abstract

C.+.41 (2 3'$ ,.23 (,/.13-3 ".,,$1"(+ 24&1 331(!43$ !43 (- )4("$ "+1(%("3(.- (32 1$,.5+ (2 424++8 -.3
".-2(#$1$# ,.-& /1(,18 .!)$"3(5$2. H.6$5$1, !2$# .- 1$24+32 /1$2$-3$#, ++ 23-#1# "+1(%("3(.- /1."$#41$2
'5$ 3'$ /.3$-3(+ %.1 2(&-(%("-3+8 '(&'$1 1$,.5+ .% ".+.41 3'- (2 1$+(2$# (- 3'$ (-#4231(+ /1"3("$. F.41 /1(-"(/+
)4("$ "+1(%("3(.- /1."$#41$2, 5(9. #$%$"3(.- !8 '.3 +(,(-&, 24+%(33(.-, "1!.-3(.- -# #.4!+$-"1!.-3(.- 6$1$
3$23$# -# 51(.42 2/$"32 .% ".+.41 !$'5(.41 (-5$23(&3$#. C1!.-3(.- (2 -.3 6(#$+8 42$# (- 3'$ "-$ 24&1
(-#42318, !43 /$1(.#(" 2/(*$2 (- 24+%41 /1("$2, 24&1 04+(38 (224$2 -# $-5(1.-,$-3+ ".-"$1-2 '5$ 23(,4+3$#
$%%.132 3. ".-2(#$1 1$/+"(-& .1 24//+$,$-3(-& 24+%41 #(.7(#$ 6(3' "1!.- #(.7(#$ 3'3 ,8 !$ 5(+!+$ ".23-%1$$
%1., 3'$ %$1,$-33(.- /+-3. T'$ ".+.41 1$,.5+, viz. 3'$ 1$+3(5$ #(%%$1$-"$ !$36$$- ".+.41 .% 16 -# "+1(%($#
)4("$, .!3(-$# (- .41 3$232 62 .- 5$1&$ 35, 47, 44 -# 74% %.1 #$%$"3(.-, 24+%(33(.-, 2(-&+$-"1!.-3(.- -#
,.#(%($# #.4!+$-"1!.-3(.-, 1$2/$"3(5$+8. S$5$1+ %"3.12 %%$"3(-& "+1(%($# )4("$ ".+.41 (- '.3 +(,(-& 6$1$ 3$23$#,
viz. 3'$ 3(,$ -# 3$,/$1341$ #41(-& 2$33+(-&; !&"(++. -# 2.(+ ".-3$-3, -# /'.2/'3$ -# /1.3$(- ##(3(.-. A3 +.6
+(,$ #.2$, !$+.6 !.43 1 *& CO/3.--$ "-$ (#$%$"3(.-, 24+%(33(.- -# "1!.-3(.-), 2(&-(%("-3 /.13(.- .% ".+.41
1$,.5+ 1$24+32 %1., #2.1/3(.- .- 3'$ '$3-".&4+3$# "-$ /1.3$(-, (- ##(3(.- 3. (32 "/341$ !8 3'$ -2"$-3
"+"(4, /'.2/'3$ /1$"(/(33$. H.6$5$1, 3'$ #2.1/3(5$ "/"(38 .% 3'$ /1$"(/(33$ %.1 "-$ ".+.1-32 //$12 .-+8
/13(++8 $7'423$# (- 3'$ -.1,+ /1."$#41$. A+3'.4&' 3'$ #$".+.41(23(.- $%%$"32 .% 24+%(33(.- -# "1!.-3(.-
6$1$ %.4-# 3. !$ !.43 $04+, 3'$ //1$-3+8 +.6$1 3'$1,+ 23!(+(38 .% "+1(%($# )4("$ -# 2814/ /1.#4"$# !8
"1!.-3(.- ,8 1$04(1$ %413'$1 234#8. L.6$1(-& 3'$ "+1(%($1 3$,/$1341$ !8 11:C 62 %.4-# 3. +(,(3 3'$ )4("$ ".+.41
(-"1$2$ (- 3'$ "+1(%($12 3. -$1+8 9$1.. T'(2 62 3$23$# (- %"3.18 31(+. T'$ (-3$1-+ "+1(%($1 3$,/$1341$ 62
1$#4"$# !8 1$-1.43(-& %(+3$1 )4("$ #(1$"3+8 3. 3'$ (-+$3 .% 3'$ "+1(%($1. S+(&'3 1$#4"3(.- .% "+1(%($# )4("$ ".+.41 62
.!2$15$#, 6(3' -. -$&3(5$ $%%$"3 .- "+1(%($# )4("$ 341!(#(38.
