Upload
gladys-gonzalez-gonzalez
View
216
Download
0
Embed Size (px)
Citation preview
8/18/2019 Do Nascimento, 2015
1/12
Correlations
between
subjective quality
and
physicochemicalattributes
of
fresh
fruits
and
vegetables
M. Cecilia do Nascimento Nunes*
University of South Florida, Department of Cell Biology, Microbiology and Molecular Biology, Tampa, FL 33602, USA
A
R
T
I
C
L
E
I
N
F
O
Article history:
Received
9
November
2014
Received in revised form 30 April 2015Accepted 3 May 2015
Available
online
19
May 2015
Keywords:
Quality rating charts
Color
Texture
Weight loss
Anthocyanins
Chlorophyll
A
B
S
T
R
A
C
T
Color charts and rating scales have been developed for several fresh fruits and vegetables (FFVs) but
limited information is available regarding the correlation between subjective evaluations and
physicochemical attributes. The objective of this work was to correlate subjective quality data with
quantitative analytical data collected for several fruits exposed to different environmental conditions.
Avocados, blueberries, peppers, strawberries and tomatoes were exposed to a range of different
temperaturesandhumidity conditions forvariedperiods of time, andquality evaluated using both rating
scales and physicochemical analysis. The strength of the relationship between variables wasmeasured
using thePearson correlation coef cient (r ) and the coef cient of determination(r 2) and, thesignicance
of the relationship was expressed by probability levels ( p=0.05). Overall, there was a signicant
correlationbetweenmostof the subjectivequality attributesevaluatedand the physicochemicalanalysis
performed. Subjectivecolorwas signicantlycorrelated with hue angle for all fruits evaluated except for
blueberries for which subjective color had a stronger correlation with L* values. Correlations between
subjective color and anthocyanins, ascorbic acid or chlorophyll contents were also signicant. Shriveling
or stem freshnesswas strongly correlatedwithweightlosswhereas subjectivermnesswassignicantly
correlated with instrumental texture. Results fromthis work showed that subjective quality evaluations
using ratingscalescan be a reliableandsimplemethodto estimate changes in color, softening, waterloss,
and ultimately changes in specic chemical components when FFVs are exposed to different
environmental conditions. A color chart is proposed for the visual evaluation of strawberry quality.ã 2015 Elsevier B.V. All rights reserved.
1. Introduction
Visual appearance of fresh fruits and vegetables (FFVs) has the
greatest impact on retailers buying decisions and on consumer
choices and purchases. Attributes such as appearance, freshness
and color are considered the foremost criteria used to evaluate the
immediate quality of FFVs (Clydesdale, 1991; Mitcham et al., 1996;
Barrett et al., 2010). As a result, they are used as quality indicators
throughout the supply chain, from the farm to the consumer, and
ultimately determine product acceptance or rejection. Texture,taste and aroma are also important sensory attributes but are
mostly related with subsequent purchases (Clydesdale, 1991;
Barrett et al., 2010). Nutrient content is not visible or touchable,
and therefore is often disregarded as a quality attribute when it
comes to food choices and purchase decisions. However, FFVs are
important contributors to a well-balanced diet and to human
wellbeing as they supply important macronutrients, such as
carbohydrates and ber, and micronutrients such as vitamins and
minerals as well as polyphenols.
Subjective quality evaluations are often used to rate the
appearance, texture and avor of FFVs and, unlike formal sensory
panels, these are usually performed by trained individuals but not
on a sensory panel setting (e.g., quality control, and to estimate
ripeness stage and maturity at harvest). Although these evalua-
tions are criticized by some as being inexact, in the absence of
formal analytical or affective sensory panels they are valuable toquality control, and to determine the ripeness stage or the end of
shelf life of FFVs. In addition, they are faster, easier and less
expensive than sensory panels or instrumental measurements
which may require extensive training and complex logistics, or
expensive equipment (Mitcham et al., 1996; Barrett et al., 2010).
Consequently, many researchers frequently use somewhat quanti-
tative scoring systems either alone or combined with drawings or
photographs to evaluate the sensory quality of FFVs. For example,
Kader and Cantwell (2006) developed several color charts along
with rating scales and descriptors for physical damage of produce.* Tel.: +1 813 974 9307; fax: +1 813 905 9919.
