Journal of Food Engineering 89 (2008) 429–434

Contents lists available at ScienceDirect

Journal of Food Engineering

journal homepage: www.elsevier .com/locate / j foodeng

Evolution of colorants in sugarbeet juices during decolorization using styrenic resins

Mónica Coca *, M. Teresa García, Silvia Mato, Ángel Cartón, Gerardo GonzálezDepartamento de Ingeniería Química y Tecnología del Medio Ambiente, Universidad de Valladolid P� Prado de la Magdalena s/n, 47011 Valladolid, Spain

a r t i c l e i n f o

Article history:Received 26 November 2007Received in revised form 13 March 2008Accepted 17 May 2008Available online 2 July 2008

Keywords:MelanoidinsSugar technologySize exclusion chromatographyColor removalIon exchange

0260-8774/$ - see front matter � 2008 Elsevier Ltd. Adoi:10.1016/j.jfoodeng.2008.05.025

* Corresponding author. Tel.: +34 983 423 166; faxE-mail address: monica@iq.uva.es (M. Coca).

a b s t r a c t

Molecular size distribution of coloring impurities in sugar beet juices was studied in order to get a betterunderstanding of the evolution of colorants during ion exchange decolorization processes using styrenicresins as well as to provide useful information about the influence of operating decolorization conditionsand regenerant consumptions on the removal of harmful colorants. A study of resin life was also per-formed. Size exclusion chromatography (SEC) of sugar beet thin juices confirmed the presence of colo-rants with molecular masses above 100 kDa, 20 kDa and 2 kDa. The global color reduction percentagesachieved in the decolorization stage were about 75–80%. The colorants with a molecular mass of20 kDa were completely removed whereas components above 100 kDa and 2 kDa presented lowerremoval efficiencies, showing lower affinity for the styrenic resin. Colored impurities are likely to berelated to melanoidins, Maillard reaction products. Low regenerant consumptions, about 57 L of solutionper m3 of treated juice, removed adsorbed colorants from the styrenic resin without reducing its decol-orization capacity considerably.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

Many factors affect the quality and quantity of beet sugar. Thesefactors are related to the formation of non-sugars, mainly coloredcompounds. These components are formed through beet process-ing as a result of pH changes, thermal and autocatalytic effects.These impurities are of high molecular masses, polymeric and withtendency to occlude within the sugar crystal (Bento and Sá, 1998).The nature of beet coloring matter is quite different to cane colo-rants (Godshall et al., 2002). The mechanisms concerning color for-mation in beet sugar processing are very complicated because ofthe many parameters involved. The main mechanisms related tocolor formation during purification stage are Maillard reactionand alkaline degradation of invert sugars. Maillard reaction prod-ucts, melanoidins, are formed by the reaction of monosaccharidesand carbonyl compounds with amino acids. Melanoidins are recog-nized as being acidic and polymeric compounds, with a highlycomplicated structure (Cämmerer and Kroh, 1995; Cämmerer etal., 2002; Yaylayan and Kaminsky, 1998). The formation of mela-noidins proceeds much faster at high temperature and basic pHconditions. In concentrated juices, the Maillard reaction also pro-ceeds faster. Alkaline degradation products of hexoses (HADP)may be responsible for up to 80% of color in sugar beet juices(Heitz, 1995). The production of colored HADP takes place at thecommon pH of a beet sugar factory (8–11). The formation of deg-radation products takes place mainly in the purification step where

ll rights reserved.

: +34 983 423 616.

temperature increases up to 85 �C and pH rises up to strong basicvalues (11–12). The nature and structure of colored HADP havenot been elucidated but they are probably due to the extensivealdolisation of intermediate di-carbonyl compounds in alkalinesolutions (De Bruijn, 1986).

