Impact of Almond Form and Moisture Content on Texture Attributes and Acceptability Zata Vickers, Amanda Peck, Theodore Labuza, and Guangwei Huang
Abstract: The objectives of this study were to measure sensory texture attributes of 5 types of almonds (blanched slivered, natural sliced, whole blanched, whole dry roasted, and natural whole) conditioned at 4 different moisture levels, to measure liking of a subset of these products, and to compare the sensory texture measurements with consumer liking ratings. Thirteen panelists trained to evaluate almond texture rated the texture attributes of the 20 almond samples. A panel of 113 almond consumers rated their liking of a subset of 8 of these almonds. Compared with the whole almonds, sliced and slivered almonds had less hardness, less crunchiness, less cohesiveness, less tooth packing, and required fewer chews and swallows to consume. Compared with slivered almonds, sliced almonds were more powdery, had more surface roughness, more loose particles, and were crisper. Compared with slivered almonds, sliced almonds were less hard, broke into fewer pieces, had less moistness and cohesiveness of mass, less fatty film, and required fewer chews and fewer swallows to consume. Dry roasted almonds were generally harder, more crisp, more crunchy, and produced more loose particles than natural almonds, which were, in turn, more hard, crisp, and crunchy than blanched almonds. As moisture content increased, moistness of mass and cohesiveness of mass increased. Crispness, number of pieces, hardness, crunchiness, persistence of crunch, and particulate mass decreased with increasing moisture content. Consumer texture liking ratings were highly positively correlated with the attributes crispness, crunchiness, and persistence of crunch. Keywords: almonds, acceptability, descriptive analysis, moisture content, sensory texture
Introduction Almonds can be consumed plain or used as ingredients in a wide variety of sweet and savory foods. They provide a nutrient dense, antioxidant rich, and “heart healthy” food suitable for many diets (Spiller and others 1992, 1998; Hyson and others 2002; Jambazian and others 2005;Milbury and others 2006). Almond sensory properties are desirable alone and add character and interest when used as ingredients in other foods. Almond texture, specifically hardness and crunchiness, is its most salient sensory attribute (Civille and others 2010). Sensory texture can be evaluated by gently rubbing the surface of the almond across the lips to perceive surface texture attributes, by using the molars to crush the almonds to evaluate textures perceived on the first chew, by further chewing the almond to evaluate texture attributes perceived during/after repeated chewing, and by assessing any residual texture sensations (for example, oily film, moistness, and particles) left in the mouth after expectoration or swallowing (Civille and others 2010). Variations in texture arise from different almond varieties (Civille and others 2010), roasting times and temperatures (Guerrero and others 1997; Gou and others 2000; Varela and others 2006), and storage conditions (Kennedy and Wright 2008). MS 20131355 Submitted 9/23/2013, Accepted 4/27/2014. Authors Vickers, Peck, and Labuza are with Dept. of Food Science and Nutrition, Univ. of Minnesota, 1334 Eckles Ave., St. Paul, MN 55108, U.S.A. Author Huang is with Almond Board of California, 1150 9th St., #1500, Modesto, CA 95354, U.S.A. Direct inquiries to author Zata Vickers (E-mail: [email protected]).
R C 2014 Institute of Food Technologists
doi: 10.1111/1750-3841.12509 Further reproduction without permission is prohibited
Roasting almonds increases brittleness, crispness, crunchiness, and hardness, and reduces deformability and toughness. (Saklar and others 1999; Varela and others 2006). Almonds roasted at higher temperatures or for longer times also contain less moisture. However, moisture content and roasting conditions have been confounded in previous studies. This study examines the influence of moisture content on almond texture across almond forms, roasting and blanching treatments. The objectives of this study were to measure sensory texture attributes of 5 types of almond forms (blanched slivered, natural sliced, whole blanched, natural whole, and dry roasted natural whole) conditioned at 4 different moisture levels, to measure liking of a subset of the whole almonds, and to compare sensory texture measurements and moisture content of these whole almonds with consumer liking ratings.
Materials and Methods Descriptive analysis Subjects. Thirteen subjects, members of the Univ. of Minnesota Sensory Center trained descriptive analysis panel, participated. All were 6-n-propothiouracil (PROP) tasters, and none had any food allergies or sensitivities. All recruiting and experimental procedures were approved by the Univ. of Minnesota’s Institutional Review Board. Products tested. Five almond forms (natural whole, dry roasted natural whole, blanched whole, blanched slivered, and natural sliced), all processed from a single variety (Monterey) by Paramount Farms, Inc. (Lost Hills, Calif., U.S.A.), were provided Vol. 79, Nr. 7, 2014 r Journal of Food Science S1399
S: Sensory & Food Quality
Practical Application: In addition to roasting and blanching, water content of almonds is important for key texture properties. Almond producers can use this information to provide customers with almonds having texture properties important to consumers.
Almond texture, water activity, and liking . . . Table 1–Almond types and moisture treatments used in the test water in the 500 μL. The amount of water per sample weight in and coding used in the report. μg water per μg extraction solution was determined after subType
Moisture treatment
Raw whole Raw whole Raw whole Raw whole Dry roasted whole Dry roasted whole Dry roasted whole Dry roasted whole Blanch Whole Blanch Whole Blanch Whole Blanch Whole Blanch Sliver Blanch Sliver Blanch Sliver Blanch Sliver Sliced Sliced Sliced Sliced
Lowest Low Normal High Lowest Low Normal High Lowest Low Normal High Lowest Low Normal High Lowest Low Normal High
tracting the average of the triplicate blanks as methanol alone is very hygroscopic. Appropriate dilution factors were then used to Raw LL calculate the grams water per gram almond solids for the almond Raw Low Raw Norm samples. Raw High Water activity was measured using the Decagon CX 2 (Pullman, DryR LL Wash., U.S.A.) on each of the triplicate samples for each of the 4 DryR Low moisture levels of each of the 5 almond forms. DryR Norm Panelists received 3 almonds of each whole almond sample (natDryR High ural, blanched, and roasted), 3 slivered almonds, or about 1 g of Blch LL Blch Low sliced almonds. Almond samples for sensory testing were wrapped R Blch Norm in Saran to maintain moisture content, and served at room temBlch High perature in coded 60 mL cups with lids. Sliver LL Coding
Sliver Low Sliver Norm Sliver High Sliced LL Sliced Low Sliced Norm Sliced High
S: Sensory & Food Quality
by the Almond Board of California. These are the most commonly consumed almond forms. The Monterey variety almonds are often utilized to make these whole and cut forms. The dry roasted natural almonds were originally roasted in a hot air conveyor roaster at 245 to 265 °F for 90 min. The blanched almonds were originally prepared by scalding brown skin almonds in hot water (205 °F) for 3 min, and then drying at 245 °F for 60 min following skin removal. Once received, the almonds were then dried or humidified to obtain 3 targeted levels of moisture simulating those encountered in commercial products, A 4th high moisture level was added to mimic the wettest condition a commercial product could be without safety concern for panelists’ tasting. Preparation of these samples was as follows: 1 kg of each almond form was spread out onto an aluminum tray created to fit into 20 gallon fish tanks. Fish tanks contained either desiccant (Drierite 0.05% relative humidity [RH]) or a saturated solution of sodium chloride (75% RH). The tray was then placed on a rack above the desiccant or salt solution and the tank covered with a plastic plate. The tray in the desiccant tank was held for the proposed period (2 wk for the lowest moisture, 1 wk for low moisture) at room temperature. For the high moisture level, the tray containing the almonds was placed over the salt solution for 48 h. The normal moisture refers to the almonds as obtained from the processor. Table 1 lists the almond forms, moisture treatments, and coding used in the tables and the plots of this report. Moisture content and water activity were measured in triplicate for each almond form. Moisture content was measured by the Karl Fischer method using the Aquatest cma Karl Fischer Coulometric Titrator (Photovolt Co., Indianapolis, Ind., U.S.A.). Basically, a sample of the almond form was first ground in a Waring Blender, Triplicate samples were prepared by weighing approximately 1 g (actual weight measured) into a 50 mL Erlenmeyer flask. To this, 50 g of methanol was added and a rubber septum was put onto the top opening. Triplicate blanks were also prepared with only methanol. The samples were then placed on a rotating shaker and extracted for 18 h. After settling, approximately 500 μL of the methanol extraction (weighed syringe before and after) was injected into the instrument for determination of the amount of S1400 Journal of Food Science r Vol. 79, Nr. 7, 2014
Experimental procedure Training. Prior to testing, the panelists participated in 4 training sessions during which they tasted representative samples of the almonds, and, working from the texture descriptors from Civille and others (2010), selected appropriate sensory texture attributes and eating techniques relevant to the almond samples to be tested. They modified the definitions and added/deleted attributes as necessary to describe differences among the almonds. Panelists practiced rating these attributes on the scales used for the actual testing. We provided feedback to them on their individual performance and further discussed attributes panelists were having difficulty understanding. (We shared with panelists plots of their replicate ratings so that they could see how well they replicated their judgments. We also shared with them plots of their mean ratings compared to the mean ratings of all the participants. These plots were constructed using PanelCheck 1.3.2 (PanelCheck 2014) Testing. Panelists participated in 5 test sessions. Over the 5 sessions, they evaluated all 20 samples in replicate. Within a session, serving orders were balanced for order and carryover effects. During these testing sessions, participants were asked to place an almond in their mouth, bite down, and chew with their molars. Each panelist evaluated each sample by rating the intensity of each attribute on 20 point line scales labeled “none” at the left end and “intense” at the right end. See Table 2 for the list of attributes, references, and definitions that comprised the lexicon for these products Data analysis R We used analyses of variance (SAS PROC GLM, version 9.2) to determine whether almond samples differed in any of the specific texture attributes. The attribute intensity was the dependent variable; panelist, almond, replicate, and taste position were fixed predictors (α = 0.05). We used Student–Neuman Keuls (SNK) Tests to determine which products differed significantly from which others for a specific attribute. To summarize these data, we used a principal components analR ysis (PCA) with Varimax rotation (XLSTAT ) using the mean trained panel ratings of all 20 almond samples and all attributes that differed significantly among those almond samples. To choose whole almonds to use in the Consumer Liking Study, we conducted a PCA with Varimax rotation using the mean values for each attribute that differed significantly among only the whole almonds. We selected almonds for the Consumer Liking Study that best represented the positive and negative extremes of the first 4 principal components. Thus, the almonds selected for the Consumer study best represented the breadth or range of the sensory texture qualities of these almonds.