K"64,/!0: "1!.-3(.-, "+1(%("3(.-, ".+.41, 24&1"-$, 24+%(33(.-
R$,.5+ .% ".+.41 (- 24&1 "-$ )4("$ "+1(%("3(.- !8
#$%$"3(.-, 24+%(33(.- -# "1!.-3(.-
By M. S2*
1*
, B.S. Z.22(
2
-# H. L(4
3
1
Audubon Sugar lnstitute, Louisiana State University Agricultural Center, St. Gabriel, Louisiana, USA.
2
Estacion Experimental Agroindustrial Obispo Colombres (EEAOCj, Tucuman, Argentina.
3
Guangxi University, Nanning, People's Republic of China.
*
Contact author: msaska@agcenter.lsu.edu
I-31.#4"3(.-
The colour of sugar, be it raw, direct mill white or refined is its most
important commercial attribute, and much resource is spent by the
millers and refiners to comply with the market requirements on the
colour of their product. Crystallisation itself, apart from producing a
stable, marketable product is also 95-99% effective in partitioning
colour and is, in the production of low colour sugar, supplemented
by a number of carbon and ion-exchange resin-based adsorption
processes and, to a lesser degree, by methods based on chemical
reactions that render colourless the colorant molecules.
The main objectives of juice clarification are to raise its pH and
eliminate suspended solids. Colour removal is at best considered a
secondary objective, rarely monitored by the mill laboratory and to
our knowledge never used as a criterion to assess, let alone control,
the process. The use of SO
2
is widespread in clarifying juice in
production of plantation white sugar, however periodic spikes in
sulfur prices and sugar quality issues have stimulated efforts to
reduce or even eliminate its use. With that in mind, a carbonation
process has been tested and compared with standard sulfitation
and defecation. Besides eliminating the use of sulfur, carbonation
would also provide a means to utilise and sequester some of the
excess CO
2
that may be available cost-free in some sugar factories
from molasses or juice fermentation. The traditional double-
carbonation process was used initially in cane juice clarification in
Java (Honig, 1959j, later practised for many years in South Africa
(Rault, 1960j, and is still used nowadays in some cane factories in
China, Taiwan (Sheen et al., 2003j and elsewhere. Although it is
reported to provide excellent, low colour clarified juice, the very high
lime consumption of 11-15 kg CaO / tonne cane is making the
economics increasingly un-sustainable. ln this program, some mod-
ifications were tested to reduce lime consumption and replace the
filtration of all first carbonation juice by settling and filtration of the
concentrated mud.
M3$1(+2 -# ,$3'.#2
ln most cases, raw juice was prepared by milling samples of fresh
cane brought to Audubon Sugar lnstitute, Louisiana, USA. Usually,
0.6 to 0.8 L of the mill juice was diluted to 1 L volume to bring the
brix closer to that of factory mixed juice, and clarified by four
methods: defecation by hot liming, sulfitation, single carbonation or
double carbonation.
For defecation, the diluted juice was brought to and kept boiling
for about one minute in a microwave oven, then quickly limed while
stirring with Ca(OHj
2
slurry to a pH between 7 and 8.2. Then 2.5 ppm
of Magnafloc LT340 flocculant were added while stirring, the whole
volume of limed juice transferred to a covered 1 L glass beaker and
allowed to settle in a 96C water bath, usually for 60 minutes, before
sampling the supernatant for analysis.
Sulfitation was done in a 6 L stirred jacketed glass reactor
provided with lime slurry and gas inlets, and pH and temperature
readouts; usually by first liming 1 L of diluted juice at about 50C to
a pH of 8-9 and then gassing with SO
2
to a pH of about 7.
Alternatively, gassing was done first to pH 3-4 followed by liming to
pH 7. The clarification performance was about equal, but the former
was preferred as it was considered to be more closely comparable
with the carbonation tests where gassing with CO
2
must be done at
alkaline pH (Figure 1j because of the negligible rate of absorption of
carbon dioxide below pH 6. ln either case, the sulfited and limed
juice was heated and kept boiling for one minute, flocculant added
and settling done as in defecation.