E-mail address: [email protected] (M. C.d.N. Nunes).
http://dx.doi.org/10.1016/j.postharvbio.2015.05.001
0925-5214/ã 2015 Elsevier B.V. All rights reserved.
Postharvest Biology and Technology 107 (2015) 43–54
Contents
lists
available
at
ScienceDirect
Postharvest
Biology
and
Technology
j ou rna l homepage : www.elsev ier .com/locate/postharvbio
mailto:[email protected]:[email protected]://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://www.sciencedirect.com/science/journal/09255214http://www.elsevier.com/locate/postharvbiohttp://www.elsevier.com/locate/postharvbiohttp://www.elsevier.com/locate/postharvbiohttp://www.elsevier.com/locate/postharvbiohttp://www.sciencedirect.com/science/journal/09255214http://dx.doi.org/10.1016/j.postharvbio.2015.05.001http://dx.doi.org/10.1016/j.postharvbio.2015.05.001mailto:[email protected]://crossmark.crossref.org/dialog/?doi=10.1016/j.postharvbio.2015.05.001&domain=pdf
8/18/2019 Do Nascimento, 2015
2/12
White et al. (2005) developed a color chart to describe skin color
and avocado ripening. Our group also have developed a system that
uses scores and descriptors to rate individual sensory quality
attributes of various fruits and vegetables (Laurin et al., 2003;
Nunes et al., 2003a,b,c; Nunes et al., 2004, 2006, 2007, 2011, 2012,
2013; Proulx et al., 2005; Nunes and Emond, 2007; Nunes, 2008;
Proulx et al., 2010; Chilson et al., 2011). In the produce industry,
color charts are also frequently used to assess the stage of ripeness
(e.g., banana, tomato and avocado color charts) or to grade (e.g.,
size, color, and defects) or even to decide if a load of produce should
be accepted or rejected based on visual inspection.
Many color charts and rating scales have been developed for
several types of produce, and various studies have shown that
correlations exist between sensory and physical and/or chemical
attributes of FFVs (Ressureccion and Shewfelt, 1985; Maul et al.,
2000,b; Harker et al., 2002a,b; Safner et al., 2008; Gunness et al.,
2009; Pace et al., 2011; Corollaro et al., 2014). For example,
rmness and color of tomatoes were highly correlated with
sensory attributes (Ressureccion and Shewfelt, 1985) whereas
perceived sweetness or sourness was correlated with specic
volatile compounds (Maul et al., 2000). Pace et al., (2011) reported
a signicant correlation between b*, chroma, pH, titratable acidity
and appearance of fresh-cut nectarines. In apples, titratable acidity
was suggested to be a good predictor of acid taste (Harker et al.,2002a,b) and texture analysis correlated well with sensory
perception of apple texture (Harker et al., 2002a,b; Corollaro
et al., 2014). However, these studies used consumer or trained
sensory panels and to our knowledge there is a lack of published
studies showing that subjective quality data collected by trained
individuals (not in a sensory panel setting) can also be, in the
absence of formal trained sensory panel, a reliable way the
determined changes in the overall quality of FFVs. The objectives of
this work were to: (1) correlate subjective quality data, such as
color, rmness and shriveling, with quantitative analytical data
collected for different FFVs, and to show that in the absence of a
formal sensory panel the use of color charts and rating scales can
be used by trained individuals as an accurate way of determining
changes in overall quality of FFVs; and (2) give an example of aunique color chart designed for the evaluation of visual quality of
strawberry based on correlations between individual subjective
quality characteristics and physicochemical attributes, and that
was validated in research and commercial settings.
2. Material and methods
2.1.
Sampling
Fresh fruits and vegetables were harvested twice from
commercial operations in Florida and transported to the laboratory
within minimal delay after harvest (i.e., 1 to 6 h, depending on the
distance between the eld and the laboratory) (Tables 1 and 2). All
FFVs were harvested at the commercial maturity stage, and cluster
tomatoes were harvested from a greenhouse at the light-red stage.