Almond texture, water activity, and liking . . . Table 2–Sensory texture attributes, their definitions, and references. Numbers by references indicate the 20-point scale value of that reference.
Surface Powdery/Fuzzy Macroroughness Loose particles Oiliness First bite (made with molars) Hardness to split/crack Crispness Number of pieces Hardness to grind pieces Chewdown Number of chews to Bolus Moistness of mass (5 chews)
Definition
Reference
Action: Run the sample gently over the lips Perceive: Amount of small, fine particles detected. Action: Run sample gently over the lips Perceive: The amount of irregular structure in the surface (may be grainy or bumpy) Action: Run the sample gently over the lips Perceive: Amount of loose particles on the surface Action: Run the sample gently over the lips Perceive: The amount of wetness/oiliness on the surface Action: Using the molars, bite through the sample Perceive: The force required to bite into and crack the sample Action: Using the molars, bite through the sample Perceive: Amount of High pitched sound Action: Using the molars, bite through the sample Perceive: Number of pieces the almond fractures into Action: Using the molars, bite through the cracked sample after the first chew Perceive: Amount of work taken to create smaller pieces Action: Chew using your molars Perceive: How many chews until you feel the sample is ready to swallow/expectorate Action: Chew using your molars Perceive: Amount of wetness/oiliness in the chewed mass
Cohesiveness of mass
Action: Chew using your molars Perceive: The degree to which the mass holds together
Particulate mass (at the time of swallowing)
Action: Chew using your molars Perceive: The amount of particles in the mass Action: Chew using your molars Perceive: The amount of fibrous material between teeth (in the area where you floss) Action: Chew using your molars Perceive: Amount of low pitched sound Action: Chew using your molars Perceive: The number of chews that the sample still has a crunchy sound Action: Once completely chewed, swallow the sample Perceive: Number of swallows needed to clear mouth of almond mass.
Fibers between teeth Crunchiness Persistence of crunch Number of swallows Residual Toothpack Loose particles Fatty/oily film
Action: Swallow or expectorate the sample Perceive: The amount of product left in the teeth crevices after it has left the mouth Action: Swallow or expectorate the sample Perceive: The amount of chalky/grainy particles remaining in the mouth after expectoration Action: Swallow or expectorate the sample Perceive: The amount of fatty/oily residue felt by the tongue when moved over the surface of the mouth
Consumer liking study Subjects. We recruited 113 judges from students and staff of the Univ. of Minnesota who expressed an interest in participating in sensory tests. Panelists were 18 y or older, had no food allergies, and had consumed almonds in the previous month. They were paid for participating. The Univ. of Minnesota’s Institutional Review Board approved all recruiting and experimental procedures.
R Cocoa almond (Emerald Cocoa roast Almonds) = 8 R Almond (Trader Joe’s raw almonds) = 7
R Salted almond (Blue Diamond Roasted Salted Almonds) = 8 R Cocktail peanut (Market Pantry Target Corporation, Minneapolis, Minn., U.S.A.) = 7 R Zone Perfect chocolate peanut butter bar (Abbot Nutrition, Columbus, Ohio, U.S.A.) = 4 Celery = 18 R Saltine Cracker (Market Pantry ) = 10
R Zone Perfect bar = 4 R Pure Protein bar (chocolate deluxe. Sport Nutritional Supplements, Inc. Bay Port, N.Y., U.S.A.) = 9 R Power Bar chocolate (Powerbar, Inc. Florham Park, N.J., U.S.A.) = 12 Pure protein bar = 15 Zone perfect bar = 3
Celery = 12 Almond (Trader Joe’s raw almonds) = 7
Zone perfect bar = 12 Pure protein bar = 12 R Olive oil (Pompeian ) = 14
Products. We chose the following products for the consumer liking study because they were positioned most distant from the origin on the principal components: Blanched Whole High Moisture (Blch High), Blanch Whole Lowest Moisture (Blch LL), Natural Whole Low Moisture (Raw Low), Natural Whole Lowest Moisture (Raw LL), Natural Whole High Moisture (Raw High), Dry Roasted Natural Whole High Moisture (DryR High), Dry Roasted Natural Whole Low Moisture
Vol. 79, Nr. 7, 2014 r Journal of Food Science S1401
S: Sensory & Food Quality
Attribute
Almond texture, water activity, and liking . . . (DryR Low), and Dry Roasted Natural Whole Normal Moisture (DryR Norm). Each panelist received 3 almonds of each product, wrapped in Saran, in a lidded 60 mL plastic cup labeled with a random 3-digit number.
Experimental procedure We served all 8 samples, balanced for order and carryover effects, to panelists in a single session. Subjects were asked to take a few bites of each almond sample prior to rating that sample for overall liking, liking of texture, liking of flavor, hardness, crispness, crunchiness, and tooth packing. Liking ratings were made on 20 point scales labeled dislike extremely at the left end and like extremely at the right end. Ratings of the texture attributes were made on 20 point scales labeled none at the left end and extremely at the right end. Data analysis. To determine whether samples differed in any of the attributes, we used analyses of variance (ANOVA) (SAS PROC Mixed, version 9.2) with the attribute intensity as the dependent variable and subject, almond, and taste position as predictors (α = 0.05). Subject was a random predictor in the models. The SNK Test was used to determine which almonds differed significantly from which others for a specific attribute.
Almond
Blanched Slivered
moistness
Sliver High Sliver LL
2
Blanched 0
-2
-4 -8
-6
Sliver Norm Sliver Low
numberofchews residualparƟcles Hardnesstogrind Hardness to split Raw High numberofpieces Raw LL Raw Norm DryR LL Natural DryR High DryRLow Sliced LL Sliced Raw Low DryR Norm Sliced High Natural Whole Sliced Low Sliced Norm macroroughness looseparƟcles powdery -4 -2 0 2 4 6
S1402 Journal of Food Science r Vol. 79, Nr. 7, 2014
0.31 0.27 0.29 0.22 0.26
0.041 0.041 0.027 0.030 0.039
0.40 0.35 0.37 0.27 0.19
0.042 0.045 0.023 0.036 0.032
0.51 0.64 0.30 0.59 0.51
0.053 0.063 0.029 0.052 0.047
0.68 0.72 0.66 0.69 0.40
0.066 0.075 0.061 0.070 0.076
Water activity and moisture content Water activities (moisture content—grams water per gram almond solids) ranged from 0.22 to 0.31 (0.027 to 0.041) in the lowest (LL) moisture group; 0.19 to 0.40 (0.023 to 0.045) in the low moisture group; 0.30 to 0.64 (0.029 to 0.063) in the normal moisture group; and 0.40 to 0.72 (0.061 to 0.076) in the high moisture group (Table 3).