Carbonation was done by liming the diluted juice in the 6 L stirred
reactor kept at 50 to 60C, to pH 8-9, kept at the high pH for 2 to 5
minutes and then gassed with CO
2
to pH 7-8, over about 5 minutes.
The heating, flocculating and settling followed as before. The
quantity of lime was maintained about the same in both sulfitation
and carbonation to allow direct comparison between the two
methods, and at levels comparable with those used in plantation mill
white factories throughout South / Central America. ln the following
text, the pH measured at room temperature (about 25Cj is denoted
as pH25.
Double carbonation was done by preheating 1 L of diluted raw
juice in the 6 L reactor to 50C and liming with 3-5 g CaO to a
pH 10.5-11. The pH was then reduced with CO
2
to about 10,
temperature raised to 60C, flocculent applied, and the mud settled
for about 60 minutes to 25-30% of the initial volume. No precautions
were taken to remove dissolved air from the juice. The supernatant
from settling and filtrate from vacuum filtration of the thickened mud
were combined and pH of the combined clear juice then reduced
with CO
2
to pH 6.5-7. The carbonated juice was brought to boil and
the small amount of the second carbonation precipitate removed by
filtration under vacuum. The filterability of the first carbonation mud
was measured with the same apparatus and mostly following the
same procedure as before (Saska, 2005j in measuring filterability of
clarifier mud. A temperature of 60C was chosen, at 0.7, 1.4 and 2.1
x 105 Pa pressure, a filtration area of 3.1 cm
2
with the support
formed by a 20 m stainless steel mesh pre-coated with HyfloSuper
Cel filter aid.
lt is sometimes recommended that lime suspension be aged"
before its use in juice clarification but in these tests the analytical
grade Ca(OHj
2
that was used was mixed with water at about 1:10
ratio and used immediately. No adverse effects on settling or turbid-
ity removal were noted. All tests reported in this paper were done
during the 2008 - 2009 season with juice extracted from the main
cane cultivars currently grown in Louisiana (i.e. cv. L 97-128, HoCP
96-540, LCP 85-384, etc.j. However, other tests at EEAOC in
Argentina (Zossi and Cardenas, 2009j have confirmed the validity of
the conclusions regarding the colour behaviour in clarification of
juice produced from the two main cane varieties grown in Tucuman.
R$24+32
Colour removal in clarification
The colour of raw juice ranged from 10000 to 20000 lU with the
variations reflecting effects of different cane cultivars, cane conditions
and the varied quantities of tops and green leaves crushed
with the clean billets. The relative decolourisation, defined as 100 x
(Colour of raw juice - Colour of clarified juicej / Colour of raw juice for
defecation by hot liming, averaged 32%; sulfitation and carbonation
were nearly equally efficient in terms of colour removal, with decolouri-
sation of 45 and 42%, respectively with lime (CaOj consumption of
about 0.7 and 0.9 g/L respectively versus the 0.5 g/L used in normal
defecation. Although the colours of clarified and in particular of mixed
juice are rarely measured in the factories, the limited available data
(Eggleston, 2000; Sahadeo et al, 2002j indicate that the factory
performance in terms of colour removal is substantially lower than the
laboratory test results reported in Table 1. Amongst the reasons for the
difference between factory performance and laboratory decolourisa-
tion may be localised overheating or overliming in the industrial
process, excessive residence times of juice or mud in the industrial
F(&41$ 1. Ab0,/-1&,+ /a1" ,# a/b,+ !&,5&!" a+!
02)-%2/ !&,5&!" &+ )&*"! '2&"
heaters, clarifiers or filters, or other factors.
No systematic measurements were done of the mud settling
characteristics. However, all three procedures produced well-settling
mud with no apparent differences in settling rates among the three
methods. Clarified juice turbidity varied mostly within the 50-150 NTU
range, similar to the range of industrial clarified juice and, as with mud
settling rates, no systematic differences among the three clarification
methods were observed. The large standard deviation of the CaO
dose comes from the intentional variations introduced in the
procedure to test the robustness and response of the process.