Upon arrival to the laboratory, FFVs were visually selected for
uniformity of color/ripeness stage, size and freedom of defects.
Sample sizes were chosen based on the size and variability
within each commodity (i.e., the smaller the size of the fruit the
larger the number of fruits per replicates). Thus, three avocados,
three peppers, two clusters of three tomatoes each, and three
replicated samples of 15 or 20 strawberries and blueberries each,
respectively, were used for initial subjective quality evaluation, and
for instrumental color and texture analysis, and immediately
frozen to be later used for chemical compositional analysis. A total
of 20 avocados or 20 peppers (three fruit per RH), and 15 clusters of
three tomatoes each (3 clusters per RH), and a total of 15 clamshells
(3 clamshells per RH) containing 15 or 20 strawberries or
blueberries, respectively, were distributed among the ve RH-
controlled rooms and reused daily or every two days for non-
destructive quality evaluations (i.e., subjective quality evaluations
and weight loss). For destructive quality evaluations (i.e., texture
analysis and chemical analysis) and for non-destructive evalua-
tions that required manipulation of the fruit to an extent that could
cause minor bruising (i.e., instrumental color) 165 avocados or
peppers (33 fruits per RH), and 110clusters of three tomatoes each(22 clusters per RH), and 120 or 135 clamshells (24 clamshells per
RH; 27 clamshells per RH, respectively) containing 20 or 15 blue-
berries and strawberries each, respectively were distributed
among the ve RH-controlled rooms. However, every day three
clamshells of these strawberries or every two days, three of these
avocados or peppers, and two clusters or tomatoes, and three
clamshells of these blueberries were removed from their respec-
tive RH and immediately frozen, to be later used for chemical
compositional analysis. Avocados were stored for 20 d, blueberries
were stored for 16 d, and peppers and tomatoes were stored for 22
days and quality evaluated every two days. Strawberries were
stored for nine days and quality evaluated every day (Table 2). For
temperature treatments the experimental setup was similar to that
used for RH treatments except that only avocados, strawberriesand tomatoes were used in this part of the experiments (Table 1).
Since all non-destructive quality analysis (i.e., subjective quality
evaluations and weight) were assessed using always the same FFVs
samples, those were conducted within approximately 30 min after
the products were removed from storage, to minimize temperature
uctuations that could affect the quality.
2.2.
Storage
conditions
Storage conditions (i.e., temperature and relative humidity
treatments) and experimental setup used in this study were
similar to those previously described in detail by Nunes et al.
Table 1
Optimum storage conditions, cultivar, harvest location and date, and storage conditions during the temperature experiments.
Commodity Optimum temperature (C) Optimum RH (%) Cultivar Origin Harvest date Storage duration (d)a
Avocados 5–12b 85–95b ‘Choquette’ Homestead, Florida October 1, 2008 22
‘Choquette’ Homestead, Florida November 19, 2009 22
Strawberries 0c 90–95c ‘Albion’ Floral City, Florida December 12, 2008 10
‘Albion’ Floral City, Florida March 9, 2009 10
Tomatoes 7–13d 90–95d ‘Success’ Wellborn, Florida January 15, 2009 22
‘Success’ Welborn, Florida February 24, 2008 22
Storage conditions: (A) 1.8 0.8 C; (B) 5.2 0.2 C; (C) 10.6 0.6 C; (D) 15.2 0.4 C; (E) 20.2 0.2 C; 90% RH in all ve temperature-RH controlled chambers.a In some cases the experiments were terminated before the end of the storage period, at the time when at least one of the visual quality attributes evaluated reached the
maximum acceptable (rating of 3).b Woolf et al. (2014); evaluated every two days.c Mitcham (2014); evaluated every day.d Sargent
and
Moretti
(2014); light
red
greenhouse-grown
tomatoes;
evaluated
every
two
days.