Figure 1–Principal components biplot of the first 2 components computed for all almond types and moisture levels. Black bolded names indicate almond types and moisture level; red unbolded names represent locations of texture attributes. Component 1 was highly positively correlated with hardness and more chewing; whole almonds loaded highly on this component. Component 2 was positively correlated with moistness and fatty film and negatively correlated with powdery, loose particles, and surface roughness; blanched almonds loaded highly on this component.
Blch High Whole faƩyfilm Blch Low moistnessofmass Blch Norm cohesiveness of mass Blch LL
Component 1 (31 %)
Moisture content g water/g solid
Results and Discussion
Blanched Whole
4
Water activity
Lowest Blch LL Raw LL Dry R LL Sliver LL Sliced LL Low Blch Low Raw Low DryR Low Sliver Low Sliced Low Normal Blch Norm Raw Norm DryR Norm Sliver Norm Sliced Norm Higher Blch High Raw High DryR High Sliver High Sliced High
6
Component 2 (27%)
S: Sensory & Food Quality
Comparison of descriptive analysis and consumer liking data Data analysis. We conducted a PCA with Varimax rotation (XLSTAT) using the mean values for each product for each statistically significant attribute from the descriptive analysis data, but using only the almonds evaluated in the consumer liking test. We added texture liking ratings from the consumers as a supplemental variable to determine the relationship between the textural attributes from the descriptive analysis data and the consumer liking data.
Table 3–Water activity and moisture contents of the 20 almonds.
8
Almond texture, water activity, and liking . . . Descriptive analysis Almond form
less hard to grind, had less moistness of mass, less cohesiveness of mass, less fatty film, fewer number of chews, and fewer number of swallows (Figure 1). Comparison of whole, sliced, and slivered almonds. Compared with Mean values of all texture attributes, moisture content, and the whole almonds, sliced and slivered almonds were less hard, less water activity for each product are available in Table S1 to S5. We crunchy, had less cohesiveness of mass, less particulate mass, fewer have used PCA biplots to summarize these data in this manuscript. fibers between teeth, less tooth pack, less residual particles, fewer number of chews, and fewer number of swallows (Figure 1). Com- Comparisons among whole almonds. Dry roasted natural almonds pared with slivered almonds, sliced almonds were more powdery, were generally more crisp, crunchy, and powdery, and had more had more macroroughness, more loose particles, and were more loose particles and more fibers between teeth than the natural crisp (Figure 1). Compared with slivered almonds, sliced almonds almonds, which, in turn, had more of these attributes than the were less moist, less hard to split, broke into fewer pieces, were blanched almonds (Figure 2). Blanched almonds were less hard,
Figure 2–Principal component biplot showing positions of almonds and texture attributes on the first 2 principal components from the principal components analysis (PCA) on the descriptive analysis data of only the whole almonds. Component 1(horizontal axis) was highly positively correlated with crunchiness and persistence of crunch; component 2 (vertical axis) was highly positively correlated with surface roughness and highly negatively correlated with moistness and fatty film. Measures of water activity and moisture content were added to the PCA as supplementary variables.
4
macroroughness powdery
Raw High
2
fibersbetweenteeth DryR
High
Component 2 (29%)
Raw Norm numberofchews
DryR
Raw LL DryR LL crunchiness DryR Norm persistenceofcrunch
wateracvity 0
looseparcles residualparcles crispness Raw Lowhardnesstogrind hardnesstosplit Low
numberofpieces
moisturecontent
water cohesivenessofmass
More dry
moistnessofmass
Blch LL
Blch Norm
-2
Blch High
fayfilm moistness
Blch LL
-4
-6 -10
-8
-6
-4
-2
0
2
4
6
8
Component1 (36 %)
Figure 3–Principal component biplot showing positions of almonds and texture attributes on principal components 3 and 4 from the PCA done on the descriptive analysis data of only the whole almonds. Component 3 (horizontal axis) was positively related to toothpack and to fibers between teeth, and negatively related to number of chews; component 4 (vertical axis) was positively related to number of swallows and presence of particles. Measures of water activity and moisture content were added to the PCA as supplementary variables.
3
DryR Norm numberofswallows 2
Component 4 (9%)
looseparƟcles parƟculatemass
Blch High DryR high
powdery numberofpieces cohesivenessofmass DryR Low
1
DryR LL
Raw LL
fibersbetweenteeth
0
wateracƟvity Blch Low moistnessofmass Blch Norm toothpack moisturecontent
numberofchews -1
Raw Low
Raw High
Raw Norm Blch LL
-2 -4
-3
-2
-1
0
1
2
3
4
Component 3 (11 %) Vol. 79, Nr. 7, 2014 r Journal of Food Science S1403
S: Sensory & Food Quality
Blanched
Almond texture, water activity, and liking . . . Table 4–Mean ratings from consumer test (over all judges, N = 113). All ratings were made on 20 point scales where the left end was labeled “none” and the right end “intense.” All attributes differed significantly among the 8 almonds. (All P < 0.0001.) Numbers in italics below the means are standard errors of the means. Attribute Overall liking Texture liking Flavor liking Hardness Crispness Crunchiness Toothpack a bc
Raw LL
Raw Low
Raw High
DryR Low
DryR Norm
DryR High
Blched LL
Blched High
11.5b 0.39 11.9b 0.38 11.4b 0.42 14.4a 0.31 12.7ab 0.38 13.2a 0.33 10.7a 0.36
11.3b 0.43 11.5ba 0.41 11b 0.43 14.1a 0.32 12a 0.4 12.4a 0.4 10.7a 0.36
6.6c 0.42 5.4d 0.41 8.6c 0.46 6d 0.33 4.5c 0.31 4.1c 0.29 10.7a 0.45
13.4a 0.42 14a 0.34 13a 0.44 12b 0.37 13.3a 0.33 13.4a 0.32 9.8a 0.38
13.4a 0.43 14a 0.38 12.8a 0.45 11.8b 0.34 13.1a 0.37 13.1a 0.37 9.5a 0.35
7.4c 0.38 7.8c 0.39 7.5d 0.42 9.9c 0.4 7.7b 0.41 7.2b 0.39 10.6a 0.39
12.1b 0.36 13.1a 0.37 11.3b 0.4 11.3b 0.37 12.5a 0.36 12.5a 0.35 9.9a 0.39
4.5d 0.39 3.8e 0.38 5.9e 0.42 4e 0.32 2.5d 0.23 2.4d 0.26 8.2b 0.47
Means in a row with letter superscripts in common were not significantly different (P > 0.05).
broke into fewer pieces, and had less persistent crunch than the dry roasted or the natural almonds. Almonds with higher moisture content generally had more moistness and cohesiveness of mass and were less crisp, hard, and crunchy (Figure 2). Both water activity and moisture content were highly negatively correlated with component 1 (−0.92 and −0.83 respectively). Component 3 was positively related to toothpack and to fibers between teeth, and negatively related to number of chews; component 4 was positively related to number of swallows and presence of particles (Figure 3).
Figure 4–Biplot showing positions of almonds and texture attributes on the first 2 principal components from the PCA on the descriptive analysis data of only the whole almonds used in the consumer test. Component 1 was highly positively correlated with crunchiness and persistence of crunch; component 2 was highly positively correlated with number of chews and highly negatively correlated with fibers between teeth. Texture liking ratings from the consumer panelists were added to the PCA as supplementary variables.