To meet standards of whiteness, it is necessary to undertake ef-forts to reduce color levels in the end product. In addition to pre-ventive industrial actions, a decolorization stage becomesnecessary. Most of resin applications for decolorization have beenmade in cane sugar industry (Bento, 1992, 1997; Godshall, 1999).Ion exchange technology has also considerable potential to decol-orize beet sugar juices, removing colorants by adsorption as wellas by ion exchange (Broughton et al., 1991; García Agudo et al.,2002). At usually process pH conditions, most of these colorantsexhibit an anionic nature thereby anion styrenic resins are efficientdecolorizing materials (Bento and Sá, 1998; Guimaraes et al., 1996;Gula and Paillat, 2005). Styrenic matrices increase the affinity ofcolorants, showing higher decolorization efficiencies than acrylicsbut the regeneration stage is more complicated (Bento, 1997).The extent of decolorization depends on colorant properties:molecular size, ionic charge and hydrophobicity. However, colo-rants in sugar solutions are usually managed as a single substance.The global color measurement by spectrophotometric analysis at420 nm wavelength does not make possible to distinguish thedifferent characters of colorants. Therefore, it is necessary toadvance in the knowledge of beet colorants through the decolor-ization processes for a sound factory scale design.

This paper is concerned with the removal of colored compo-nents from beet sugar solutions using a commercial polystyrenic

430 M. Coca et al. / Journal of Food Engineering 89 (2008) 429–434

resin. Analysis by size exclusion chromatography coupled withdiode array detection was chosen to explore the characteristics ofcolorants through the decolorization process, providing a betterunderstanding about the nature and molecular size distributionof these impurities in juices and regeneration effluents. The affinityof colorants in beet sugar streams towards a strong styrenic resinwas analyzed together with the influence of the regenerant con-sumptions on resin life. This information is of great importancefor the design of reliable decolorization processes.

2. Materials and methods

2.1. Adsorbent

Based on previous laboratory studies, a commercial strong an-ion styrenic resin Lewatit 6368 Sulphate (Bayer) was selected asadsorbent. Prior to being used, the resin was regenerated with so-dium chloride and washed with distilled water.

2.2. Adsorbate

Sugar beet solutions were taken from a Northern Sugar factoryin the campaign 2006/2007. Thin juice (15–18% dry matter) was

Table 1Characteristics of thin juice

Color (IU) Dry matter (%) pH Purity (%) Chloride (mg/L) Sulfat

2910±100 16.7±1.0 9.1±0.4 92.9±2.0 425±15 617±

FEED JUICE

Fig. 1. Set up

prepared by dilution of thick juice containing 70% dry matter. Table1 summarizes the average characteristics of thin juice solutions.

2.3. Experimental set up

Experiments were carried out in duplicate in jacketed columnsoperating in parallel. Each column held 100 mL of resin (1 Bed Vol-ume). The experimental setup is shown in Fig. 1. First, a series ofdecolorization experiments feeding thin juice were performed at70 �C in down flow direction varying the flow rate between 7and 15 bed volumes per hour (BV/h). The cycle length wasprolonged up to reduce about 80% of color. Decolorization wasfollowed measuring absorbance at a wavelength of 420 nm,according to ICUMSA methods (ICUMSA, 1994). After decoloriza-tion, resins were washed with 13 BV/h of distilled water in up flowmode until the outlet stream had almost no sugar in solution (0% to1% dry matter). Resin regeneration was achieved with a solution ofsodium chloride (20%) and sodium hydroxide (0.7%). Regenerantwas fed through the resin in down flow at 60 �C. Regenerant vol-ume was varied between 2 and 3.5 BV. Finally, the resin waswashed with distilled water at 13 BV/h in up flow mode until a va-lue of dry matter content in the effluent lower than 5%. Washingswere performed at room temperature. After washing, the regener-

e (mg/L) Nitrate (mg/L) Citrate (mg/L) Lactate (mg/L) PCA (mg/L)

25 71±4 60±3 2180±100 1130±50

DECOLORIZED JUICE

diagram.

M. Coca et al. / Journal of Food Engineering 89 (2008) 429–434 431

ated resin was ready for further experiments. All the experimentswere carried out in duplicate and mean values are presented. Max-imum deviation was ±1%.