44 M.C.N. Nunes / Postharvest Biology and Technology 107 (2015) 43–54
8/18/2019 Do Nascimento, 2015
3/12
(2013). After sorting, FFVs designated for temperature experiments
(avocados, strawberries, tomatoes) were distributed into ve
temperature-humidity-controlled chambers (Forma Environmen-tal Chambers Model 3940 Series, Thermo Electron Corporation,
OH, USA) that were set at ve different temperatures with an
relative humidity (RH) of approximately 90% (Table 1). These
temperatures were chosen because they cover a wide range of
physiological temperatures, between optimum and ambient
temperature, with an optimum RH. For the humidity experiments,
temperature of the chambers containing blueberries and straw-
berries was maintained as close as possible to the optimum storage
temperature (1.5 C) whereas for avocados, peppers and tomatoes,
the temperature of the chambers was maintained at 15C in order
to reduce the risk of chilling injury (Table 2). The levels of humidity
used were chosen based on RH that may be encountered
throughout the supply chain (Nunes et al., 2009; Lai et al., 2011
;Pelletier et al., 2011).
2.3.
Temperature
and
relative
humidity
monitoring
The temperature inside the temperature and RH controlled
rooms was monitored throughout the study using Stow Away1
XTI02 temperature loggers from (5 C to +37 C) (Onset Computer
Corporation, Pocasset, MA). The RH was monitored with Stow
Away1 RH loggers (10 to 95% RH) (Onset Computer Corporation,
Pocasset, MA).
2.4. Subjective quality evaluation
Subjective quality evaluation of strawberries was performedevery day during nine to10 d, depending on the storage temperature
and RH (Tables 1 and 2). Subjective quality evaluations for avocados,
blueberries, peppers and tomatoes were performed every two days
with the length of storage depending on the commodity and storage
regime (Tables 1 and 2). A trained individual(s) conducted the
subjective quality evaluations using rating scales and descriptors
(Tables 3–7). Color, shriveling and stem freshness were determined
subjectively using a 1 to 5 visual rating scale and rmness was
determined based on the whole fruit resistance to slightly applied
nger pressure and recorded using a 1–5 tactile rating A score of
3 was considered the limit of acceptability for sale. Thus, when fruit
were visibly deteriorated or when at least one of the subjective
quality attributes had attained a ratingof 3 or lower, the treatments
were terminated. Some of the visual quality scores and descriptorsused (e.g., shriveling and cluster tomato quality scores and
descriptors; Table 7) weredeveloped byour team based on multiple
visual observations during postharvest storage of FFVs at different
temperatures (Nunes, 2008) whereas other scores were used based
on previously published works (Tables 3–6; see footnotes). The
individual(s) that performed the visual evaluations were trained
(using FFVs and photographs) to detect minimal changes in the
appearance and rmness of the FFVs and had extensive experience
using the rating scales.
Table 2
Optimum storage conditions, cultivar, harvest location and date, and storage conditions during the humidity (RH) experiments.
Commodity Optimum temperature (C) Optimum RH (%) Cultivar Origin Harvest date Storage duration (d)a
Avocados
5–12b 85–95b ‘Simmonds’
Homestead,
Florida
July
17,
2008
20
‘Simmonds’ Homestead, Florida August 2, 2008 20
Blueberries 0.5 to 0c >90c ‘ Jubilee’ Winter Haven, Florida April 5, 2012 16
‘ Jubilee’ Winter Haven, Florida April 26, 2012 16
Pepper 7–10d 90–95d ‘Revolution’ Immokalee, Florida November 24, 2009 22
‘Revolution’ Immokalee, Florida December 1, 2009 22
Strawberries 0e 90–95e ‘Strawberry Festival’ Floral City, Florida March 5, 2012 9
‘Strawberry Festival’ Floral City, Florida February 21, 2013 9
Tomato 7–13f 90–95f ‘Success’ Wellborn, Florida June 11, 2009 22
‘Success’ Wellborn, Florida June 27, 2009 22
Storage conditions: (A) 40.1% 3.2% RH; (B) 62.0% 2.0% RH; (C) 81.0% 2.1% RH; (D) 87.9% 1.9% RH; (E) 91.6 1.4% RH; temperature of the chambers was set at 15C for
avocado,
pepper
and
tomato
and
1.5C
for
blueberry
and
strawberry.a In some cases the experiments were terminated before the end of the storage period, at the time when at least one of the visual quality attributes evaluated reached the
maximum acceptable (rating of 3).b Woolf et al. (2014); evaluated every two days.c Perkins-Veazie (2014); evaluated every two days.d González-Aguilar (2014); evaluated every two days.e Mitcham (2014); evaluated every day.f Sargent and Moretti (2014); evaluated every two days.