4
numberofchews
2
Raw LL Raw Low
DryR High Component 2 (13%)
S: Sensory & Food Quality
Consumer Liking Study Liking ratings. Dry Roasted Norm and Dry Roasted Low were the best liked almonds, followed by Blanch LL, Raw LL, and Raw Low. Blanch High was the least liked almond (Table 4). Ratings of overall liking, texture liking, and flavor liking all followed the same pattern. Discrimination among the samples was most distinct (largest F value from the ANOVA) for the texture
liking ratings, so we included them as supplemental variables in the PCA for the consumer test data. PCA on products evaluated in the consumer test. Principal component 1 was highly positively correlated with crunchiness, persistence of crunch, and crispness (Figure 4). Blanch LL scored high on this component. Blanch High and Raw High scored low. Principal component 2 was highly positively correlated with number of chews. Raw LL and Raw Low scored high on this component; Dry Roasted Low and Raw High scored low (Figure 4). Principal component 3 was highly positively correlated with macroroughness and negatively correlated with moistness and fatty film. Raw Low scored high on this component; Blanch High and Blanch LL scored low. Principal component 4 was highly correlated with number of swallows and number of particles. Dry Roasted Norm scored high on this component; Blanch LL scored low on this component.
Blch High 0
hardnesstogrind hardnesstosplit residualparƟcles parƟculatemass crunchiness persistenceofcrunch
cohesivenessofmass faƩyfilm
Texture Liking
powdery crispness Blch LL numberofpieces looseparƟcles toothpack
DryR Norm
fibersbetweenteeth
Raw High
-2
DryR Low -4
-8
-6
-4
-2
0
Component 1 (37 %) S1404 Journal of Food Science r Vol. 79, Nr. 7, 2014
2
4
6
8
Almond texture, water activity, and liking . . .
10
Raw LL
Blch LL DryR Low
9
DryR LL
8
DryR Norm
Raw Low
Blch Low
Crunchiness
7 6
DryR High
5
Blch Norm
Raw HIgh
Raw Norm
4 3
Blch High
2 1 0 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Water acƟvity
Texture liking ratings were highly positively correlated with principal component 1 and with the attributes crispness, crunchiness, persistence of crunch, number of pieces (all factor loadings > 0.91). Texture liking ratings were unrelated to principal components 2 to 4.
Discussion Although correlations between crunchiness and moisture and between crunchiness and water activity were relatively high, they appear to be due to a grouping of higher moisture almonds with lower crunchiness and a 2nd grouping of lower moisture almonds with higher crunchiness (Figure 5). The sounds produced by a food during chewing are important for crispness/crunchiness, and the loudness of those sounds can be diminished by increasing the water activity of dry foods (Vickers and Bourne 1976a,b; Lewicki and others 2007). Relationships between water activity and sensory crispness/crunchiness or between water activity and instrumental correlates of crispness/crunchiness show that increases in water activity are commonly accompanied by decreases in these attributes. Our observation that further decreasing the moisture level below about 0.04 (or a water activity less than 0.4) did not appear to increase crunchiness is consistent with work by others indicating that water activities above 0.3 to 0.4 can be detrimental to crispness/crunchiness or its indicators (Katz and Labuza 1981; Lewicki and others 2007).
crisper. Compared with slivered almonds, sliced almonds were less hard, broke into fewer pieces, had less moistness and cohesiveness of mass, less fatty film, and required fewer chews and fewer swallows to consume. Dry roasted almonds were generally harder, more crisp, more crunchy, and produced more loose particles than natural almonds, which were, in turn, more hard, crisp, and crunchy than blanched almonds. As moisture content increased, moistness of mass and cohesiveness of mass increased. Crispness, number of pieces, hardness, crunchiness, persistence of crunch, and particulate mass decreased with increasing moisture content. Consumer texture liking ratings were highly positively correlated with the attributes crispness, crunchiness, and persistence of crunch.
Acknowledgments Funding for this research was provided by the Almond Board of California.
Author Contributions
GH conceived the objectives of the study, provided the different almond forms, and reviewed the manuscript. ZV designed the sensory testing part of the study, completed the data analysis, and drafted the manuscript. TL adjusted the moisture content of the almonds, measured moisture content and water activity, and reviewed the manuscript. AP provided the descriptive analysis training and conducted both the descriptive analysis and the Conclusion Compared with the whole almonds, sliced and slivered almonds consumer sensory tests. had less hardness, less crunchiness, less cohesiveness, less tooth packing, and required fewer chews and swallows to consume. References GV, Lapsley K, Huang G, Yada S, Seltsam J. 2010. Development of an almond Compared with slivered almonds, sliced almonds were more pow- Civille lexicon to assess the sensory properties of almond varieties. J Sens Stud 25(1):146–62. dery, had more surface roughness, more loose particles, and were doi:10.1111/j.1745-459X.2009.00261.x Vol. 79, Nr. 7, 2014 r Journal of Food Science S1405
S: Sensory & Food Quality
Figure 5–Water activity compared with crunchiness. Points represent the mean sensory scores for crunchiness of whole almonds at specific water activities. Crunchiness ratings were made on a 20 point scale where 0 = none and 20 = intense. Abbreviations by each point note the almond and treatment that point represents. The red oval encompasses the products with similarly high levels of crunchiness; the blue oval encompasses the products with “similarly” low levels of crunchiness.
Almond texture, water activity, and liking . . . Gou P, Diaz I, Guerrero L, Valero A, Arnau J, Romero A. 2000. Physico-chemical and sensory property changes in almonds of desmayo largueta variety during toasting / Cambios en las propiedades f´ısico-qu´ımicas y sensoriales de almendras de la variedad Desmayo Largueta durante el tostado. Food Sci Technol Intl 6(1):1–7. doi:10.1177/ 108201320000600101 Guerrero L, Gou P, Arnau J. 1997. Descriptive analysis of toasted almonds: a comparison between expert and semi-trained assessors. J Sens Stud 12(1):39–54. doi:10.1111/j.1745459X.1997.tb00052.x Hyson DA, Schneeman BO, Davis PA. 2002. Almonds and almond oil have similar effects on plasma lipids and LDL oxidation in healthy men and women. J Nutr 132(4):703– 7. Jambazian PR, Haddad E, Rajaram S, Tanzman J, Sabat´e J. 2005. Almonds in the diet simultaneously improve plasma alpha-tocopherol concentrations and reduce plasma lipids. J Am Diet Assoc 105(3):449–54. Katz EE, Labuza TP. 1981. Effecty of water activity on the sensory crispness and mechanical deformation of snack food products. J Food Sci 46:403–9. Kennedy J, Wright A. 2008. Storage study of various almond forms. Technical Report Natick/TR-09/011L 1-14. Lewicki P, Marzec A, Kuropatwa M. 2007. Influence of water activity on texture of corn flakes. Acta Agrophys 9(1):79–90. Milbury PE, Chen C, Dolnikowski G, Blumberg J. 2006. Determination of flavonoids and phenolics and their distribution in almonds. J Agric Food Chem 54:5027–33. doi: 10.1021/jf0603937 PanelCheck 2014. Available from: http://www.panelcheck.com. Accessed 2014 January 10. Saklar S, Ungan S, Katnas S. 1999. Instrumental crispness and crunchiness of roasted hazelnuts and correlations with sensory assessment. J Food Sci 64(6):1015–9. Spiller GA, Jenkins DJ, Cragen LN, Gates JE, Bosello O, Berra K, Rudd C, Stevenson M, Superko R. 1992. Effect of a diet high in monounsaturated fat from almonds on plasma cholesterol and lipoproteins. J Am Coll Nutr 11(2):126–30. Spiller GA, Jenkins DA, Bosello O, Gates JE, Cragen LN, Bruce B. 1998. Nuts and plasma lipids: an almond-based diet lowers LDL-C while preserving HDL-C. J Am Coll Nutr 17(3):285–90. Varela P, Chen J, Fiszman S, Povey MJW. 2006. Crispness assessment of roasted almonds by an integrated approach to texture description: texture, acoustics, sensory and
S: Sensory & Food Quality S1406 Journal of Food Science r Vol. 79, Nr. 7, 2014
structure. J Chemometrics 20(67):311–20. doi:http://dx.doi.org/10.1016/j.foodres.2008. 03.009 Vickers Z, Bourne MC. 1976a. Crispness: a review. J Food Sci 41(5):1153–7. Vickers Z, Bourne MC. 1976b. A psychoacoustical theory of crispness. J Food Sci 41(5):1158– 64.
Supporting Information Disclaimer: Supplementary materials have been peer-reviewed but not copyedited. Table S1 and S2, containing mean values and standard errors for all the sensory attributes of all the products, should be made available online. Table S3. Mean sensory texture attribute values (over all judges, N = 13) slivered almonds only plus F and P values. Table S4. Mean sensory texture attribute values (over all judges, N = 13) sliced almonds only plus F and P values. Table S5. Mean sensory texture attribute values (over all judges, N = 13) by almond type (collapsed over moisture contents) plus F and P values.
