S:Se

nsory

&Fo

odQu

ality

Sensory Differences Between Beet and CaneSugar SourcesBrittany L. Urbanus, Ginnefer O. Cox, Emily J. Eklund, Chelsea M. Ickes, Shelly J. Schmidt, and Soo-Yeun Lee

Abstract: Research concerning the sensory properties of beet and cane sugars is lacking in the scientific literature.Therefore, the objectives of this research were to determine whether a sensory difference was perceivable between beetand cane sugar sources in regard to their (1) aroma-only, (2) aroma and taste without nose clips, and (3) taste-only withnose clips, and to characterize the difference between the sugar sources using descriptive analysis. One hundred panelistsevaluated sugar samples using a tetrad test. A significant difference (P < 0.05) was identified between beet and canesugar sources when evaluated by aroma-only and taste and aroma without nose clips. However, there was no differencewhen tasted with nose clips. To characterize the observed differences, ten trained panelists identified and quantified keysensory attributes of beet and cane sugars using descriptive analysis. Analysis of variance indicated significant differences(P < 0.05) between sugar samples for 8 of the 10 attributes including: off-dairy, oxidized, earthy, and barnyard aroma,fruity and burnt sugar aroma-by-mouth, sweet aftertaste, and burnt sugar aftertaste. The sensory profile of beet sugar wascharacterized by off-dairy, oxidized, earthy, and barnyard aromas and by a burnt sugar aroma-by-mouth and aftertaste,whereas cane sugar was characterized by a fruity aroma-by-mouth and sweet aftertaste. This study shows that beet andcane sugar sources can be differentiated by their aroma and provides a sensory profile characterizing the differences. Assugar is used extensively as a food ingredient, sensory differences between beet and cane sugar sources once incorporatedinto different product matrices should be studied as a next step.

Keywords: beet, cane, descriptive analysis, sugar, tetrad

Practical Application: The results from this research document the differences between beet and cane sugar sourcesfrom a sensory perspective. Knowledge of sensory differences between the sugar sources supplement the analyticallydetermined volatile profile differences previously reported in the literature. The current research suggests that differencesbetween beet and cane sugars reside in their aromas. These findings encourage beet sugar manufacturers to continuedevelopment of a deodorization strategy to overcome the off-aromas perceived by consumers in beet sugar.

IntroductionSucrose, otherwise known as sugar, is an important commodity

worldwide, because of its influence on the sensory, physical, andchemical properties of a variety of food products. Sugar is read-ily formed in plants as a product of photosynthesis. Sugar beet(Beta vulgaris) and sugar cane (Saccharum officinarum) are the mostcommon plant sources for commercial scale sucrose extraction.Sucrose is recovered from the roots of sugar beets and from theleaves and stalks of sugar cane (Colonna and others 2000). Theextracted sucrose is processed to yield white refined granulatedsugar. The resultant sugar is composed of greater than 99% su-crose, regardless of its source, although differences in their volatileprofiles, thermal behaviors, and minor chemical components havebeen noted (Potter and Mansel 1992; Colonna and others 2000;Asadi 2005; Lu and others 2013). Remnants from the sugar plantsource and processing methods, along with water, constitute theremainder of the composition (Colonna and others 2000).

Although the chemical composition of beet and cane sugarsare nearly identical, differences in their sensory profiles have beensuggested. Monte and Maga (1982) described a preliminary tri-

MS 20140747 Submitted 5/2/2014, Accepted 6/11/2014. Authors are with Univ.of Illinois, 905 S. Goodwin Ave, Urbana, IL 61801, USA. Direct inquiries to authorLee (E-mail: soolee@illinois.edu).

angle test performed by Maga (1974, unpublished data) on beetand cane sugar solutions at varying concentrations (0.1 to 16.0wt.% in water) and temperatures (4, 22, and 35 °C). Results indi-cated a statistical difference between beet and cane sugar solutions,where temperature and concentration of the sugar solutions wereinfluential factors in panelists’ ability to detect a difference. Uponinformally questioning panelists as to the nature of the differencebetween the sugar solutions, most replied that the odor of the beetsugar solutions were “rooty and musty as compared to cane.”