2.4. Analytical procedures

Color, dry matter content, pH and purity were determinedaccording to ICUMSA methods (ICUMSA, 1994). Organic and inor-ganic anions were determined by HPLC. Organic anions were deter-mined by using an Aminex� HPX-87H column from Biorad coupledto a diode array detector (PDA 996, Waters). Organics acids weredetected at 220 nm. 0.01 N sulfuric acid solution was used as themobile phase at a flow rate of 0.6 mL/min. Analyses were per-formed at 60 �C. Inorganic anions were analyzed with an IC-PakTM

Anion HC column from Waters with a conductivity detector. Col-umn temperature was set at 35 �C. A solution of borate/gluconatewas fed at 2 mL/min.

The molecular masses of colored compounds were estimated bysize exclusion chromatography. Separation was achieved in anUltrahydrogelTM 250 column (exclusion limit 1–80 kDa) packedwith hydroxilated polymetacrylated-based gel. A guard columnof the same material was also used. Deionized water (Milli Q sys-tem) was used as mobile phase at a flow rate of 0.7 mL/min. Col-umn temperature was 35 �C. Identification was achieved by usinga Waters 996 Photodiode Array Detector coupled in series with arefractive index detector. Pullulans (non-branched polymeric sug-ars) were used as standards for the estimation of molecularmasses. Data were processed with the software Millenium 32

0.000

0.005

0.010

0.015

0.020

0.025

minutes2.00 4.00 6.00 8.00 10.00

> 1

00 k

Da

20 k

Da

a.u

Thin juice Decolorized Juice

2

nm200.00 300.00 400.00

264.6

20 kDa

2 kDa

(b)

HADP

Fig. 2. Removal of colored impurities during decolorization of thin sugar beet juice usingSpectra of 20 kDa and 2 kDa colorants in thin juice (b) comparison with synthetic HADP

Chromatography Manager. Samples were filtered through a0.22 lm membrane before injecting onto the column.

3. Results and discussion

3.1. Decolorization stage

First, a series of experiments feeding the thin juice through theresin were carried out, analyzing the influence of flow rate ondecolorization efficiencies. Three flow rates were studied: 7, 11and 15 BV/h. The volumes of treated juice were 45.5, 44 and37.5 BV, respectively. The higher color removal percentage was78.5%, corresponding to the lower flow rate whereas the lower per-centage, corresponding to the higher flow rate, was 77.2%. In allexperiments, purity, pH and dry matter content remained virtuallyconstant. Regarding inorganic content, nitrate and sulfate ionswere completely removed due to ion exchange reactions. Citrateions were also completely removed. However, the removals oflactate and pyrrolidone carboxylic acid (PCA) were considerablylower, 10%. From these results, a flow rate of 11 BV/h may beconsidered suitable for decolorizing thin juice, with a percentageof color removal about 77.5% after 240 min operation. Color con-tent was reduced from 2900 to 650 IU. The lower flow rate testedled to a slightly higher color reduction percentage after treating45.5 BV of juice in 390 min, making difficult the subsequentscale-up and industrial operation.

The molecular size distribution of sugar beet solutions wasstudied by SEC. Fig. 2a compares the chromatograms of thin juice

12.00 14.00 16.00 18.00 20.00

2 k

Da

(a)

00.00 300.00 400.00

20 kDa

2 kDa

(c)

Melanoidin

nm

styrenic resins; (a) molecular size distributions of thin juice and decolorized juice.; (c) comparison with synthetic melanoidins. Chromatogram run at 420 nm.

a.u

0.000

0.010

0.020

0.030

0.040

0.050

minutes2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00

> 1

00 k

Da

2 kD

a

Decolorized juice

Syrup B

Syrup A

Fig. 3. Evolution of colorants in decolorized juice through an evaporation stepunder vacuum. Chromatogram run at 420 nm.

a.u

0.00

0.02

0.04

0.06

0.08

0.10

minutes2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00

>10

0 kD

a

20 k

Da

2 kD

a

Syrup B

Syrup A

Thick juice

Fig. 4. Chromatograms at 420 nm corresponding to the evaporated decolorizedjuices and thick beet juice.

