying from 0.232 to 46.1 s−1 . (after 75 and 120 min) and milk 85 min after the start of the
To determine the Casson plastic viscosity (ηCA ) and Cas- process (Fig. 1A). Milk addition after 85 min of refining
son yield value (τ CA ), Casson’s equation was used [11]. allowed a smaller mass of product with a lower viscosity
The analysis was run in triplicate for different chocolate to be processed so that it could be recycled through the ball
aliquots, then separate fittings to Casson equation were layer a large number of times. There was a rapid reduction
considered and the average values were reported. of the chocolate particle size during the first 85 min of the
A micrometer produced by NSK Ltd. (Tokyo, Japan) process; after that time, reduction became progressively
was used to assess the particle size of the raw materials slower. An increase in the particle size was noted upon the
 Table 1 Formulation and experimental timing of the tests

Lecithin and milk addition are indicated by arrows; sampling times are reported in bold
a
1, 4, 7: roller-dried milk; 2, 5, 8: HFF spray-dried milk; 3, 6, 9: spray-dried milk

addition of milk up to a dimension of 68 µm for roller milk, an intermediate trend with respect to the series of tests 1,
122 µm for HFF milk and 101 µm for spray milk. 2, 3 and 4, 5, 6 in the first hour of refining. After 130 min
When the milk was present in the formulation at the the curves of all the tests could be superimposed. At the
beginning of the process (tests 4, 5 and 6), the three samples end of the process, the final size of the particles was close
had a higher particle size during the first 60 min of refining to the threshold for sensory perception, which is generally
that appeared different for the three types of powdered milk considered to be 25–40 µm [12].
(Fig. 1B). As the process proceeded and lecithin was added,
the reduction in particle size became rapid, and the curves
of the various formulations became very similar. Rheological and moisture evaluation
The reduction in particle size in the experiments per-
formed with all the ingredients mixed together at the be- In tests 1, 2 and 3 an increase in viscosity was observed up
ginning of the process (test 7, 8 and 9, Fig. 1C) presented to the first addition of lecithin (Fig. 2A). This increase was
 (A) 500 (A)
450 Test 1 2.0

Casson plastic viscosity (Pa s)

400 Test 2 1.8
particle size (µm)

350 Test 3 1.6

300 1.4
250 1.2
1.0
200
0.8
150
0.6 Test 1
100
0.4 Test 2
50
0.2 Test 3
0
0.0
0 20 40 60 80 100 120 140 160 180 200
40 60 80 100 120 140 160 180 200
refining time (min)
refining time (min)
(B) 500 (B)
450 Test 4 4.0

Casson plastic viscosity (Pa s)

400 Test 5 3.6 Test 4
particle size (µm)

350 Test 6 3.2 Test 5

300 2.8 Test 6
250 2.4
200 2.0
150 1.6
100 1.2
50 0.8
0 0.4
0 20 40 60 80 100 120 140 160 180 200 0.0
40 60 80 100 120 140 160 180 200
refining time (min)
refining time (min)
(C) 500
Test 7 (C)
450 2.0
Casson plastic viscosity (Pa s)

400 Test 8 1.8

particle size (µm)

350 Test 9 1.6

300 1.4
250 1.2
200 1.0
150 0.8
100 0.6 Test 7
50 0.4 Test 8
0 0.2 Test 9
0.0
0 20 40 60 80 100 120 140 160 180 200
40 60 80 100 120 140 160 180 200
refining time (min)
refining time (min)
Fig. 1 Particle size reduction during refining
Fig. 2 Casson plastic viscosity during refining