Table 3
Visual quality scores and descriptors for avocado.
Scores
and
description
1
2
3a 4
5
Very poor Poor Acceptable Good Excellent
Colorb Very dark skin with no traces of
green;
fruit
appears
completely
brown or black
Some green on brown or
black;
approximately
75% colored
Some yellow/black or brown
on
green;
approximately
25% colored
Darker lime green and not so
glossy;
beginning
of
color
changes;
some yellowing
Full dark lime green
color;
very
glossy;
freshly harvested
Firmnessb Extremely
soft
on
touch;
easily
yields to slight hand pressure
Soft;
whole
fruit
deforms
with slight hand pressure
Elastic;
slightly
yield
to
hand pressure
Hard
but
less
resistant
to
extreme
nger pressure
Extremely
hard;
does
not yield to extreme
nger pressure
Shriveling Skin appears extremely dry Serious shriveling Shriveling evident, but not
serious
Minor signs of shriveling Field-fresh, no signs of
shriveling
a Score of 3 was considered to be the minimum acceptable quality before avocados become unmarketable.b Modied from White et al. (2005).
M.C.N. Nunes / Postharvest Biology and Technology 107 (2015) 43–54 45
8/18/2019 Do Nascimento, 2015
4/12
2.5. Instrumental surface color
A total of two color measurements per FFVs were taken on
opposite sides at the equatorial region. A hand-held tristimulus
reectance colorimeter (Model CR-300, Minolta Co., Ltd., Osaka,
Japan) equipped with a glass light-protection tube (CR-A33,
Minolta Co., Ltd., Osaka, Japan) was used. Color was recorded
using the CIE-L*a*b* uniform color space (CIE-Lab), L* (lightness),
a* (redness) and b* (yellowness) values. Numerical values of a* and
b* were converted into hue angle and chroma using the Minolta
Color Management Software (1996–1999CyberSoft SpectraMatch/
QC software version 3.3, CyberChrome, Inc., Stone Ridge, N.Y.).
2.6. Texture analysis
Before texture was analyzed, FFVs from each treatment were
conditioned at room temperature for approximately 1 h. Thereaf-
ter, rmness of each individual FFVs was measured using a TA.XT
plus Texture Analyzer (Texture Technologies Corp., NY, USA)
equipped with a 50 kg load cell.
Table 4
Visual quality scores and descriptors for blueberry.
Scores and description
1 2 3a 4 5
Very poor Poor Acceptable Good Excellent
Colorb Extremely dark;
overripe or
senescent
Very dark blue/purplish Fully dark blue More blue, less
bright
Bright blue color
Firmnessc Berry rupture ontouch
Berry surface very depressed on touchbut
no
ruptureBerry surface depressed on touch,softer
than rmer
Slight depressionon
touchFirm berry, not yielding to touch
Shriveling Extremely wilted
and dry
Severe shriveling Shriveling evident but not serious Slight signs of
shriveling
Field-fresh, fruit appear very
fresh and turgid
a Score of 3 was considered to be the minimum acceptable quality before blueberries become unmarketable.b Modied from Sanford et al. (1991).c Modied from Beaudry et al. (1998).
Table 5
Visual quality scores and descriptors for pepper.
Scores and description
1
2
3a 4
5
Very poor Poor Acceptable Good Excellent
Colorb Extremely dull
green or 25%
colored
Dull green or slight coloration Green but showing loss of glossiness Less bright dark green Completely bright
dark green; very
glossy
Firmnessb Extremely soft on
touch
Soft on touch, yields easily to nger
pressure; not crisp; 75% of the fruit is
soft
Fruit starts to soften and yield to
nger pressure; 50% of the fruit is
soft
Firm but less resistant to nger
pressure; 25% of the fruit is soft
Very rm, turgid and
crisp
Shriveling Extremely
shriveled
and
dry
Serious shriveling Shriveling evident, but not serious Slight signs of shriveling Field-fresh, no signs
of
shriveling
a Score of 3 was considered to be the minimum acceptable quality before peppers become unmarketable.b Modied from Lownds et al. (1994).