432 M. Coca et al. / Journal of Food Engineering 89 (2008) 429–434

and decolorized juice at a flow rate of 11 BV/h. Both chromato-grams were obtained at 420 nm, the wavelength useful to measurecolor in beet sugar solutions. Regarding the molecular size distri-bution of thin juice, there were observed three peaks with differentmolecular masses, corresponding to the peak apices. The first peak,appearing at 7.5 min, corresponded to colorants with molecularmass above 100 kDa. The peaks with retention times of 10.2 and13.2 min corresponded to colored components with molecularmasses of 20 and 2 kDa, respectively. Spectra of colorants withmolecular masses of 20 kDa and 2 kDa were compared with thoseof melanoidins and HADP previously synthesized according toShore’s method (Shore et al., 1984). By comparing the spectra ofcolorants in sugarbeet juices with those of synthetic colorants,the characterization of colored impurities in beet juices may be ef-fected. Fig. 2b and c compare spectra of colorants in the thin juicewith those of synthetic colorants. Melanoidins are characterized bya stable absorbance region between 278.8 and 326.2 nm whereasHADP showed a strong absorption peak at 264.6 nm (Coca et al.,2004). Spectra of synthetic colorants are similar to those describedby other authors (Guimaraes et al., 1996; Rafik et al., 1997). As canbe observed in Fig. 2b and c, spectrum of 20 kDa peak presents astable absorption region at 268.7 nm. The 2 kDa molecular masscolorant shows a slight absorption maximum at 275.8 nm. Sizeexclusion chromatograms were processed with the Millenium 32Chromatography Manager software, which is able to match thespectra of the colorants in beet juice with those of synthetic colo-rants. However, the software did not find any match between col-orants in the sugarbeet thin juice and the synthetic colorants. Thisfact may be probably due to the complex interactions taking placebetween color precursors throughout the sugarbeet processingwhere there are almost limitless possibilities of polymeric colo-rants formation.

Regarding the decolorized juice, chromatograms shown in Fig.2a confirm that the concentration of coloring impurities responsi-ble for the juice color was considerably reduced. The results corre-sponding to 15 and 7 BV/h were quite similar. The colorants with amolecular mass of 20 kDa were completely removed, showing highaffinity for the styrenic resin. The removal of colorants with amolecular mass above 100 kDa was 75–80%. The colorants of lowermolecular mass presented considerably lower decolorization re-moval efficiencies: 50–70%. Therefore, the colorants with molecu-lar mass of 2 kDa exhibit lower affinity for the resin. Quantificationwas made by calculating the peak areas at 420 nm. According toreferences (Guimaraes et al., 1996) melanoidins and HADP, whichwere synthesized according to the Shore’s method (Shore et al.,1984), were well retained by styrenic resins, reporting removalpercentages about 97.5% and 98%, respectively. However, theexperimental results described above have shown that the removalof colorants depends on the molecular mass. This fact may be dueto the different conditions of temperature and reactants betweenthe Shore’s method and the sugar beet refinery. Decolorizationstudies on cane sugar affination liquor reported that melanoidinswith molecular masses between 10 and 40 kDa were completelyremoved by polystirenic resins (Bento and Sá, 1998). Therefore,the colorants with molecular mass of 20 kDa, with anionic charac-ter, may be related to melanoidins.

Evaporation is the beet processing step in which the formationof color due to the Maillard reaction is more important due to thehigh temperatures, up to 120 �C. To analyze the evolution of decol-orized juice through the evaporation stage, experimental runswere carried out in a rotary evaporator working under vacuum(temperature about 85 �C). The concentration ratio was about4:1. As a result of the evaporation step, color increased about40–60%, depending on heating time and temperature. The dry mat-ter content increased from 17% to 69–78%. Purity and pH remainedalmost constant. Purity and pH values were about 94–96% and

8.8–9.0, respectively. Regarding molecular size distribution of col-orants, Fig. 3 compares the SEC chromatograms of the decolorizedjuice and syrups obtained after different evaporation runs. Thesyrups designated as ‘‘A” and ‘‘B” presented dry matter contentsof 69% and 78%, respectively. The color contents were about1630 IU and 970 IU, respectively. As can be seen, there was a sharpincrease in the concentration of colorants, especially those withmolecular mass of 2 kDa. Coloring impurities with molecular massof 20 kDa, which were completely removed during the decoloriza-tion stage, were not formed. Evaporation conditions favour brown-ing, especially of components above 100 kDa and 2 kDa. Therefore,these compounds may also be related to melanoidins becausethese kinds of colorants are mainly formed during the evaporationstage (Mersad et al., 2001).