Beet sugar is often described in the literature as having an ob-jectionable earthy and musty aroma based on instrumental anal-ysis (Acree and others 1976; Parliment and others 1977; Monteand Maga 1982; Marsili and others 1994; Pihlsgard 1997; Magneand others 1998). Analytical flavor chemistry techniques have at-tributed these off-aromas in beet sugar to geosmin and a varietyof volatile fatty acid compounds (Marsili and others 1994; God-shall and others 1995; Moore and others 2004). These volatilecompounds are associated with the microbial contamination fromthe soil, the beet root itself, and the decomposition of plant parts(Marsili and others 1994; Godshall and others 1995; Clarke andothers 1995; Lu and others 2003). Although the flavor profiledifferences between beet and cane sugars have been explored us-ing analytical chemistry techniques, the differences have yet to becharacterized using sensory descriptive analysis.

Eating is a cross-modal experience that combines aroma, taste,and tactile perceptions. Flavor perception is influenced by the

C© 2014 Institute of Food Technologists R©doi: 10.1111/1750-3841.12558 Vol. 79, Nr. 9, 2014 � Journal of Food Science S1763Further reproduction without permission is prohibited

S:Sensory&Food

Quality

Beet and cane sensory differences . . .

interaction between these different sensory modalities (Laing andJinks 1996; Belitz and others 2004). Due to the complexity offlavor, aromas are frequently confused as tastes. To isolate tastefrom the other sensory perceptions, nose clips are often used(Murphy and others 1977; Murphy and Cain 1980; O’Mahony1991; Abegaz and others 2004). The nose clip closes the nostrils,blocking aroma input, and, in turn, retronasal aroma perception.Therefore, before characterizing the sugar using descriptive anal-ysis, it was first important to determine whether panelists couldperceive a difference between beet and cane sugars with and with-out nose clips.

The objectives of this research were to determine whether asensory difference was perceivable between beet and cane sugarsources in regard to their (1) aroma-only, (2) aroma and tastewithout nose clips, and (3) taste-only with nose clips and to char-acterize the difference between the sugar sources using descriptiveanalysis. In view of the analytically determined volatile profile dif-ferences between beet and cane sugars, it was hypothesized thatthe sugars would differ in their aromas and that sugars from beetsources would receive higher ratings for off-aromas compared tocane sugars in the descriptive analysis.

Materials and Methods

Sample selectionIn the United States, there are a large number of sugar brands,

but a relatively limited number of separate beet and cane sugarmanufacturers. For example, Florida Crystal, Domino Sugar, andC&H Sugar are 3 very large cane sugar brands, but they are allowned by the same manufacturer (ASR Group 2013). Two manu-facturers of beet sugar, Michigan Sugar Company (Pioneer Sugar,Bay City, Mich., U.S.A.) and United Sugar Corporation (UnitedSugar, Minneapolis, Minn., U.S.A.) and 2 manufacturers of canesugar, ASR Group (C&H, Crockett, Calif., U.S.A.) and UnitedSugar Corporation (United Sugar), were used in this study. Thebeet and cane sugars from United Sugar Corporation were do-nated and the Pioneer Sugar and the C&H were purchased froma local grocery store (Urbana, Ill., U.S.A.).

Tetrad testSample preparation. Initial instrumental screening and pre-

vious R-index results indicated that sugars of like sources fromdifferent manufacturers exhibit similar sensory and thermal char-acteristics (Lu and others 2013; Urbanus 2014). Therefore, PioneerSugar and C&H were selected for use in the tetrad test as repre-sentative beet and cane sugar samples, respectively. To prepare thesugar samples, plastic 29.5-mL cups with lids (Solo Cup Co., Inc.,Chicago, Ill., U.S.A.), labeled with a randomized 3-digit code,were filled with approximately 1 g (1/4 tsp) of sugar.

Panelists. A total of 100 panelists (77F, 23M, age range 18 to55 y) participated in the tetrad test. The panelists were screenedbased on interest and availability. Panelists were instructed not toeat or drink at least 30 min prior to their scheduled session times.All panelists were compensated monetarily at the completion ofthe study.