Table 6
Visual quality scores and descriptors for strawberry.a
Scores and description
1 2 3b 4 5
Very poor Poor Acceptable Good Excellent
Color Very dark purplish-red;
extremely overripe or
senescent
Overripe; very dark
red
Fully red Fully light red Three-quarter to fully light red
Firmness Extremely soft and
deteriorated
Soft and leaky Minor signs of softness Firm but less turgid Very rm and turgid
Shriveling
Extremely
wilted
and
dry
Severe
shriveling,
fruit
is shriveled and
calyx is wilted and dry
Shriveling
evident,
fruit
and
calyx
show
evident signs of moisture loss
Minor
signs
of
shriveling,
calyx slightly wilted
Field-fresh,
fruit
and
calyx
appear very fresh and turgid
a Nunes et al. (2003c).b Score of 3 was considered to be the minimum acceptable quality before strawberries become unmarketable.
46 M.C.N. Nunes / Postharvest Biology and Technology 107 (2015) 43–54
8/18/2019 Do Nascimento, 2015
5/12
For avocado, the instrument was tted with a 76.2 mm diameter
stainless compression plate and rmness was measured on two
cheeks of each non-pealed fruit. The probe was then driven with a
crosshead speed of 2 mm s1, and the compression force was
recorded at 3.0 mm deformation ( Jeong et al., 2002). For blueberry,
three replicated samples of 30 g each were placed into three-
100 mL plastic beakers and the fruit was compressed to a depth of
30 mm using a 38 mm diameter and 20 mm high acrylic cylinder
probe. The probe was then driven with a crosshead speed of
1 mm s1, and the compression force was recorded after the probe
had compressed the fruit by 30 mm (Sanford et al., 1991). Peppers
were rst cut transversely, approximately three-quarter way from
the shoulder end and then placed, with the cut end down, on the
at
platform
of
the
texture
analyzer.
A
7.95
mm
stainless
convexprobe was centered on a shoulder lobe with two measurements
being taken on opposite lobes. The probe was then driven with a
crosshead speed of 1 mm s1, and the compression force was
recorded at 5.0 mm deformation (Nunes et al., 2012). Strawberryrmness was measured at the equatorial part of the fruit using a
7.95 mm stainless convex probe. The probe was then driven with a
crosshead speed of 1 mm s1, and the compression force was
recorded at 3.0 mm deformation (Whitaker et al., 2012). Tomato
was placed on the at surface of the texture analyzer with stem-
end down, so the pressure was applied on the blossom-end part of
the fruit. The instrument was tted with a 76.2 mm diameter
stainless compression plate and the probe was then driven with a
crosshead speed of 1 mm s1, and the compression force was
recorded
at
10.0
mm
deformation
(Chilson
et
al.,
2011).
2.7.
Weight
loss
Weight loss of each individual avocado, pepper or tomato and of
each individual triplicated clamshell containing 20 blueberries or
15 strawberries each was calculated from the initial weight and
after every day or every twodays during ve to 22 d, depending on
the fruit and on the temperature or RH regimes. Concentrations of
chemical constituents were expressed in terms of dry weight in
order to show the differences between temperatures or RHs that
might be obscured by differences in water content. The following
formula was used for water loss corrections: [chemical compo-
nents (fresh weight) 100 g/average dry weight (avocado = 12.9 g;
blueberry = 13.2 g; pepper = 6.1 g; strawberry = 9.4 g; tomato = 5.5
g) + weight loss during storage (g)]. The dry weight was deter-
mined by drying three weighed aliquots of homogenized fruit
tissue at 80 C, until weight stabilized.
2.8. Total anthocyanins content
Three replicated samples of blueberry or strawberry fruit were
homogenized in a laboratory blender (Waring Products Div.