Colorant distributions in syrups were compared with which ofthick juice from the sugar factory. As is observed in Fig. 4, colorantsin thick juice were divided into two groups with different molecu-lar masses. The first group comprised compounds whose molecularmass was 20 kDa. The second group comprised compounds with amolecular mass about 2 kDa. There was also a large shoulder thatcorresponded to the elution of colorants with molecular massabove 100 kDa. Dry matter contents were similar, about 70–80%,but color of thick juice was 3300 IU, considerably higher than colorof syrups from the evaporation of decolorized juices, which wereabout 1000–1600 IU.

a.u

0.000

0.002

0.004

0.006

0.008

minutes2.00 4.00 6.00 8.0 10.00 12.00 14.00 16.00 18.00 20.00

CYCLE 1 CYCLE 10 CYCLE 15

>10

0 kD

a

2.5

kDa

Fig. 6. Resin life: chromatograms at 420 nm of decolorized juices from differentcycles. Percentages of color removal of cycles 1%, 10% and 15: 77%, 75% and 76%,respectively.

M. Coca et al. / Journal of Food Engineering 89 (2008) 429–434 433

3.2. Regeneration stage

After analyzing the decolorization stage, it was necessary toperform a study of regeneration of exhausted resin. A series ofexperimental runs were performed feeding 40 BV of thin juice at70 �C at a flow rate of 11 BV/h through the styrenic resin bed. Afterdecolorization, a solution of sodium chloride (20%) with sodiumhydroxide (0.7%) was used as regenerant. Sodium hydroxide wasadded to reinforce the action of sodium chloride. Both chemicalsare considerably cheaper than other reagents also suitable forregenerating styrenic resins after decolorizing cane juices, suchas ethanol, calcium salts and sucrose (Bento, 1996). Moreover, pre-vious studies with a mixture of calcium hydroxide in sucrose solu-tion led to the formation of precipitates during the regeneration ofstyrenic resins after decoloring beet juices (data not shown). In or-der to minimize the reagent consumptions, the regenerant volumewas varied between 2 and 3.5 BV. Regeneration experiments wereperformed at 60 �C.

Experimental results show that decolorization percentageswere about 75% and 78%, independently of the regenerant con-sumption in the further regeneration step. Fig. 5 shows the molec-ular size distribution of the regeneration effluents. As can be seenfrom the figure, the effluents presented colorants with molecularmasses above 100 kDa, 2.8 kDa and 0.5 kDa. In the raw thin juice,there were colorants with molecular masses above 100 kDa,20 kDa and 2 kDa. The lower molecular mass colorant in theregeneration effluent did not appear in the thin juice before decol-orization. The spectrum shape of the 0.5 kDa colorant in the regen-eration effluent showed a stable absorbance region at about268.7 nm and a strong absorption maximum at 214.5 nm. The20 kDa colorant in the raw thin juice also presented a stable absor-bance region at 268.7 nm. This might be interpreted as follows: thehigh salt composition of the regenerant may produce the depoli-merization of the adsorbed colorants. Moreover, the colored com-pounds may be difficult to be removed from the resin matrixdue to interactions during the regeneration stage at high saltconcentration.

Regarding the influence of regenerant consumption on the effi-ciency of regeneration, SEC chromatograms showed hardly differ-ences. Analyzing the area under the absorption curves in Fig. 5, itcan be seen that the concentration of colorants removed fromthe resin was 8–10% higher feeding 3 BV of regenerant solution,especially the concentration of colored impurities of lower molec-ular mass. As there were almost no differences, it can be selected avolume of 2.5 BV of regenerant to perform the analysis on resin life.This consumption corresponds to a ratio of 17.6 L of juice per literof regenerant solution. Salt consumptions were 12 kg NaCl and0.42 kg NaOH per m3 of juice treated, respectively.

a.u

0.00

0.10

0.20

0.30

0.40

0.50

minutes2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00

2 BV2.5 BV3 BV

>10

0 kD

a

2.8

kDa

0.5

kDa

Fig. 5. Influence of the regenerant consumption on the removal of adsorbedcolorants from the exhausted resin. Chromatograms at 420 nm.