Test design. The tetrad test took place in a room with par-titioned booths maintained at 22 °C and 33% relative humid-ity. Panelists evaluated the samples under incandescent lighting.Panelists attended one, 15- to 20-min session. The samples werepresented to the panelists in 3 sets of tetrad. Each set of samplescorresponded to different evaluation conditions: aroma-only byorthonasal sniffing, taste and aroma-by-mouth without nose clips,

and taste-only by mouth with nose clips (Bettertimes, Santa FeSprings, Calif., U.S.A.). The nose clips closed the panelist’s nos-trils to prevent retronasal olfactory perception. The order of thesets and the samples within each set were randomized per pan-elist. Randomization was generated by the Compusense five Plus(Version 5.0: Guelph, ON, Canada) data acquisition system usinga William’s Latin square design.

Each panelist was served their 1st tray of samples along withwarm and room temperature rinse water. For samples evaluatedby taste and aroma-by-mouth without nose clips and taste-onlyby mouth with nose clips, the panelists were instructed to rinsewith warm (26 to 29 °C) and room temperature purified water(Absopure, Urbana, Ill., U.S.A.) before the 1st sample and betweensubsequent samples, expectorate all rinses and samples, and to tastethe entire contents of the cup at once. Panelists were served the 4samples of tetrad side-by-side and asked to evaluate the samples inthe order from left to right. Their task was to sort the samples into2 groups of 2 samples based on similarity. There was a 1-minutebreak between each tetrad set. This protocol was repeated withthe 2nd and 3rd tetrad set. No replicate testing was done for thetetrad test

Data analysis. The raw data were collected using Compusensefive Plus (Version 5.0) and analyzed using IFPrograms software(Version 8.1: Richmond, Va., U.S.A.). The program computedvalues of d’, and binomial probabilities (P values) for sampledifferences. d’ is a measure of the estimated sensory differencebetween samples. The calculated P values were compared to a0.05 significance level to determine if a significant differenceexisted.

Descriptive analysis panelSample and reference preparation. Pioneer Sugar and

United Sugar were selected as representative beet sugars and C&Hand United Sugar were selected as representative cane sugars andwere used in the descriptive analysis test. To prepare the sugar sam-ples, plastic 29.5-mL cups with lids (Solo Cup Co., Inc.), labeledwith a randomized 3-digit code, were filled with approximately12 g (1 tablespoon) of sugar.

Reference samples were prepared in plastic cups with lids(Solo Cup Co., Inc.) labeled with the reference identity. Acomplete list of the finalized attributes, definitions, references,and reference preparation methods are provided in Table 1.All references were prepared no more than 24 h prior toevaluation.

Panelists. Ten panelists (8F, 2M, age range 23 to 45 y) partici-pated in the descriptive analysis portion of the study. Panelists wereselected based upon interest, experience, aptitude, and availabil-ity. All panelists were University of Illinois graduate students whohad prior experience participating in sensory descriptive analysispanels.

Test design. Quantitative descriptive analysis (Stone andothers 1974) was used to evaluate the sensory attributes of beetand cane sugars. The descriptive analysis panel took place over 6sessions. The sessions were held on 5 consecutive days. Panelistsattended one, 1-h session per day with the exception of day 5,in which panelists attended 2, 30-min sessions. Training was ac-complished in 4 sessions, which took place in a conference room(Larson-Powers and Pangborn 1978; Piggott and Mowat 1991;Shamaila and others 1992). Sample evaluation took place duringsessions 5 and 6 in a room with partitioned booths. The boothswere maintained at 22 °C and 33% relative humidity with incan-descent lighting. In between each sample and reference, panelists

S1764 Journal of Food Science � Vol. 79, Nr. 9, 2014

S:Se

nsory

&Fo

odQu

ality

Beet and cane sensory differences . . .

Table 1–Descriptive attributes, definitions, references, and intensities as generated by a trained descriptive panel evaluating beetand cane sugar samples. Reference intensities were determined by panel average.

Reference Reference ReferenceModality Attribute Definition product preparation rating

Aroma Off-dairy Off-aroma associate withdairy

Powdered milk (Nestle;Solon, Ohio, U.S.A.)