Abstract: The objectives of this study were to measure sensory texture attributes of 5 types of almonds (blanched slivered, natural sliced, whole blanched, whole dry roasted, and natural whole) conditioned at 4 different moisture levels, to measure liking of a subset of these products, and to compare the sensory texture measurements with consumer liking ratings. Thirteen panelists trained to evaluate almond texture rated the texture attributes of the 20 almond samples. A panel of 113 almond consumers rated their liking of a subset of 8 of these almonds. Compared with the whole almonds, sliced and slivered almonds had less hardness, less crunchiness, less cohesiveness, less tooth packing, and required fewer chews and swallows to consume. Compared with slivered almonds, sliced almonds were more powdery, had more surface roughness, more loose particles, and were crisper. Compared with slivered almonds, sliced almonds were less hard, broke into fewer pieces, had less moistness and cohesiveness of mass, less fatty film, and required fewer chews and fewer swallows to consume. Dry roasted almonds were generally harder, more crisp, more crunchy, and produced more loose particles than natural almonds, which were, in turn, more hard, crisp, and crunchy than blanched almonds. As moisture content increased, moistness of mass and cohesiveness of mass increased. Crispness, number of pieces, hardness, crunchiness, persistence of crunch, and particulate mass decreased with increasing moisture content. Consumer texture liking ratings were highly positively correlated with the attributes crispness, crunchiness, and persistence of crunch. Keywords: almonds, acceptability, descriptive analysis, moisture content, sensory texture
Introduction Almonds can be consumed plain or used as ingredients in a wide variety of sweet and savory foods. They provide a nutrient dense, antioxidant rich, and “heart healthy” food suitable for many diets (Spiller and others 1992, 1998; Hyson and others 2002; Jambazian and others 2005;Milbury and others 2006). Almond sensory properties are desirable alone and add character and interest when used as ingredients in other foods. Almond texture, specifically hardness and crunchiness, is its most salient sensory attribute (Civille and others 2010). Sensory texture can be evaluated by gently rubbing the surface of the almond across the lips to perceive surface texture attributes, by using the molars to crush the almonds to evaluate textures perceived on the first chew, by further chewing the almond to evaluate texture attributes perceived during/after repeated chewing, and by assessing any residual texture sensations (for example, oily film, moistness, and particles) left in the mouth after expectoration or swallowing (Civille and others 2010). Variations in texture arise from different almond varieties (Civille and others 2010), roasting times and temperatures (Guerrero and others 1997; Gou and others 2000; Varela and others 2006), and storage conditions (Kennedy and Wright 2008). MS 20131355 Submitted 9/23/2013, Accepted 4/27/2014. Authors Vickers, Peck, and Labuza are with Dept. of Food Science and Nutrition, Univ. of Minnesota, 1334 Eckles Ave., St. Paul, MN 55108, U.S.A. Author Huang is with Almond Board of California, 1150 9th St., #1500, Modesto, CA 95354, U.S.A. Direct inquiries to author Zata Vickers (E-mail: [email protected]).
R C 2014 Institute of Food Technologists
doi: 10.1111/1750-3841.12509 Further reproduction without permission is prohibited
Roasting almonds increases brittleness, crispness, crunchiness, and hardness, and reduces deformability and toughness. (Saklar and others 1999; Varela and others 2006). Almonds roasted at higher temperatures or for longer times also contain less moisture. However, moisture content and roasting conditions have been confounded in previous studies. This study examines the influence of moisture content on almond texture across almond forms, roasting and blanching treatments. The objectives of this study were to measure sensory texture attributes of 5 types of almond forms (blanched slivered, natural sliced, whole blanched, natural whole, and dry roasted natural whole) conditioned at 4 different moisture levels, to measure liking of a subset of the whole almonds, and to compare sensory texture measurements and moisture content of these whole almonds with consumer liking ratings.
Materials and Methods Descriptive analysis Subjects. Thirteen subjects, members of the Univ. of Minnesota Sensory Center trained descriptive analysis panel, participated. All were 6-n-propothiouracil (PROP) tasters, and none had any food allergies or sensitivities. All recruiting and experimental procedures were approved by the Univ. of Minnesota’s Institutional Review Board. Products tested. Five almond forms (natural whole, dry roasted natural whole, blanched whole, blanched slivered, and natural sliced), all processed from a single variety (Monterey) by Paramount Farms, Inc. (Lost Hills, Calif., U.S.A.), were provided Vol. 79, Nr. 7, 2014 r Journal of Food Science S1399
S: Sensory & Food Quality
Practical Application: In addition to roasting and blanching, water content of almonds is important for key texture properties. Almond producers can use this information to provide customers with almonds having texture properties important to consumers.
Almond texture, water activity, and liking . . . Table 1–Almond types and moisture treatments used in the test water in the 500 μL. The amount of water per sample weight in and coding used in the report. μg water per μg extraction solution was determined after subType
Moisture treatment
Raw whole Raw whole Raw whole Raw whole Dry roasted whole Dry roasted whole Dry roasted whole Dry roasted whole Blanch Whole Blanch Whole Blanch Whole Blanch Whole Blanch Sliver Blanch Sliver Blanch Sliver Blanch Sliver Sliced Sliced Sliced Sliced
Lowest Low Normal High Lowest Low Normal High Lowest Low Normal High Lowest Low Normal High Lowest Low Normal High
tracting the average of the triplicate blanks as methanol alone is very hygroscopic. Appropriate dilution factors were then used to Raw LL calculate the grams water per gram almond solids for the almond Raw Low Raw Norm samples. Raw High Water activity was measured using the Decagon CX 2 (Pullman, DryR LL Wash., U.S.A.) on each of the triplicate samples for each of the 4 DryR Low moisture levels of each of the 5 almond forms. DryR Norm Panelists received 3 almonds of each whole almond sample (natDryR High ural, blanched, and roasted), 3 slivered almonds, or about 1 g of Blch LL Blch Low sliced almonds. Almond samples for sensory testing were wrapped R Blch Norm in Saran to maintain moisture content, and served at room temBlch High perature in coded 60 mL cups with lids. Sliver LL Coding
Sliver Low Sliver Norm Sliver High Sliced LL Sliced Low Sliced Norm Sliced High
S: Sensory & Food Quality
by the Almond Board of California. These are the most commonly consumed almond forms. The Monterey variety almonds are often utilized to make these whole and cut forms. The dry roasted natural almonds were originally roasted in a hot air conveyor roaster at 245 to 265 °F for 90 min. The blanched almonds were originally prepared by scalding brown skin almonds in hot water (205 °F) for 3 min, and then drying at 245 °F for 60 min following skin removal. Once received, the almonds were then dried or humidified to obtain 3 targeted levels of moisture simulating those encountered in commercial products, A 4th high moisture level was added to mimic the wettest condition a commercial product could be without safety concern for panelists’ tasting. Preparation of these samples was as follows: 1 kg of each almond form was spread out onto an aluminum tray created to fit into 20 gallon fish tanks. Fish tanks contained either desiccant (Drierite 0.05% relative humidity [RH]) or a saturated solution of sodium chloride (75% RH). The tray was then placed on a rack above the desiccant or salt solution and the tank covered with a plastic plate. The tray in the desiccant tank was held for the proposed period (2 wk for the lowest moisture, 1 wk for low moisture) at room temperature. For the high moisture level, the tray containing the almonds was placed over the salt solution for 48 h. The normal moisture refers to the almonds as obtained from the processor. Table 1 lists the almond forms, moisture treatments, and coding used in the tables and the plots of this report. Moisture content and water activity were measured in triplicate for each almond form. Moisture content was measured by the Karl Fischer method using the Aquatest cma Karl Fischer Coulometric Titrator (Photovolt Co., Indianapolis, Ind., U.S.A.). Basically, a sample of the almond form was first ground in a Waring Blender, Triplicate samples were prepared by weighing approximately 1 g (actual weight measured) into a 50 mL Erlenmeyer flask. To this, 50 g of methanol was added and a rubber septum was put onto the top opening. Triplicate blanks were also prepared with only methanol. The samples were then placed on a rotating shaker and extracted for 18 h. After settling, approximately 500 μL of the methanol extraction (weighed syringe before and after) was injected into the instrument for determination of the amount of S1400 Journal of Food Science r Vol. 79, Nr. 7, 2014
Experimental procedure Training. Prior to testing, the panelists participated in 4 training sessions during which they tasted representative samples of the almonds, and, working from the texture descriptors from Civille and others (2010), selected appropriate sensory texture attributes and eating techniques relevant to the almond samples to be tested. They modified the definitions and added/deleted attributes as necessary to describe differences among the almonds. Panelists practiced rating these attributes on the scales used for the actual testing. We provided feedback to them on their individual performance and further discussed attributes panelists were having difficulty understanding. (We shared with panelists plots of their replicate ratings so that they could see how well they replicated their judgments. We also shared with them plots of their mean ratings compared to the mean ratings of all the participants. These plots were constructed using PanelCheck 1.3.2 (PanelCheck 2014) Testing. Panelists participated in 5 test sessions. Over the 5 sessions, they evaluated all 20 samples in replicate. Within a session, serving orders were balanced for order and carryover effects. During these testing sessions, participants were asked to place an almond in their mouth, bite down, and chew with their molars. Each panelist evaluated each sample by rating the intensity of each attribute on 20 point line scales labeled “none” at the left end and “intense” at the right end. See Table 2 for the list of attributes, references, and definitions that comprised the lexicon for these products Data analysis R We used analyses of variance (SAS PROC GLM, version 9.2) to determine whether almond samples differed in any of the specific texture attributes. The attribute intensity was the dependent variable; panelist, almond, replicate, and taste position were fixed predictors (α = 0.05). We used Student–Neuman Keuls (SNK) Tests to determine which products differed significantly from which others for a specific attribute. To summarize these data, we used a principal components analR ysis (PCA) with Varimax rotation (XLSTAT ) using the mean trained panel ratings of all 20 almond samples and all attributes that differed significantly among those almond samples. To choose whole almonds to use in the Consumer Liking Study, we conducted a PCA with Varimax rotation using the mean values for each attribute that differed significantly among only the whole almonds. We selected almonds for the Consumer Liking Study that best represented the positive and negative extremes of the first 4 principal components. Thus, the almonds selected for the Consumer study best represented the breadth or range of the sensory texture qualities of these almonds.