Changes of juice pH and colour during settling
The composition of clarified juice and consequently the
clarification method may have an impact on the rate of colour
increase later in settling, during evaporation and in the
vacuum pans. Thus thermal stability of the juice needs to be
considered when evaluating clarification. To that effect, the
residence time was varied from to 4 hours for the three
types of clarified juices (Table 2j, at 96C i.e. close to the
average temperature in the industrial clarifiers, and at 85C for
clarified juice from hot liming.
At 95C, the colour increase in defecation and carbonation
juices was found to be about equal, approximately 600 lU/h,
but lower for juice produced by sulfitation. The pH drop during
settling of defecation juices was found significantly higher than
in sulfitation and carbonation.
Reducing the temperature from 96C to 85C in settling of
defecation juice had a dramatic effect on colour formation rate
and pH drop. Colour increase was reduced six-fold to less
than 100 lU/h and the pH drop about two-fold. Although not
studied here, it is probable that comparable
improvements would ensue in sulfitation and
carbonation. lt was suggested (Zossi et al.,
2009j that reduced clarifier temperature
could be accomplished in the factory without
any additional heat exchangers by re-routing
the vacuum filtrate that is now usually sent
back to the mixed juice tank. The limed juice
would be heated and flashed as usual but its
temperature would then be reduced down-
stream of the flash tank by the cooler filtrate.
This would be expected to reduce juice
colour and sucrose inversion, similar to the
benefits of short residence time clarifiers.
Thermal stability of syrup
The rates of Maillard and other mechanisms
responsible for colour formation are known
to increase at lower water concentrations.
Therefore, a series of thermal storage" tests
at 70C were also done with syrup following
the same methodology as in the recent raw
sugar storage tests (Saska and Kochergin,
2009j. Syrup from each clarified juice
was prepared in a glass laboratory rotovap"
evaporator, under vacuum at 50-55C.
Colour increased in storage in each case (Figure 2j; with the
approximate slopes of 15, 9, 29 and 13 lU/h for defecation,
sulfitation, carbonation and double carbonation, respectively.
Because of the lower temperature, the rate of colour formation is
much less than in settling (Table 2j, but the sulfitation syrup again
exhibits the slowest rate of colour increase. The pH drop (Figure 3j
is about equal in all four cases, and apparently independent of the
initial pH.
Mechanism of colour removal
lt is usually assumed that, in juice clarification, any colour removal is
due to adsorption of colorants on the nascent crystals of calcium
C(%(-*, IU CaO, /L Dec(%(-*"+a,"('
*
A. SD A. SD %
Ra/ $-#ce (N = 26) 14367 4113
Deeca,#(' (N = 18) 9766 3668 0.5 0.3 35
a
S-%#,a,#(' (N = 20) 7858 2821 0.7 0.5 47
b
Ca*b('a,#(' (N = 17) 8351 3127 0.9 0.7 44
b
*
d#e*e', +-)e*+c*#), %e,,e*+ (a (* b) #'d#ca,e a +#!'##ca', d#e*e'ce () 1 0.05) be,/ee' ,/( da,a +e,+;
/"#%e +a&e %e,,e*+ #'d#ca,e a +,a,#+,#ca%%0 #'+#!'##ca', d#e*e'ce
T!+$ 1. A3"/a$" ,),2/ b"#,/" a+! a#1"/ '2&" )a/&#&a1&,+, 1%" )&*" !,0"
20"! a+! /")a1&3" !",),2/&0a1&,+ a%&"3"!
C(%(-* c!a'e, IU/! )H25 c!a'e, 1/!
Te&)e*a,-*e,1C C%a*"f"ca,"(', b0 A.
*
SD A.