1/2 tsp in 2 oz. cup 7.8

Oxidized Aroma of wet cardboard Wet cardboard 1/2 cm square in 5 oz. cup,dipped in water

13.6

Earthy Aroma of root vegetable Raw beet 1 g in 2 oz. cup 9.0Barnyard Aroma of barnyard Goat milk (Mayenberg;

Turlock, Calif., U.S.A.)1/2 tsp in 2 oz. cup 13.1

Vanilla Aroma of vanillin Vanilla powder (SensientFlavors LLC; Indianapolis,Ind., U.S.A.)

Pinch of powder in 2 oz. cup 15.0

Aroma-by-mouth

Fruity Fruity aroma-by-mouth Dissolved cotton candy(Charms; Covington,Tenn., U.S.A.)

2 g in 100-mL water, 10-mLin 2 oz. cup

9.9

Burnt sugar Aroma-by-mouth of burntsugar

Dilute molasses (B&G Foods,Inc., Roseland, N.J.,U.S.A.)

10 drops in 100-mL water,10-mL in 2 oz. cup

10.7

Taste Sweet Sweet taste Half concentrated sugarsolution (Domino Foods,Inc., Yonkers, N.Y., U.S.A.)

100-mL sat. soln. with100-mL water, 1 t-bsp in2 oz. cup

12.6

Aftertaste Sweet Sweet aftertaste Half concentrated sugarsolution (Domino Foods,Inc.)

100-mL sat. soln. with100 mL water, 1 tbsp in2 oz. cup

9.9

Burnt sugar Aftertaste of burnt sugar Dilute molasses (B&G Foods,Inc., Roseland, N.J.,U.S.A.)

10 drops in 100-mL water,10-mL in 2 oz. cup

10.2

were instructed to rinse with warm water and room tempera-ture water (Absopure). All samples, references, and rinses wereexpectorated.

During the 1st session, panelists were refreshed on the use ofscales and references to describe attributes of the sugar samples.Next, panelists evaluated the aroma, aroma-by-mouth, taste, andaftertaste of the sugar samples and generated verbal descriptors foreach sample that fit into those sensory modality categories. Pan-elists also generated references, which served as physical represen-tations for the descriptors generated. These tasks were performedindividually, followed by a group discussion.

During the 2nd session, panelists evaluated the physical ref-erences generated from session one, in order to determine if thereferences adequately represented the sensory attributes of the sam-ples. Panelists continued term generation and reference refinementduring the 2nd session. This procedure was repeated during the3rd session until panelists reached a consensus of generated termsand references by omitting redundant terms and refining the listto include only the most pertinent attributes used to describethe differences among the samples. The finalized list of termsand references was compiled (Table 1). Panelists then rated all ofthe finalized reference samples relative to sugar sample intensitiesfor the specific attribute. An 11-point scale, 0 to 10, representedthe range of intensity of the attribute that panelists perceived inthe sugar samples. Reference samples with intensities perceivedto be greater than the strongest sugar sample for that attributereceived scores greater than 10. Panelists’ scores were compiledand the average rating for each reference was calculated. The av-erage rating of each reference was used to anchor the scale thatpanelists used for sample rating (Table 1). Session 4 was used forpanelist calibration in order to ensure consistency across the pan-elists. Panelists practiced rating samples with respect to referenceintensities.

During the 5th and 6th sessions, panelists were instructed toreview the references before entering the booths for sample rating.Once in the booths, panelists evaluated sugar samples for each ofthe attributes using an 11-point scale ranging from 0 to 10. Verbaland written instructions for test procedures were provided to thepanelist prior to evaluation. Panelists were presented with the sugarsamples and asked to rate their perceived intensities on the specificattributes as compared to the anchored references. Samples wereevaluated in duplicate over 2 sessions. Sample randomization andevaluation, and data collection were done using Compusense fivePlus (Version 5.0).

Data analysisData collected from the panelists were compiled into a spread-

sheet and analyzed using Microsoft Excel and XLSTAT (Version2009: Addinsoft USA, New York, N.Y., U.S.A.). Analysis of vari-ance (ANOVA) was conducted for each of the 10 sensory at-tributes to assess differences in the mean scores of the 4 types ofsugar samples for each attribute. The calculated probabilities ob-tained from the analysis were compared to the significance level of0.05. Adjusted F-values by mixed model ANOVA were calculatedfor attributes with significant judge-by-sample interaction. Judge-by-sample interaction was used as the error term for all attributesin this calculation. Fisher’s least significant difference (LSD) wasconducted on attributes determined as being significantly differentby ANOVA. Pearson correlation coefficients were computed toshow the relationship among the significant attributes.