3.3. Resin life

Resin life depends on quality of juice charged to resin, nature ofcolorants, color loaded per cycle, resin type and consumption ofregenerant (Bento, 1997). To analyze resin life, there were per-formed different experiments charging 40 BV of thin juice throughthe resin bed. Regeneration was carried out feeding 2.5 BV of theregenerant solution. Experimental conditions were those selectedfrom previous experiments. The color extinction in IU relates to100 g dry substance. When this value is multiplied by the dry sub-stance throughput per liter of resin, divided by 100, the result is theabsolute IU load per liter of resin. Multiplication with the degree ofdecolorization gives the absolute IU removed per liter of resin(Perschak, 1998). This is a useful parameter to define the decolor-ization capacity in plant design.

After fifteen cycles, the percentage of color removed was about73–79%, showing that the decolorization capacity was not reduced.The absolute IU load per liter of resin was about 4500–5000 per cy-cle. The absolute IU removed per liter of resin was about 3400–4000 per cycle. Using this parameter, the regeneration percentageswere about 45–50%, showing a strong adsorption of colorants inthe resin matrix. This fact may be due to hydrophobic interactionsbetween colorants and the resin matrix during the regeneration athigh salt concentration. Although the regeneration step could notdesorb completely colorants from the resin, the low volume ofregenerant used (2.5 BV per cycle) did not produce a decrease inthe decolorization efficiencies, which were about 73–79% in allthe experiments performed. According to references (Bento,1997), styrenic resins are prone to be fouled by organic compoundsand colorants, producing a quick decrease on resin capacity. How-ever, in the experimental runs carried out, there was not observeda sharp decrease in the resin decolorization capacity. Fig. 6 showssize exclusion chromatograms of decolorized juices after differentcycles. As is shown, there were almost no differences between thechromatograms, showing that the efficacy of the styrenic resin wasnot reduced throughout decolorization cycles. The colorant of20 kDa in raw thin juice was removed in all experimental runs.The removal of colored impurities above 100 kDa was about 65–75% whereas the reduction in the concentration of the 2 kDamolecular mass colorant was slightly lower, about 55–70%. Itwas not observed a reduction in the removal percentages throughthe decolorization cycles.

4. Conclusions

This paper summarizes the results of the study on the molecularsize distribution of beet sugar colorants during a decolorizationprocess using styrenic resins. The use of styrenic resins for decolor-izing thin juices has been successful, providing color reduction

434 M. Coca et al. / Journal of Food Engineering 89 (2008) 429–434

percentages about 75–80%. SEC study was useful to analyze theevolution of colorants. Beet colorants were adsorbed in the resinmatrix depending on their molecular masses. Colorants withmolecular mass of 20 kDa were completely removed, showing astrong anionic character. However, the removal of colorants withmolecular masses above 100 kDa and 2 kDa was considerablylower. Since some of the nonsugar substances are anionic com-pounds, part of the resin capacity may be required for the exchangeof multivalent anions that do not affect the color of the solution.The removals of nitrate, sulfate and citrate anions were completewhereas the concentration of chloride anions in decolorized juicesincreased considerably. The decolorization process does not alterjuice purity but the displacement of chloride ions from the resininto the juice would increase juice melassigenicity. Therefore,decolorization should be combined with a further demineraliza-tion step. The decolorizing capacity of the styrenic resin did notdecrease considerably after a color load about 70000 IU per literof resin. The waste produced in the regeneration stage consistedin 57 L of spent regenerant per m3 of juice treated. Further workaiming on the reductions in the consumption of chemical reagentsand the volume of wastewater using nanofiltration membranes arein progress.

Acknowledgements

The authors acknowledge the financial support of this researchto the Ministerio of Ciencia y Tecnología (MCyT, PPQ2000-0270-P4-02) and the company Azucarera Ebro, S.L. for its contributionto the development of the project.