Almond texture, water activity, and liking . . . Table 2–Sensory texture attributes, their definitions, and references. Numbers by references indicate the 20-point scale value of that reference.
Surface Powdery/Fuzzy Macroroughness Loose particles Oiliness First bite (made with molars) Hardness to split/crack Crispness Number of pieces Hardness to grind pieces Chewdown Number of chews to Bolus Moistness of mass (5 chews)
Definition
Reference
Action: Run the sample gently over the lips Perceive: Amount of small, fine particles detected. Action: Run sample gently over the lips Perceive: The amount of irregular structure in the surface (may be grainy or bumpy) Action: Run the sample gently over the lips Perceive: Amount of loose particles on the surface Action: Run the sample gently over the lips Perceive: The amount of wetness/oiliness on the surface Action: Using the molars, bite through the sample Perceive: The force required to bite into and crack the sample Action: Using the molars, bite through the sample Perceive: Amount of High pitched sound Action: Using the molars, bite through the sample Perceive: Number of pieces the almond fractures into Action: Using the molars, bite through the cracked sample after the first chew Perceive: Amount of work taken to create smaller pieces Action: Chew using your molars Perceive: How many chews until you feel the sample is ready to swallow/expectorate Action: Chew using your molars Perceive: Amount of wetness/oiliness in the chewed mass
Cohesiveness of mass
Action: Chew using your molars Perceive: The degree to which the mass holds together
Particulate mass (at the time of swallowing)
Action: Chew using your molars Perceive: The amount of particles in the mass Action: Chew using your molars Perceive: The amount of fibrous material between teeth (in the area where you floss) Action: Chew using your molars Perceive: Amount of low pitched sound Action: Chew using your molars Perceive: The number of chews that the sample still has a crunchy sound Action: Once completely chewed, swallow the sample Perceive: Number of swallows needed to clear mouth of almond mass.
Fibers between teeth Crunchiness Persistence of crunch Number of swallows Residual Toothpack Loose particles Fatty/oily film
Action: Swallow or expectorate the sample Perceive: The amount of product left in the teeth crevices after it has left the mouth Action: Swallow or expectorate the sample Perceive: The amount of chalky/grainy particles remaining in the mouth after expectoration Action: Swallow or expectorate the sample Perceive: The amount of fatty/oily residue felt by the tongue when moved over the surface of the mouth
Consumer liking study Subjects. We recruited 113 judges from students and staff of the Univ. of Minnesota who expressed an interest in participating in sensory tests. Panelists were 18 y or older, had no food allergies, and had consumed almonds in the previous month. They were paid for participating. The Univ. of Minnesota’s Institutional Review Board approved all recruiting and experimental procedures.
R Cocoa almond (Emerald Cocoa roast Almonds) = 8 R Almond (Trader Joe’s raw almonds) = 7
R Salted almond (Blue Diamond Roasted Salted Almonds) = 8 R Cocktail peanut (Market Pantry Target Corporation, Minneapolis, Minn., U.S.A.) = 7 R Zone Perfect chocolate peanut butter bar (Abbot Nutrition, Columbus, Ohio, U.S.A.) = 4 Celery = 18 R Saltine Cracker (Market Pantry ) = 10
R Zone Perfect bar = 4 R Pure Protein bar (chocolate deluxe. Sport Nutritional Supplements, Inc. Bay Port, N.Y., U.S.A.) = 9 R Power Bar chocolate (Powerbar, Inc. Florham Park, N.J., U.S.A.) = 12 Pure protein bar = 15 Zone perfect bar = 3
Celery = 12 Almond (Trader Joe’s raw almonds) = 7
Zone perfect bar = 12 Pure protein bar = 12 R Olive oil (Pompeian ) = 14
Products. We chose the following products for the consumer liking study because they were positioned most distant from the origin on the principal components: Blanched Whole High Moisture (Blch High), Blanch Whole Lowest Moisture (Blch LL), Natural Whole Low Moisture (Raw Low), Natural Whole Lowest Moisture (Raw LL), Natural Whole High Moisture (Raw High), Dry Roasted Natural Whole High Moisture (DryR High), Dry Roasted Natural Whole Low Moisture
Vol. 79, Nr. 7, 2014 r Journal of Food Science S1401
S: Sensory & Food Quality
Attribute
Almond texture, water activity, and liking . . . (DryR Low), and Dry Roasted Natural Whole Normal Moisture (DryR Norm). Each panelist received 3 almonds of each product, wrapped in Saran, in a lidded 60 mL plastic cup labeled with a random 3-digit number.
Experimental procedure We served all 8 samples, balanced for order and carryover effects, to panelists in a single session. Subjects were asked to take a few bites of each almond sample prior to rating that sample for overall liking, liking of texture, liking of flavor, hardness, crispness, crunchiness, and tooth packing. Liking ratings were made on 20 point scales labeled dislike extremely at the left end and like extremely at the right end. Ratings of the texture attributes were made on 20 point scales labeled none at the left end and extremely at the right end. Data analysis. To determine whether samples differed in any of the attributes, we used analyses of variance (ANOVA) (SAS PROC Mixed, version 9.2) with the attribute intensity as the dependent variable and subject, almond, and taste position as predictors (α = 0.05). Subject was a random predictor in the models. The SNK Test was used to determine which almonds differed significantly from which others for a specific attribute.
Almond
Blanched Slivered
moistness
Sliver High Sliver LL
2
Blanched 0
-2
-4 -8
-6
Sliver Norm Sliver Low
numberofchews residualparƟcles Hardnesstogrind Hardness to split Raw High numberofpieces Raw LL Raw Norm DryR LL Natural DryR High DryRLow Sliced LL Sliced Raw Low DryR Norm Sliced High Natural Whole Sliced Low Sliced Norm macroroughness looseparƟcles powdery -4 -2 0 2 4 6
S1402 Journal of Food Science r Vol. 79, Nr. 7, 2014
0.31 0.27 0.29 0.22 0.26
0.041 0.041 0.027 0.030 0.039
0.40 0.35 0.37 0.27 0.19
0.042 0.045 0.023 0.036 0.032
0.51 0.64 0.30 0.59 0.51
0.053 0.063 0.029 0.052 0.047
0.68 0.72 0.66 0.69 0.40
0.066 0.075 0.061 0.070 0.076
Water activity and moisture content Water activities (moisture content—grams water per gram almond solids) ranged from 0.22 to 0.31 (0.027 to 0.041) in the lowest (LL) moisture group; 0.19 to 0.40 (0.023 to 0.045) in the low moisture group; 0.30 to 0.64 (0.029 to 0.063) in the normal moisture group; and 0.40 to 0.72 (0.061 to 0.076) in the high moisture group (Table 3).