Results and DiscussionFor the tetrad test analysis, values of d’ and binomial probabil-

ities for sample differences are provided in Table 2. A significantdifference was identified between beet and cane sugars when the

Vol. 79, Nr. 9, 2014 � Journal of Food Science S1765

S:Sensory&Food

Quality

Beet and cane sensory differences . . .

Table 2–Tetrad results for beet and cane sugars by evaluationconditions: percent of correct responses, d’, and binomial prob-abilities for sample differences.

Evaluation condition Modality % Correcta d’ P value

Without nose clips Aroma-only 75 1.89 0.00∗Aroma and taste 61 1.40 0.00∗

With nose clips Taste-only 38 0.51 0.19

∗Indicates significance at P < 0.05.apercent correct based on the total sample size of 100 panelists.

Table 3–ANOVA table for 10 attributes describing 4 sugarsamples.

Interaction AdjustedAttribute Judge Sugar (J × S) F-value

AromaOff-dairy 1.706 9.815∗∗∗ 2.202∗ 4.46∗Oxidized 5.767∗∗∗ 25.437∗∗∗ 3.878∗∗∗ 6.56∗Earthy 1.002 16.061∗∗∗ 2.145∗ 7.49∗∗Barnyard 2.287∗ 35.89∗∗∗ 2.699∗∗ 13.30∗∗Vanilla 4.891∗∗∗ 10.29∗∗∗ 3.752∗∗∗ 2.74

Aroma-by-mouthFruity 3.461∗∗ 21.499∗∗∗ 1.992∗ 10.79∗∗Burnt sugar 3.272∗∗ 25.642∗∗∗ 1.693

TasteSweet 10.541∗∗∗ 5.054∗∗ 1.941∗ 2.60

AftertasteSweet 10.42∗∗∗ 4.637∗∗ 1.412Burnt sugar 3.346∗∗ 30.472∗∗∗ 2.077∗ 14.67∗∗∗

∗, ∗∗, and ∗∗∗ indicate significance at P < 0.05, P < 0.01, and P < 0.001, respectively.

samples were evaluated by aroma-only and taste and aroma with-out nose clips. However, samples could not be differentiated whentasted with nose clips. Because nose clips are intended to stopvolatiles from entering the olfactory receptors in order to isolatetaste perception, the data suggest that beet and cane sugars cannotbe differentiated by their taste only. Differences between beet andcane sugars that were evaluated by tasting samples without noseclips can be attributed to retronasal aroma differences.

The descriptive analysis study was carried out to characterizethe nature of the observed difference found in the tetrad test. In thedescriptive analysis study, 10 terms, including off-dairy aroma, ox-idized aroma, earthy aroma, barnyard aroma, vanilla aroma, fruityaroma-by-mouth, burnt sugar aroma-by-mouth, sweet taste, sweetaftertaste, and burnt sugar aftertaste, were generated by paneliststo describe the sample set. The attribute definitions and referenceintensity ratings for each attribute given by the panelists are shownin Table 1.

ANOVA was conducted for each of the 10 sensory attributesgenerated by the descriptive analysis panel and the results are sum-marized in Table 3. Judges scores were significantly different (P <

0.05) for 8 of 10 attributes. This source of variation is common indescriptive analysis testing and may be due to panelists differing intheir use of the scale when rating samples. All of the 10 attributeswere found to be significantly different (P < 0.05) across thesugar samples. Judge-by-sample interaction (J × S) was significant(P < 0.05) for 8 of the attributes, which is telling of inconsistencyamong the panelists. To account for the variation due to panelistsacross the samples, adjusted F-values were calculated for attributeswhose interaction was significant. Of the 8 calculated adjustedF-values, 6 were noted as being significantly different, includingoff-dairy aroma, oxidized aroma, earthy aroma, barnyard aroma,fruity aroma-by-mouth, and burnt sugar aftertaste. The adjustedF-values are shown in Table 3.