connected with the exponential growth in the specific sur-

face of the solid particles which, as they were not immedi- increased from about 0.5–0.6% to 1.0–1.3% upon the ad-
ately covered with the cocoa butter, caused the chocolate to dition of milk (Fig. 3A). The addition of roller milk caused
flow with great difficulty. When the first amount of lecithin the highest moisture increase and a greater difficulty in
was introduced (75 min after the start of the refining pro- removing water during the subsequent part of the refin-
cess), the viscosity decreased rapidly as the emulsifier with ing process. The viscosity of the products decreased again
its hydrophilic groups combined with the water molecules when a second aliquot of lecithin was added (120 min after
absorbed on the surface of the sugar, producing a lubri- the refining start). At the end of the process, the size reduc-
cating action [13]. After the addition of milk, an increase tion and the consequent increase in specific surface of the
in viscosity was observed due to the presence of the milk chocolate solids caused the viscosity to remain constant or
solids that gave the chocolate mass a small amount of water, slightly increase, a phenomenon that could be attributed to
critical to the chocolate viscosity. In fact, the three types a re-aggregation trend [14].
of milk contained 2.5, 2.0 and 2.2 g/100 g moisture for Viscosity in tests 4, 5 and 6 showed higher values at the
roller-dried, high free fat and spray-dried, respectively. In- beginning of the process due to the presence of powdered
deed, the moisture content of the mass under refinement milk in the formulation. Also in these tests, a reduction in
 (A) (A)
2.0 40.0
1.8 Test 1 Test 1
35.0

Casson yield value (Pa)

Test 2 Test 2
1.6
Test 3 30.0 Test 3
1.4
moisture (%)

25.0
1.2
1.0 20.0
0.8 15.0
0.6
10.0
0.4
5.0
0.2
0.0 0.0
40 60 80 100 120 140 160 180 200 40 60 80 100 120 140 160 180 200

refining time (min) refining time (min)

(B) (B)
2.0 40.0
1.8 Test 4 Test 4
35.0

Casson yield value (Pa)

1.6 Test 5 Test 5
Test 6 30.0
1.4 Test 6
moisture (%)

1.2 25.0
1.0 20.0
0.8 15.0
0.6
10.0
0.4
5.0
0.2
0.0 0.0
40 60 80 100 120 140 160 180 200 40 60 80 100 120 140 160 180 200
refining time (min) refining time (min)

(C) 2.0 (C) 40.0

1.8 Test 7 Test 7
35.0
Casson yield value (Pa)

1.6 Test 8 Test 8

Test 9 30.0
1.4 Test 9
moisture (%)

1.2 25.0
1.0 20.0
0.8 15.0
0.6
10.0
0.4
0.2 5.0
0.0 0.0
40 60 80 100 120 140 160 180 200 40 60 80 100 120 140 160 180 200
refining time (min) refining time (min)
Fig. 3 Moisture variation during refining Fig. 4 Casson yield value during refining