*
SD
96 Deeca,#(' (N = 19) 616
a
209 -0.34
a
0.09
96 S-%#,a,#(' (N = 6) 355
b
56 -0.07
b
0.03
96 Ca*b('a,#(' (N = 5) 581
ab
190 -0.08
b
0.07
85 Deeca,#(' (N = 9) 83 131 -0.17 0.08
* d#e*e', +-)e*+c*#), %e,,e*+ (a (* b) #'d#ca,e a +#!'##ca', d#e*e'ce () 1 0.05) be,/ee' ,/( da,a +e,+;
/"#%e +a&e %e,,e*+ #'d#ca,e a +,a,#+,#ca%%0 #'+#!'##ca', d#e*e'ce
T!+$ 2. C,),2/ &+/"a0" a+! -H !/,- !2/&+$ 0"11)&+$ a1 967C a+! 857C, #,/
!&##"/"+1 )a/&#&a1&,+ ,+!&1&,+0
F(&41$ 2. C,),2/ &+/"a0" ,# 60 - 70 b/&5 06/2- -/,!2"! #/,*
1%" #,2/ !&##"/"+1 )a/&#&a1&,+ -/,"!2/"0 a+! 01,/"! a1 707C
phosphates and other sparingly soluble anions, in analogy
to the more frequently studied colorant behaviour in sugar
refining by phosphatation or carbonation. However, the few
experiments that are summarised in Figures 4 and 5 indicate
that adsorption on heat-coagulated cane protein may be a sig-
nificant mechanism in colour removal during juice clarification.
ln the experiments in Figure 4, bringing the raw juice to boiling
without any lime addition removed nearly 7000 lU or 39% of
the initial colour. The full hot-liming and clarification done on
the same juice only added another 6% to the total 45%
decolourisation. When 100 mg/L aliquots of phosphoric acid
were added to the mixed juice prior to liming, the colour
removal by hot-liming increased by about 600 lU per aliquot.
ln the experiment in Figure 5, intrinsic cane protein was
supplemented by sequential additions of bovine serum albumin
(BSAj, a structurally similar protein. An addition to diluted raw
juice of 1.5 g/L of protein increased the colour removal by hot
liming by an additional 1500 lU. Another 3 g/L of BSA was
added (for a total of 4.5 g/Lj to the clear juice from this exper-
iment, and the spiked juice was briefly boiled again; however,
no additional lime was added. The colour decreased by another
1300 lU. This was repeated one more time, for a total of 7.5 g/L
of protein, with an additional 600 lU removed. lt is therefore clear
that both BSA and cane protein when heat-coagulated or dur-
ing heat-coagulation have strong affinity for cane colorants.
The decolourisation effect, however, decreases with increas-
ing dose of the protein; perhaps because the affinity of the
protein is specific for only certain fractions among the wide
variety of cane juice colorants. Unlike BSA, egg-white albumin
was found ineffective in juice colour removal.
Bagacillo or soil particles in only coarsely screened indus-
trial cane juice have a substantial effect on the mud behaviour
in clarifiers, most notably the former providing the bulk of
the mud volume. Whether they affect colour of clarified juice
as has been sometimes alleged is less certain. The experi-
ments summarised in Figure 6 indicate that neither has any
detectable influence above the normal experimental
variations. The added amounts indicated in Figure 6 are given
in g dry matter/100 mL raw juice.
The clarifier underflow is made up of 90-95% clarified juice
with colour that one could assume identical to that of the clar-
ifier overflow provided the residence times were comparable.
However, our previous observations in factory tests indicated
that the colour of the mud juice", that is the juice entrained in,
and recovered from the mud in vacuum filters, was on occasions
actually lower than that of the overflow from the clarifiers. This
prompted tests to determine the state of saturation of the
adsorptive capacity of mud particles for colorants. ln the
experiments reported in Table 4, cane juice was clarified by
hot liming and its colour determined as usual (CJ colour and
Decolourisation 1 in Table 4j. The settled mud was then
blended for a few seconds in a standard kitchen blender,
solids separated by centrifugation, and the colour of the
supernatant juice again determined by the standard method.
ln all, twenty experiments were done and are reported in Table
4; the colour of the supernatant from the blended mud (Mud
colour and Decolourisation 2 of Table 4j always decreased,
sometimes by up to 2700 lU. This is evidence that the
F(&41$ 3. -H !/,- !2/&+$ 01,/a$" ,# 06/2- -/,!2"! b6 1%"
#,2/ !&##"/"+1 )a/&#&a1&,+ -/,"!2/"0
F(&41$ 4. C,),2/ ,# /a4 '2&", /a4 '2&" a#1"/ b,&)&+$ a+!
)a/&#&"! '2&" b6 %,1 )&*&+$ a#1"/ a!!&1&,+ ,# -%,0-%,/& a&!