Mean separation analysis by LSD was conducted and the at-tribute means for each sugar sample are given in Table 4. Pioneerand United Sugar Corporation beet sugars were significantly dif-ferent for the attributes of: off-dairy aroma, oxidized aroma, barn-yard aroma, and burnt sugar aroma-by-mouth. Mean attributescores for C&H and United Sugar Corporation cane sugars weresignificantly different for 2 of the 8 attributes: barnyard aroma andfruity aroma-by-mouth. When comparing beet sugar versus canesugar sources, significant differences existed between at least onebeet and cane sample for all 8 attributes. The findings providedevidence that there is more variation between sugar sources thanbetween manufacturers of like sugar sources.

Pearson correlation coefficients are shown in Table 5. Signifi-cant positive correlations were observed for off-aroma attributes,including off-dairy aroma, oxidized aroma, and earthy aroma, andbetween barnyard aroma and earthy aroma. Positive correlationswere also observed between burnt sugar aroma-by-mouth andburnt sugar aftertaste. Oxidized aroma and sweet aftertaste weresignificantly correlated in the negative direction.

Data from the tetrad test suggested that beet and cane sugarsources can be differentiated by aroma, but not by taste-only. Thisis in agreement with the descriptive analysis results, which foundno significant differences in taste attributes between the sugars.Hence, the differences between beet and cane sugar sources re-side in their aroma characteristics, which, as suggested by theliterature, can be attributed to their volatile profiles. Analyticalflavor chemistry techniques have been previously used to iden-tify the compounds responsible for the characteristic off-aromas inbeet sugar. The geosmin (trans-1,10-dimethyl-trans-(9)-decalol),in combination with volatile fatty acids, including butanoic acidisovaleric acids (3-methyl-butanoic acid), were found to have anaroma identical to the off-aroma perceived in beet sugar (Marsiliand others 1994; Godshall and others 1995; Moore and others2004). These compounds emit an earthy, musty aroma, which isperceivable in beet sugar. The terms generated in the descriptiveanalysis study herein are in agreement with the analytically deter-mined flavor attributes reported in the literature to characterizebeet sugar aroma.

Off-aromas in beet sugar originate from soil microorganisms,the beet itself, and from the degradation of the tops, leaves, orroot of the beet (Marsili and others 1994; Clarke and others1995; Godshall and others 1995; Lu and others 2003). Crystalgrowth takes place in a supersaturated solution. Once the crystalreaches the desired size, the syrup is separated from the crystals bycentrifugation. Although most of the syrup is removed, most ofthe volatiles present in beet sugar are thought to reside in a thinouter layer of concentrated syrup surrounding the crystal (Clarkeand others 1995; Godshall and others 1995; Colonna and others1996).

Little published work is available on the elimination of off-aromas in beet sugar. Air circulation and ventilation of sugar dur-ing storage resulted in partial elimination of the off-odors in beetsugar (Clarke and others 1995; Colonna and others 1996; Duf-faut and others 2004). Treatment suggestions, such as polishingcarbon, additional washing in the centrifuge, and introducing awarm, dry stream of helium to the surface of the crystal, have alsobeen reported to achieve partial off-aroma elimination (Clarkeand others 1995; Colonna and others 1996; Duffaut and others2004).

The next step in this research should focus on the impact ofthe sensory differences between beet and cane sugar sources indifferent product matrices, since sugar is used extensively as a food

S1766 Journal of Food Science � Vol. 79, Nr. 9, 2014

S:Se

nsory

&Fo

odQu

ality

Beet and cane sensory differences . . .

Table 4–Mean intensity ratings and Fisher’s least significant difference (LSD) for significant attributes of 4 sugar samples rated byan 11-point scale from 0 to 10.

Aroma Aroma-by-mouth Aftertaste

Off-dairy Oxidized Earthy Barnyard Fruity Burnt sugar Sweet Burnt sugar

Pioneer (Beet) 6.20b 5.40b 5.35b 6.15c 3.55a 7.90c 7.35ab 7.40b

United Sugar Corporation (Beet) 8.10c 7.95c 6.75b 7.75d 4.45ab 6.35b 6.45a 6.30b

C&H (Cane) 4.75a 3.60a 2.35a 1.75a 8.15c 2.70a 8.15b 2.55a

United Sugar Corporation (Cane) 5.00ab 4.30a 3.10a 3.60b 5.30b 3.45a 7.40b 3.00a

Means within a column that are noted with the same superscript letter indicate no significant difference (P < 0.05) between sugar type for a given attribute.