References

Bento, L.S.M. 1992. Sugar liquor decolourization by ion exchange resins. In:Proceedings of SPRI Conference, New Orleans, USA.

Bento, L.S.M. 1996. Sugar colourants and ion exchange resins: influence of calciumand sucrose in sugar colourants removal from ion exchange resins. In:Proceedings of SPRI Conference, pp. 121–136.

Bento, L.S.M. 1997. Ion exchange resins for sugar liquors decolourization. In:Proceedings of SIT Conference, Montreal, Canada. Symposium A: Sugar Colour.

Bento, L.S.M, & Sá, S. 1998. Study of HMW compounds in sugar liquors fromcarbonation and ion exchange resins using GPC chromatography with anevaporative light scattering detector. In: Proceedings of SPRI Conference,Savannah, GA, USA, pp. 233–247.

Broughton, N.W., Sargent, D., Jones, G.C., Brown, B.W., 1991. Resin decolorization ofbeet sugar factory juices. In: Proceedings of CITS, pp. 477–502.

Cämmerer, B., Kroh, L.W., 1995. Investigations of the influence of reactionconditions on the elementary composition of melanoidins. Food Chemistry53, 55–59.

Cämmerer, B., Jalyschkov, V., Kroh, L.W., 2002. Carbohydrate structures as part ofthe melanoidin skeleton. International Congress Series 1245, 269–273.

Coca, M., García, M.T., González, G., Peña, M., García, J.A., 2004. Study of coloredcomponents formed in sugar beet processing. Food Chemistry 86, 421–433.

De Bruijn, J.M., 1986. Monosaccharides in alkaline medium: isomerization,degradation and oligomerization. PhD thesis, University of Technology, Delf.

García Agudo, J.A., García Cubero, M.T., González Benito, G., Peña Miranda, M., 2002.Removal of colored compounds from sugar solutions by adsorption onto anionicresins: equilibrium and kinetic study. Separation and Purification Technology29, 199–205.

Godshall, M.A., 1999. Removal of colorants and polysaccharides and the quality ofwhite sugar. In: Proceedings of 6� Symposium Association AVH, Reims, pp. 28–35.

Godshall, M.A., Vercellotti, J.R., Triche, R., 2002. Comparison of cane and beet sugarmacromolecules in processing. International Sugar Journal 104, 228–233.

Guimaraes, C., Bento, L.S.M., Mota, M., 1996. A study of sugar colorants through ionexchange and salt regeneration. International Sugar Journal 98, 584–587.

Gula, F., Paillat, D., 2005. Decolorization of refinery liquors: a technical andeconomic comparison between the different systems using activated carbon orresins. International Sugar Journal 107, 235–240.

Heitz, F., 1995. Les colorants en sucrerie. Actualités Techniques et Industrielles,538–544.

ICUMSA Methods Book, 1994. England, ICUMSA.Mersad, A., Lewandowski, R., Heyn, B., Decloux, M., 2001. Formation des colorants

de fabrication en sucrerie: reconstitution au laboratoire et analyses parspectrométrie. Industries Alimentaries et Agricoles 118, 42–51.

Perschak, F., 1998. Ion exchange and decolorization processes. In: van der Poel, P.W.,Schiweck, H., Schwartz, T. (Eds.), Sugar Technology: Beet and Cane SugarManufacture. Verlag Dr. Albert Bartens KG, Berlin, pp. 37–114.

Rafik, M., Mas, A., Elharfi, A., Schue, F., 1997. Decoloration de solutions sucrees parultrafiltration sur une membrane a base de poly(organocyclophosphazene).European Polymer Journal 33, 679–690.

Shore, M., Broughton, N.W., Dutton, J.V., Sissons, A., 1984. Factors affecting whitesugar color. Sugar Technology Reviews 12, 1–99.

Yaylayan, V.A., Kaminsky, E., 1998. Isolation and structural analysis of Maillardpolymers: caramel and melanoidin formation in glycine/glucose model system.Food Chemistry 63, 25–31.