F(&41$ 5. C,),2/ ,# /a4 '2&", )a/&#&"! '2&" b6 %,1 )&*&+$
(,+1/,)), a+! )a/&#&"! '2&" a#1"/ 0".2"+1&a) a!!&1&,+0 ,# b,3&+"
0"/2* a)b2*&+
adsorptive power of the hot-liming precipitate is not exhausted in
the standard process.
lt is possible that blending the mud exposes internal surfaces of
the precipitate, rendering them available to absorb more colorant. lt
would seem, therefore, that potential exists for further improvements
in decolourisation above the levels reported in Table 1.
Double-carbonation
ln the modified double-carbonation procedure, the CaO
dose was reduced on average to 3.1 g/L (Table 5j, or about
four-times greater than the present industrial process.
Decolourisation at these conditions was 74% on average, with
some values exceeding 80%. Clarified juice turbidity was
less than 10 NTU, and Ca and Mg ions determined by ion-
chromatography were about 500 and 200 mg/L, respectively.
The filterability of the thickened first carbonated mud may
be the critical parameter for the scale-up of the modified
process. Therefore, the effects of various parameters, e.g. the
quantity of lime used were measured. The filterability of the
juice, thickened by settling to about 30% of the original
volume, was found unaffected (Figure 7j, and comparable
with filterability of the standard hot liming defecation mud
determined under identical conditions (Saska, 2005j. The
volume to be handled by the filter station is greatly reduced,
and its capacity in terms of tonnes of cane per day increased.
Reducing colour and sucrose inversion by re-routing filter juice
The finding regarding the affect of temperature on clarified juice
colour can readily be exploited in the factory. Re-routing the cooler
filter juice directly to the inlet of the clarifiers (Figure 8j, rather than
F(&41$ 6. C,),2/ ,# /a4 '2&" (13 200 IU), '2&" )a/&#&"! b6 %,1
)&*&+$ (,+1/,)), a+! '2&" )a/&#&"! b6 %,1 )&*&+$ a#1"/ a!!&1&,+0
,# ba$a&)), ,/ !/6 *2! 1, 1%" /a4 '2&"
RJ c(%(-* CJ c(%(-* M-d c(%(-* Dec(%(-*"+a,"(' 1
*
Dec(%(-*"+a,"(' 2
*
D"ffe*e'ce
IU IU IU % % %
15 957 9986 7764 37 51 14
10429 7742 35 51 17
9811 7575 39 53 14
10000 7701 37 52 14
17 390 8689 7473 50 57 7
8838 7331 49 58 9
8919 7519 49 57 8
17 390 8013 6132 54 65 11
7993 6176 54 64 10
8189 6287 53 64 11
8356 5936 52 66 14
12 712 6772 6467 47 49 2
7010 6691 45 47 3
7241 6492 43 49 6
7165 6655 44 48 4
A.! 46
a
A.! 55
b
*d#e*e', +-)e*+c*#), %e,,e*+ (a (* b) #'d#ca,e a +#!'##ca', d#e*e'ce () 1 0.05) be,/ee' ,/( da,a +e,+; /"#%e +a&e %e,,e*+ #'d#ca,e a +,a,#+,#ca%%0
#'+#!'##ca', d#e*e'ce
T!+$ 4. C,),2/ ,# /a4 '2&", )a/&#&"! '2&" a#1"/ %,1 )&*&+$ a+! ,# 1%" 8*2! '2&"9 a#1"/ b)"+!&+$ 1%" *2!
C(%(-*, IU Dec(%(-*"+a,"('
Ca0 )H
25
RJ CJ % IU
added /L (f CJ
A.! (N = 43) 3.1 7.4 12302 2877 74 9432
SD 0.9 1.2 3917 660 12 3489
T!+$ 5. A3"/a$" -"/#,/*a+" ,# 1%" !,2b)" a/b,+a1&,+ )a/&#&a1&,+ ,# a+" '2&"
sending it to the mixed juice tank as usual would reduce the internal
temperature within the clarifier. This would eliminate the cost of re-
circulating the filter juice through the juice heaters and shift heating
duty to the clarified juice heater. ln addition to lower colour, inversion
of sucrose would be reduced. By reference to standard tables, a 10
C temperature reduction lowers sucrose inversion by 0.3 - 0.6 kg /
tonne cane. ln Table 6 the approximate expected filter and clarified
juice temperatures were calculated if all filter juice were sent
directly to the clarifier and thermal losses were negligible. The effect
depends on the temperature and volume of the filter cake wash
water but it is clear that temperatures are still above the range
conducive to microbial activity that of course must be avoided.