Table 5–Pearson correlation matrix of descriptive analysis attributes from the descriptive analysis panel.

Variables Off-dairy A Oxidized A Earthy A Barnyard A Fruity ABM Burnt sugar ABM Sweet AT Burnt sugar AT

Off-dairy A 1.000Oxidized A 0.995 1.000Earthy A 0.967 0.958 1.000Barnyard A 0.940 0.941 0.990 1.000Fruity ABM −0.624 −0.638 −0.785 −0.853 1.000Burnt sugar ABM 0.700 0.667 0.853 0.860 −0.878 1.000Sweet AT −0.921 −0.953 −0.903 −0.928 0.734 −0.614 1.000Burnt sugar AT 0.752 0.717 0.887 0.884 −0.853 0.996 −0.646 1.000

A, aroma; ABM, aroma-by-mouth; AT, aftertaste.Values in bold are different from 0 with a significance level α = 0.05.

ingredient. A difference test could be used to determine whetherpanelists can perceive a difference between various food productsmade with beet versus cane sugar sources. A difference wouldsuggest that additional factors, besides market price, be consideredby food manufacturers when selecting the sugar source for theirproduct.

ConclusionAlthough the composition of beet and cane sugar sources are

nearly identical (>99%), data from the tetrad test suggested thatthey can be distinguished by their sensory properties. Descriptiveanalysis findings revealed differences between the 2 different sugarsources, but similarities between sugars from different manufac-turers from like sugar sources. Beet sugar samples were character-ized by off-flavors including off-dairy, oxidized, earthy, and barn-yard aromas and by a burnt sugar aroma-by-mouth and aftertaste,whereas cane sugar was associated with a fruity aroma-by-mouthand sweet aftertaste.

This research is of importance because it documents the differ-ences between beet and cane sugar sources from a sensory per-spective. Knowledge on sensory differences between the sugarsources will supplement the analytically determined volatile pro-file differences previously reported in the literature. Findings fromthis study provide insight to sugar manufacturers, in particularbeet sugar manufacturers, by emphasizing that the difference be-tween beet and cane sugars reside in their aroma. Deodorizationstrategies such as the use of odor scavenging packaging should bestudied in the future to provide recommendations to beet sugarmanufacturers for product quality improvements. The next step inthis research will examine the influence of the sensory differencesbetween beet and cane sugar sources in different product matrices.In view of the findings from the current study, we speculate thatpanelist’s will be able to distinguish between products made withbeet and cane sugars.

AcknowledgmentsThe authors thank the sugar donation from United Sugar Cor-

poration and the vanilla flavor donation from Sensient FlavorsLLC.

ReferencesAbegaz EG, Tandon KS, Scott JW, Baldwin EA, Shewfelt RL. 2004. Partitioning taste from

aromatic flavor notes of fresh tomato (Lycopersicon esculentum, Mill) to develop predictivemodels as a function of volatile and nonvolatile components. Postharvest Biol Technol 34(3):227–35.

Acree TE, Lee CY, Butts RM, Barnard J. 1976. Geosmin, the earthy component of table beetodor. J Agric Food Chem 24(2):430–1.

Asadi M. 2005. Basics of beet-sugar technology. Anonymous beet-sugar handbook. Hoboken,N.J.: John Wiley & Sons, Inc. p 1–68.

Our Brands [Internet]. West Palm Beach, FL: ASR Group; 2013. Available from:www.asr-group.com. Accessed 2014 April 30.

Belitz HD, Grosch W, Schieberle P. 2004. Aroma compounds. Anonymous food chemistry. 3rded. New York: Springer Berlin Heidelberg. p 342–408.

Clarke MA, Godshall MA, Blanco RS, Miranda XM. 1995. Color and odor in beet sugarmanufacture and storage. Int Sugar J 97(1158):248–52.