viscosity was observed after the two additions of lecithin high free fat milk was used (HFF spray milk), the viscosity
(Fig. 2B). At the end of the refining process, the viscos- was slightly lower than in the formulations containing the
ity of all the formulations was similar, with the exception other two types of milk, especially during the first 80 min
of chocolate containing roller milk, which had higher val- of refining when milk was present at beginning of the pro-
ues throughout the refining period. This result appears to cess without any lecithin (Fig. 2B). Viscosity differences
contradict the data found in literature [1, 13]; indeed, the can again be related to the moisture content of the refining
greater percentage of free fat in the roller milk should deter- mass (Fig. 3B) that generally remains higher in roller-dried
mine a lower viscosity than that obtained using spray-dried milk.
milk as reported by Dewettinck et al. [15]. Our results must The effectiveness of lecithin in determining a rapid reduc-
probably be ascribed to the different refining system used tion in viscosity was also demonstrated in experiments 7, 8
in this research. In fact, as proved by Franke et al. [8], dur- and 9 where both lecithin and powdered milk were added at
ing ball refining milk powders release a large part of their the beginning of the refining process with the aim of keep-
internal fat, thus making the three different kinds of milk ing the viscosity low right from the beginning of the process
more similar with regard to this aspect. In general, when the (Fig. 2C). Right from the start and over the whole refining
process, viscosity values were between 1.4 and 1.6 Pa·s. Moreover, the addition of lecithin at the start of refin-
No differences were observed among the three chocolate ing was a decisive factor in keeping the viscosity and the
formulations. If lecithin is added at the beginning of the pro- yield value low. Adopting this approach, regardless of the
cess, sugar crystals are rapidly coated by fat immediately drying technology applied, the milk could be introduced
after size reduction, thanks to the emulsifying properties of into the formulation right at the beginning of the process,
lecithin. In fact, when lecithin is added, its dipolar nature simplifying the industrial production process and keeping
allows cocoa butter and sugar to become intimately associ- the viscosity low and constant throughout the process.
ated with each other. Thus, the cocoa butter coats the sugar
surface more efficiently, resulting in a portion of the cocoa
butter available for decreasing the viscosity of the choco- References
late [16]. In this last case the final moisture of the chocolate
remained constant but higher than in the other tests even if 1. Beckett ST (1999) Conching. In: Beckett ST (ed) Industrial
chocolate manufacture and use, 3rd edn. Blackwell Science,
the viscosities were very similar (Fig. 3C). This effect was UK, pp 153–181
caused by the formation of a thin film of fat on the surface 2. Haylock SJ, Dodds TM (1999) Ingredients from milk. In:
of sugar crystals limiting water removal from the mass. Beckett ST (ed) Industrial chocolate manufacture and use, 3rd
As a general trend, the Casson yield value increased as edn. Blackwell Science, UK, pp 57–77
the chocolate particle size was reduced (Fig. 4). According 3. King N (1965) Dairy Sci Abstr 27:91–140
4. Bohrin HU, Kuypers W, Meister N (1987) A process for the
to Prentice [17], when particle size decreases the number production of a milk powder. UK Patent GB 21831356 A
of bonds and consequently the amount of frictional contact 5. Koc AB, Heinemann PH, Ziegler GR (2003) J Food Sci 68:210–
between the particles increases, causing higher values of 216
τ CA . When dried milk was added after 85 min of refining 6. Lucisano M, Casiraghi E, Mariotti M (2001) In: Chiriotti (ed)
Proceedings of the 5th CISETA, V, pp 739–747
(test 1, 2 and 3), a more rapid increase of the yield value 7. Szegvari A, Yang M (1999) In: Proceedings of the seminar on
was observed in particular with roller-dried milk. In any dispersion of pigments and resins in fluid media, pp 1–7
case, the chocolate obtained with this procedure presented 8. Franke K, Scheruhn E, Tscheuschner HD (2002) Milchwis-
acceptable flow properties for the production of chocolate senschaft 57(9/10):535–539
9. International Office of Cocoa, Chocolate and Sugar Con-
bars [18]. After 3 h of refining, chocolate containing high fectionery (I.O.C.C.C.) (1952) Analytical method 3. Meulan
free fat milk had the lowest yield value. Similar results were Hardricourt, France
obtained in tests 7, 8 and 9, while very high values resulted 10. Federation International de Laiterie–International Dairy Federa-
when refining was carried out in the presence of spray or tion (FIL-IDF) (1993) Norme FIL Internationale 26A. Bruxelles,
roller-dried milk from the beginning of the process, without Belgium
11. International Office of Cocoa, Chocolate and Sugar Confec-
any lecithin added in the first 75 min of refining (test 4, 5 tionery (I.O.C.C.C.) (1973) Analytical method 10. Meulan
and 6). Hardricourt, France
12. Ziegler GR, Hogg R (1999) Particle size reduction. In:
Beckett ST (ed) Industrial chocolate manufacture and use, 3rd
edn. Blackwell Science, UK, pp 115–136
13. Fincke A (1965) Handbuch der Kakaoerzeugnisse. Springer-
Conclusion Verlag, Berlin Heidelberg New York
14. Datei L (2004) Ottimizzazione dei parametri di lavorazione di un
In all the experiments the particle size was effectively re- mulino a sfere per la raffinazione del cioccolato. Master degree
duced during the first 120 min of refining, although there thesis of Politecnico di Milano
15. Dewettinck K, De Moor H, Huyghebaert A (1996) Milchwis-
were small differences connected with the formulation and senschaft 51(1):25–28
the production process. After 2 h, the particle size assumed 16. Dimick PS, Hoskin JM (1981) Can Inst Food Sci Technol J
an asymptotic trend: obtaining a further significant and de- 14(4):269–282
sirable reduction in size, even by prolonging the refining 17. Prentice JH (1984) Measurements in the rheology of foodstuffs.
Elsevier Applied Science Publishers, London
times, becomes a problem. This problem could probably be 18. Pontillon J (1998) Cacao et chocolat. Production Utilisation Car-
solved by using two different plants in series, the second acteristiques. Tec & Doc-Lavoisier, Paris
one fitted with smaller balls down to as small as 2 mm. In
fact, as smaller balls fit into the same volume, an increased
number of collisions will be expected.

View publication stats