This arrangement was tested in 2009 at Alma sugar factory in
Louisiana. A line was installed (Figure 8j allowing re-cycling of the
filter juice from vacuum filter F1 directly to the inlet of one of the three
installed clarifiers. The old line was left in place so that either of
the two arrangements could be operated and switched by merely
opening or closing the valves. The new arrangement was operated
for about three weeks in October - November without noticing any
adverse affects on the process. Reducing the volume that had
to pass through the mixed juice tank was welcome because of
unrelated problems with the occasional mixed juice tank overflow.
The laboratory turbidity readings indicated that turbidity of clarified
juice from the Dorr was unaffected.
More detailed monitoring of the opera-
tion was done on November 3. For about
three hours filter juice F1 was continued to
be recycled directly in the Dorr and the
temperature of the overflow and under-
flow were recorded (Figure 9j. Samples of
the overflow were taken periodically to be
analyzed for color and turbidity. At around
11:30 the arrangement was switched but
the temperature recording and sampling
continued. At the same time, as a control,
sampling and temperature recording
was also done on Graver 1 clarifier which
should be unaffected by the changes on
the Dorr clarifier.
As expected, the temperature of the
juice and mud from the Dorr was lower by about 3C (Table 7j on
average in the new arrangement while the temperatures in Graver 1
were nearly unchanged. All temperatures were below normal
because of persistent problems with the factory juice heaters. ln
consequence, the juice colour changes were smaller than predicted.
lf we attribute the change in Graver clarifier (controlj juice colour
to changing cane quality in the course of the test and add it to the
difference measured in the Dorr, the net colour reduction in the Dorr
in the new filter juice arrangement was about 400 lU or about 3% of
the total colour. Turbidity was reduced on average from 95 NTU to
92 NTU; this is within the experimental errors but the important
finding was that no negative effect on juice turbidity was noticeable
throughout the three-week operation of the new filter juice
arrangement.
F(&41$ 7. F&)1"/ab&)&16 ,# 1%&("+"! *2! #/,* 1%" *,!&#&"!
-/,"00 (0,)&! )&+"0) 4&1% 3 a+! 5 $/L ,# CaO, a+! 1%a1
,# 1%" !,2b)" a/b,+a1&,+ -/,"00 4&1% 12 $/L CaO
(&+!201/&a) ,+!&1&,+0, !,11"! )&+")
F(&41$ 8. M,!&#&"! #&)1"/ '2&" a//a+$"*"+1 a1 A)*a
02$a/ #a1,/6
F"%,e* #-"ce ,e&)e*a,-*e C%a*"f"ed #-"ce ,e&)e*a,-*e
Wa+! /a,e* Wa+! /a,e* / ca$e Wa+! /a,e* / ca$e
,e&)e*a,-*e
1C 1 1.5 2 1 1.5 2
30 72.7 64.9 59.2 98.8 96.8 94.9
40 76.8 70.0 65.2 99.2 97.5 95.9
50 80.9 75.2 71.1 99.7 98.3 96.8
60 84.9 80.4 77.0 100.1 99.0 97.8
70 89.0 85.5 83.0 100.6 99.7 98.8
80 93.0 90.7 88.9 101.0 100.4 99.8
T!+$ 6. E5-"1"! )a/&#&"! '2&" 1"*-"/a12/" (7C) &# a)) #&)1"/ '2&" 4"/" /"-/,21"!
!&/"1)6 1, )a/&#&"/ &+)"1. F)a0%"! '2&" 1"*-"/a12/" 1027C
C.-"+42(.-2
The average decolourisation that was achieved in juice clarification
by defecation, sulfitation, single carbonation and modified double
carbonation was 35, 47, 44 and 74%, respectively. However,
available data indicate that decolourisation that is routinely achieved
in defecation (Eggleston, 2000; Sahadeo et al., 2002j and sulfitation
factories (Zossi and Cardenas, 2008j is considerably less.