Figure 1–Principal components biplot of the first 2 components computed for all almond types and moisture levels. Black bolded names indicate almond types and moisture level; red unbolded names represent locations of texture attributes. Component 1 was highly positively correlated with hardness and more chewing; whole almonds loaded highly on this component. Component 2 was positively correlated with moistness and fatty film and negatively correlated with powdery, loose particles, and surface roughness; blanched almonds loaded highly on this component.
Blch High Whole faƩyfilm Blch Low moistnessofmass Blch Norm cohesiveness of mass Blch LL
Component 1 (31 %)
Moisture content g water/g solid
Results and Discussion
Blanched Whole
4
Water activity
Lowest Blch LL Raw LL Dry R LL Sliver LL Sliced LL Low Blch Low Raw Low DryR Low Sliver Low Sliced Low Normal Blch Norm Raw Norm DryR Norm Sliver Norm Sliced Norm Higher Blch High Raw High DryR High Sliver High Sliced High
6
Component 2 (27%)
S: Sensory & Food Quality
Comparison of descriptive analysis and consumer liking data Data analysis. We conducted a PCA with Varimax rotation (XLSTAT) using the mean values for each product for each statistically significant attribute from the descriptive analysis data, but using only the almonds evaluated in the consumer liking test. We added texture liking ratings from the consumers as a supplemental variable to determine the relationship between the textural attributes from the descriptive analysis data and the consumer liking data.
Table 3–Water activity and moisture contents of the 20 almonds.
8
Almond texture, water activity, and liking . . . Descriptive analysis Almond form
less hard to grind, had less moistness of mass, less cohesiveness of mass, less fatty film, fewer number of chews, and fewer number of swallows (Figure 1). Comparison of whole, sliced, and slivered almonds. Compared with Mean values of all texture attributes, moisture content, and the whole almonds, sliced and slivered almonds were less hard, less water activity for each product are available in Table S1 to S5. We crunchy, had less cohesiveness of mass, less particulate mass, fewer have used PCA biplots to summarize these data in this manuscript. fibers between teeth, less tooth pack, less residual particles, fewer number of chews, and fewer number of swallows (Figure 1). Com- Comparisons among whole almonds. Dry roasted natural almonds pared with slivered almonds, sliced almonds were more powdery, were generally more crisp, crunchy, and powdery, and had more had more macroroughness, more loose particles, and were more loose particles and more fibers between teeth than the natural crisp (Figure 1). Compared with slivered almonds, sliced almonds almonds, which, in turn, had more of these attributes than the were less moist, less hard to split, broke into fewer pieces, were blanched almonds (Figure 2). Blanched almonds were less hard,
Figure 2–Principal component biplot showing positions of almonds and texture attributes on the first 2 principal components from the principal components analysis (PCA) on the descriptive analysis data of only the whole almonds. Component 1(horizontal axis) was highly positively correlated with crunchiness and persistence of crunch; component 2 (vertical axis) was highly positively correlated with surface roughness and highly negatively correlated with moistness and fatty film. Measures of water activity and moisture content were added to the PCA as supplementary variables.
4
macroroughness powdery
Raw High
2
fibersbetweenteeth DryR
High
Component 2 (29%)
Raw Norm numberofchews
DryR
Raw LL DryR LL crunchiness DryR Norm persistenceofcrunch
wateracvity 0
looseparcles residualparcles crispness Raw Lowhardnesstogrind hardnesstosplit Low
numberofpieces
moisturecontent
water cohesivenessofmass
More dry
moistnessofmass
Blch LL
Blch Norm
-2
Blch High
fayfilm moistness
Blch LL
-4
-6 -10
-8
-6
-4
-2
0
2
4
6
8
Component1 (36 %)
Figure 3–Principal component biplot showing positions of almonds and texture attributes on principal components 3 and 4 from the PCA done on the descriptive analysis data of only the whole almonds. Component 3 (horizontal axis) was positively related to toothpack and to fibers between teeth, and negatively related to number of chews; component 4 (vertical axis) was positively related to number of swallows and presence of particles. Measures of water activity and moisture content were added to the PCA as supplementary variables.
3
DryR Norm numberofswallows 2
Component 4 (9%)
looseparƟcles parƟculatemass
Blch High DryR high
powdery numberofpieces cohesivenessofmass DryR Low
1
DryR LL
Raw LL
fibersbetweenteeth
0
wateracƟvity Blch Low moistnessofmass Blch Norm toothpack moisturecontent
numberofchews -1
Raw Low
Raw High
Raw Norm Blch LL
-2 -4
-3
-2
-1
0
1
2
3
4
Component 3 (11 %) Vol. 79, Nr. 7, 2014 r Journal of Food Science S1403
S: Sensory & Food Quality
Blanched
Almond texture, water activity, and liking . . . Table 4–Mean ratings from consumer test (over all judges, N = 113). All ratings were made on 20 point scales where the left end was labeled “none” and the right end “intense.” All attributes differed significantly among the 8 almonds. (All P < 0.0001.) Numbers in italics below the means are standard errors of the means. Attribute Overall liking Texture liking Flavor liking Hardness Crispness Crunchiness Toothpack a bc
Raw LL
Raw Low
Raw High
DryR Low
DryR Norm
DryR High
Blched LL
Blched High
11.5b 0.39 11.9b 0.38 11.4b 0.42 14.4a 0.31 12.7ab 0.38 13.2a 0.33 10.7a 0.36
11.3b 0.43 11.5ba 0.41 11b 0.43 14.1a 0.32 12a 0.4 12.4a 0.4 10.7a 0.36
6.6c 0.42 5.4d 0.41 8.6c 0.46 6d 0.33 4.5c 0.31 4.1c 0.29 10.7a 0.45
13.4a 0.42 14a 0.34 13a 0.44 12b 0.37 13.3a 0.33 13.4a 0.32 9.8a 0.38
13.4a 0.43 14a 0.38 12.8a 0.45 11.8b 0.34 13.1a 0.37 13.1a 0.37 9.5a 0.35
7.4c 0.38 7.8c 0.39 7.5d 0.42 9.9c 0.4 7.7b 0.41 7.2b 0.39 10.6a 0.39
12.1b 0.36 13.1a 0.37 11.3b 0.4 11.3b 0.37 12.5a 0.36 12.5a 0.35 9.9a 0.39
4.5d 0.39 3.8e 0.38 5.9e 0.42 4e 0.32 2.5d 0.23 2.4d 0.26 8.2b 0.47
Means in a row with letter superscripts in common were not significantly different (P > 0.05).
broke into fewer pieces, and had less persistent crunch than the dry roasted or the natural almonds. Almonds with higher moisture content generally had more moistness and cohesiveness of mass and were less crisp, hard, and crunchy (Figure 2). Both water activity and moisture content were highly negatively correlated with component 1 (−0.92 and −0.83 respectively). Component 3 was positively related to toothpack and to fibers between teeth, and negatively related to number of chews; component 4 was positively related to number of swallows and presence of particles (Figure 3).
Figure 4–Biplot showing positions of almonds and texture attributes on the first 2 principal components from the PCA on the descriptive analysis data of only the whole almonds used in the consumer test. Component 1 was highly positively correlated with crunchiness and persistence of crunch; component 2 was highly positively correlated with number of chews and highly negatively correlated with fibers between teeth. Texture liking ratings from the consumer panelists were added to the PCA as supplementary variables.
4
numberofchews
2
Raw LL Raw Low
DryR High Component 2 (13%)
S: Sensory & Food Quality
Consumer Liking Study Liking ratings. Dry Roasted Norm and Dry Roasted Low were the best liked almonds, followed by Blanch LL, Raw LL, and Raw Low. Blanch High was the least liked almond (Table 4). Ratings of overall liking, texture liking, and flavor liking all followed the same pattern. Discrimination among the samples was most distinct (largest F value from the ANOVA) for the texture
liking ratings, so we included them as supplemental variables in the PCA for the consumer test data. PCA on products evaluated in the consumer test. Principal component 1 was highly positively correlated with crunchiness, persistence of crunch, and crispness (Figure 4). Blanch LL scored high on this component. Blanch High and Raw High scored low. Principal component 2 was highly positively correlated with number of chews. Raw LL and Raw Low scored high on this component; Dry Roasted Low and Raw High scored low (Figure 4). Principal component 3 was highly positively correlated with macroroughness and negatively correlated with moistness and fatty film. Raw Low scored high on this component; Blanch High and Blanch LL scored low. Principal component 4 was highly correlated with number of swallows and number of particles. Dry Roasted Norm scored high on this component; Blanch LL scored low on this component.