Colonna WJ, McGillivray T, Samaraweera U, Torgeson T. 1996. Odor in beet sugar: somecausative agents and preventative measures. Proceedings of the Conference by The SugarProcessing Research Institute; April 14–17; New Orleans, La., p 198–220.

Colonna WJ, Samaraweera U, Clarke MA, Cleary M, Godshall MA, White JS. 2000. Sugar.In: Anonymous Kirk-Othmer encyclopedia of chemical technology. New York, N.Y.: JohnWiley & Sons, Inc.

Duffaut E, Godshall MA, Grimm C. 2004. The effect of ozone and air on off-odors in beetsugar. Proceedings of the Conference by The Sugar Processing Research Institute; April 4-7;Atlanta, Ga., p 193–208.

Godshall MA, Grimm CC, Clarke MA. 1995. Sensory properties of white beet sugars. Int SugarJ 97(1159B):296–343.

Laing D, Jinks A. 1996. Flavour perception mechanisms. Trends Food Sci Technol 7(12):387–89.

Larson-Powers N, Pangborn RM. 1978. Descriptive analysis of the sensory properties of bever-ages and gelatins containing sucrose or synthetic sweeteners. J Food Sci 43(1):47–51.

Lu G, Edwards CG, Fellman JK, Mattinson DS, Navazio J. 2003. Biosynthetic origin of geosminin red beets (Beta vulgaris L.). J Agric Food Chem 51(4):1026–9.

Lu L, Lee JW, Thomas L, Schmidt SJ. 2013. Differences in the thermal behavior of beet andcane sugars. Proceedings of the 74th Annual Institute of Food Technologists, Chicago, Ill.

Magne V, Mathlouthi M, Robilland B, Magne M, Mathlouthi B. 1998. Determination ofsome organic acids and inorganic anions in beet sugar by ionic HPLC. Food Chem 61(4):449–53.

Marsili RT, Miller N, Kilmer GJ, Simmons RE. 1994. Identification and quantitation of theprimary chemicals responsible for the characteristic malodor of beet sugar by purge-and-trapGC-MS-OD techniques. J Chromatogr Sci 32(5):165–71.

Monte WC, Maga JA. 1982. Flavor chemistry of sucrose. Sugar Technol Rev 8(3):181–204.Moore SJ, Godshall MA, Grimm CC. 2004. Comparison of Two methods of volatile analysis

for determining the causes of off-odors in white beet sugars SPME and headspace. Int SugarJ 105(1253):224–9.

Vol. 79, Nr. 9, 2014 � Journal of Food Science S1767

S:Sensory&Food

Quality

Beet and cane sensory differences . . .

Murphy C, Cain WS. 1980. Taste and olfaction: independence vs interaction. Physiol Behav24(3):601–5.

Murphy C, Cain WS, Bartoshuk LM. 1977. Mutual action of taste and olfaction. Sens Processes1(3):204–11.

O’Mahony M. 1991. Taste perception, food quality and consumer acceptance. J Food Qual14(1):9–31.

Parliment T, Kolor M, Maing I. 1977. Identification of the major volatile components cookedbeets. J Food Sci 42(6):1592–3.

Piggott JR, Mowat RG. 1991. Sensory aspects of maturation of cheddar cheese by descriptiveanalysis. J Sens Stud 6(1):49–62.

Pihlsgard P. 1997. The properties of sugar focusing on odours and flavours—a literature review.SIK Rapport (634).

Potter R, Mansel R, inventors; University of South Florida, assignee. 1992. Assay for theDetection of Beet Sugar Adulteration of Food Products. U.S. patent 5128243.

Shamaila M, Powrie WD, Skura BJ. 1992. Sensory evaluation of strawberry fruit stored un-der modified atmosphere packaging (MAP) by quantitative descriptive analysis. J Food Sci57(5):1168–84.

Stone H, Sidel J, Oliver S, Woolsey A, Singleto RC. 1974. Sensory evaluation by quantitativedescriptive analysis. Food Technol 28(11):24–33.

Urbanus BL. 2014. Sensory differences between beet and cane sugars. [MSc thesis]. Urbana-Champaign, Ill.: University of Illinois. 156p.

S1768 Journal of Food Science � Vol. 79, Nr. 9, 2014