Replacement of SO
2
by CO
2
(carbonationj is feasible, achieving
comparable decolourisation and mud settling characteristics, but
the apparently lower thermal stability of clarified juice and syrup from
carbonation requires more study.
The colour increase at conditions typical in industrial clarifiers
was found to be about 600 lU/h for defecation and carbonation, and
about 350 lU/h for sulfitation clarified juice. By reducing the settling
temperature by 11C, the colour increase could be reduced to
less than 100 lU/h.
An indication was obtained that, in juice clarification, most
decolourisation comes from adsorption of cane colorants onto
heat-coagulated cane protein rather than onto the nascent calcium
precipitate, and that some of the adsorption capacity
remains unused in the standard process.
No significant effect was observed from either bagacillo
or soil on colour removal in clarification.
Significant reductions of CaO consumption in the
commercial double carbonation process to about 3 kg
CaO/tonne cane are possible, in conjunction with thickening
of the first carbonation mud by settling and polishing
filtration after the second carbonation.
The internal clarifier temperature can be reduced at no
cost by re-routing filter juice directly back in the inlet of
a clarifier. This is expected to produce similar effects as
shorter residence time: a clarified juice with less colour
and smaller sucrose inversion loss. No negative effect on
clarified juice turbidity was observed in a factory test.
R$%$1$-"$2
Eggleston, G. (2000j Hot and cold lime clarification in raw
sugar manufacture. l: Juice quality differences. Int. Sugar J., 102
(122j: 406-416.
Honig, P. (1959j Principles of sugar technology. Elsevier,
Amsterdam.
Rault, J. (1960j The juice carbonatation process and
repercussions of economics on technology. Proc. S. Afr. Sug. Technol.
Assoc., 34: 120-127.
Sahadeo, P., Lionnet, G.R.E. and Davis, S.B. (2002j Mixed juice
clarification revisited. Proc. S. Afr. Sug. Technol. Assoc., 76: 421-432.
Saska, M. (2005j Composition of clarifier mud and its filterability.
Sugar Journal, 67 (10j: 10-15.
Saska, M. and Kochergin, v. (2009j Quality changes during storage of
raw and vLC sugar: Effects of pH and moisture. Int. Sugar J., 111 (1324j:
234-238.
Sheen, H.K., Huang, C.M., Chang, R.Y., Chen, W.C., Lin, L.H., Hsiung,
S.Y. and Liang, J.H. (2003j Modification of carbonation process in sugar
mill for the production of edible B-grade white crystal sugar. Taiwan Sugar,
50: 8-14. http://www.cabdirect.org:80/abstracts/20033169851.html
Zossi, S., Liu, H. and Saska, M. (2009j. Colour and pH Phenomena in
Cane Juice Clarification by Defecation, Sulfitation and Carbonation, Proc.
Sug. Ind. Technol., lnc., New Orleans.
Zossi, S. and Cardenas, G. (2009j Estacion Experimental
Agroindustrial Obispo Colombres (EEAOCj, Tucuman, Argentina,
unpublished.
F(&41$ 9. T"*-"/a12/"0 (7C) ,# )a/&#&"/ ,3"/#),4 (b)2") a+! 2+!"/#),4 (/"!) #/,* 1%" D,// )a/&#&"/
Te&)e*a,-*e, 1C
F"%,*a,e ,( D(** F"%,*a,e ,( MJ ,a'$
D(** CJ 91.7 94.5
c%a*##e* M-d 88.8 90.9
C(',*(% CJ 94.9 94.1
(G*a.e* c%a*##e*) M-d 90.7 92.4
T!+$ 7. A3"/a$" )a/&#&"! '2&" a+! *2! 1"*-"/a12/"0 &+
1%" #a1,/6 )a/&#&"/0
C(%(-*, IU T-*b"d",0, NTU
F"%,*a,e F"%,*a,e ,( F"%,*a,e ,( F"%,*a,e ,(
,( D(** MJ ,a'$ ,( D(** MJ ,a'$
D(** CJ 16534 16748 91.7 94.5
M-d 15007 15601 - -
C(',*(% CJ 16946 16758 94.9 94.1
(G*a.e* 1) M-d 15337 15556 - -
T!+$ 8. A3"/a$" )a/&#&"! '2&" ,),2/ a+! 12/b&!&16 &+ 1%"
#a1,/6 )a/&#&"/0