Blch High 0
hardnesstogrind hardnesstosplit residualparƟcles parƟculatemass crunchiness persistenceofcrunch
cohesivenessofmass faƩyfilm
Texture Liking
powdery crispness Blch LL numberofpieces looseparƟcles toothpack
DryR Norm
fibersbetweenteeth
Raw High
-2
DryR Low -4
-8
-6
-4
-2
0
Component 1 (37 %) S1404 Journal of Food Science r Vol. 79, Nr. 7, 2014
2
4
6
8
Almond texture, water activity, and liking . . .
10
Raw LL
Blch LL DryR Low
9
DryR LL
8
DryR Norm
Raw Low
Blch Low
Crunchiness
7 6
DryR High
5
Blch Norm
Raw HIgh
Raw Norm
4 3
Blch High
2 1 0 0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Water acƟvity
Texture liking ratings were highly positively correlated with principal component 1 and with the attributes crispness, crunchiness, persistence of crunch, number of pieces (all factor loadings > 0.91). Texture liking ratings were unrelated to principal components 2 to 4.
Discussion Although correlations between crunchiness and moisture and between crunchiness and water activity were relatively high, they appear to be due to a grouping of higher moisture almonds with lower crunchiness and a 2nd grouping of lower moisture almonds with higher crunchiness (Figure 5). The sounds produced by a food during chewing are important for crispness/crunchiness, and the loudness of those sounds can be diminished by increasing the water activity of dry foods (Vickers and Bourne 1976a,b; Lewicki and others 2007). Relationships between water activity and sensory crispness/crunchiness or between water activity and instrumental correlates of crispness/crunchiness show that increases in water activity are commonly accompanied by decreases in these attributes. Our observation that further decreasing the moisture level below about 0.04 (or a water activity less than 0.4) did not appear to increase crunchiness is consistent with work by others indicating that water activities above 0.3 to 0.4 can be detrimental to crispness/crunchiness or its indicators (Katz and Labuza 1981; Lewicki and others 2007).
crisper. Compared with slivered almonds, sliced almonds were less hard, broke into fewer pieces, had less moistness and cohesiveness of mass, less fatty film, and required fewer chews and fewer swallows to consume. Dry roasted almonds were generally harder, more crisp, more crunchy, and produced more loose particles than natural almonds, which were, in turn, more hard, crisp, and crunchy than blanched almonds. As moisture content increased, moistness of mass and cohesiveness of mass increased. Crispness, number of pieces, hardness, crunchiness, persistence of crunch, and particulate mass decreased with increasing moisture content. Consumer texture liking ratings were highly positively correlated with the attributes crispness, crunchiness, and persistence of crunch.
Acknowledgments Funding for this research was provided by the Almond Board of California.
Author Contributions
GH conceived the objectives of the study, provided the different almond forms, and reviewed the manuscript. ZV designed the sensory testing part of the study, completed the data analysis, and drafted the manuscript. TL adjusted the moisture content of the almonds, measured moisture content and water activity, and reviewed the manuscript. AP provided the descriptive analysis training and conducted both the descriptive analysis and the Conclusion Compared with the whole almonds, sliced and slivered almonds consumer sensory tests. had less hardness, less crunchiness, less cohesiveness, less tooth packing, and required fewer chews and swallows to consume. References GV, Lapsley K, Huang G, Yada S, Seltsam J. 2010. Development of an almond Compared with slivered almonds, sliced almonds were more pow- Civille lexicon to assess the sensory properties of almond varieties. J Sens Stud 25(1):146–62. dery, had more surface roughness, more loose particles, and were doi:10.1111/j.1745-459X.2009.00261.x Vol. 79, Nr. 7, 2014 r Journal of Food Science S1405
S: Sensory & Food Quality
Figure 5–Water activity compared with crunchiness. Points represent the mean sensory scores for crunchiness of whole almonds at specific water activities. Crunchiness ratings were made on a 20 point scale where 0 = none and 20 = intense. Abbreviations by each point note the almond and treatment that point represents. The red oval encompasses the products with similarly high levels of crunchiness; the blue oval encompasses the products with “similarly” low levels of crunchiness.
Almond texture, water activity, and liking . . . Gou P, Diaz I, Guerrero L, Valero A, Arnau J, Romero A. 2000. Physico-chemical and sensory property changes in almonds of desmayo largueta variety during toasting / Cambios en las propiedades f´ısico-qu´ımicas y sensoriales de almendras de la variedad Desmayo Largueta durante el tostado. Food Sci Technol Intl 6(1):1–7. doi:10.1177/ 108201320000600101 Guerrero L, Gou P, Arnau J. 1997. Descriptive analysis of toasted almonds: a comparison between expert and semi-trained assessors. J Sens Stud 12(1):39–54. doi:10.1111/j.1745459X.1997.tb00052.x Hyson DA, Schneeman BO, Davis PA. 2002. Almonds and almond oil have similar effects on plasma lipids and LDL oxidation in healthy men and women. J Nutr 132(4):703– 7. Jambazian PR, Haddad E, Rajaram S, Tanzman J, Sabat´e J. 2005. Almonds in the diet simultaneously improve plasma alpha-tocopherol concentrations and reduce plasma lipids. J Am Diet Assoc 105(3):449–54. Katz EE, Labuza TP. 1981. Effecty of water activity on the sensory crispness and mechanical deformation of snack food products. J Food Sci 46:403–9. Kennedy J, Wright A. 2008. Storage study of various almond forms. Technical Report Natick/TR-09/011L 1-14. Lewicki P, Marzec A, Kuropatwa M. 2007. Influence of water activity on texture of corn flakes. Acta Agrophys 9(1):79–90. Milbury PE, Chen C, Dolnikowski G, Blumberg J. 2006. Determination of flavonoids and phenolics and their distribution in almonds. J Agric Food Chem 54:5027–33. doi: 10.1021/jf0603937 PanelCheck 2014. Available from: http://www.panelcheck.com. Accessed 2014 January 10. Saklar S, Ungan S, Katnas S. 1999. Instrumental crispness and crunchiness of roasted hazelnuts and correlations with sensory assessment. J Food Sci 64(6):1015–9. Spiller GA, Jenkins DJ, Cragen LN, Gates JE, Bosello O, Berra K, Rudd C, Stevenson M, Superko R. 1992. Effect of a diet high in monounsaturated fat from almonds on plasma cholesterol and lipoproteins. J Am Coll Nutr 11(2):126–30. Spiller GA, Jenkins DA, Bosello O, Gates JE, Cragen LN, Bruce B. 1998. Nuts and plasma lipids: an almond-based diet lowers LDL-C while preserving HDL-C. J Am Coll Nutr 17(3):285–90. Varela P, Chen J, Fiszman S, Povey MJW. 2006. Crispness assessment of roasted almonds by an integrated approach to texture description: texture, acoustics, sensory and
S: Sensory & Food Quality S1406 Journal of Food Science r Vol. 79, Nr. 7, 2014
structure. J Chemometrics 20(67):311–20. doi:http://dx.doi.org/10.1016/j.foodres.2008. 03.009 Vickers Z, Bourne MC. 1976a. Crispness: a review. J Food Sci 41(5):1153–7. Vickers Z, Bourne MC. 1976b. A psychoacoustical theory of crispness. J Food Sci 41(5):1158– 64.
Supporting Information Disclaimer: Supplementary materials have been peer-reviewed but not copyedited. Table S1 and S2, containing mean values and standard errors for all the sensory attributes of all the products, should be made available online. Table S3. Mean sensory texture attribute values (over all judges, N = 13) slivered almonds only plus F and P values. Table S4. Mean sensory texture attribute values (over all judges, N = 13) sliced almonds only plus F and P values. Table S5. Mean sensory texture attribute values (over all judges, N = 13) by almond type (collapsed over moisture contents) plus F and P values.