Dynamics Corp. of America, New Hartford, CO) at high speed for
2 min. The homogenate (2 g) was mixed with 28 mL of 0.5% (v/v)
HCl in methanol and held for 1 h at 4 C for pigment extraction. The
occulate was removed by ltering the extract through a single
layer of facial tissue, and absorbance of the resulting liquid
containing
the
pigments
was
measured
at
520
nm
(maximumabsorbance for anthocyanins) using a BioTek microplate reader
(BioTek Instruments, Inc., Highland Park, Vermont, USA). Pigment
content was calculated using the following formula: A520
dilution factor [molecular weight (MW) of PGN/molar extinction
coef cient] where MW of PGN = 433.2 and the molar extinction
coef cient = 2.908 10,000. The amount of total anthocyanins was
expressed in terms of dry weight (g kg1) to compensate for water
loss during storage.
2.9. Total chlorophylls content
Chlorophyll content of avocado puree was extracted in the dark
with N ,N -dimethylformamide and absorbance of the ltrate
measured
at
625,
647
and
664nm
using
a
BioTek
microplatereader (BioTek Instruments, Inc., Highland Park, Vermont, USA).
Total chlorophylls content was calculated according to Moran
(1982) and the results expressed in terms of dry weight (g kg1) to
compensate for water loss during storage.
2.10.
Total
ascorbic
acid
Total ascorbic acid was quantied by mixing 2 g of strawberry,
pepper or tomato homogenates with 20 mL metaphosphoric acid
mixture (6% HPO3 containing 2 N acetic acid). Samples were then
ltered (0.22 mm) prior to HPLC analysis. Ascorbic acid analysis
was conducted using a Hitachi LaChromUltra UHPLC system with a
diode array detector and a LaChromUltra C18 2 mm column
(2 50 mm) (Hitachi, Ltd., Tokyo, Japan). The analysis was
Table 7
Visual quality scores and descriptors for greenhouse-grown cluster tomato.
Scores and description
1 2 3a 4 5
Very poor Poor Acceptable Good Excellent
Color Very dark red, overripe Dark red; dull color Vivid red; loss of glossiness Light red but less glossy Light red with no trace of
green; very glossy
Firmness Extra-soft, overripe, fruit yields
very readily to slight pressure
Soft, fruit yields readily
to slight pressure
Firm, fruit yields slightly to
moderate pressure
Hard, fruit yields only slightly
to considerable pressure
Extra hard, fruit does not yield
to considerable pressure
Shriveling Extremely shriveled and dry; fruit
appears old and deteriorated
Severe shriveling Shriveling evident, but not
serious
Slight signs of shriveling Field-fresh, no signs of
shriveling
Stem freshness Stem is completely dry
and dark brownish-
green
Stem is dry, wilted and
brownish-green
Signs of dryness are evident
but not objectionable
Stem appears slightly less
green and less turgid
Stem is
very
bright
green and
turgid
a Score of 3 was considered to be the minimum acceptable quality before cluster tomatoes become unmarketable.
M.C.N. Nunes / Postharvest Biology and Technology 107 (2015) 43–54 47
8/18/2019 Do Nascimento, 2015
6/12
performed under isocratic mode at a ow rate of 0.5 mL/min with a
detection of 254 nm. Sample injection volume was 5 mL, each with
duplicate HPLC injections. Mobile phase was buffered potassium
phosphate monobasic (KH2PO4, 0.5%, w/v) at pH 2.5 with
metaphosphoric acid (HPO3, 0.1%, w/v). The retention time of
the ascorbic acid peak was 2.5 min. After comparison of retention
time with the ascorbic acid standard, the peak was identied. The
amount of total ascorbic acid content in strawberry was quantied
using calibration curves obtained from different concentrations
(0.01 g L 1, 0.02 g L 1, 0.03 g L 1, 0.05 g L 1, 0.10g L 1, 0.15 g L 1,
0.20 g L 1 and 0.30 g L 1) of ascorbic acid standards. Total ascorbic
acid content was expressed in terms of dry weight (g kg1) to
compensate for water loss during storage.
2.11. Statistical analysis
The SigmaPlot Version 12.0 (Systat Sotware, San Jose CA) was
used for the analysis of the data. Since there were no signicant
differences between harvests in respect to the strength and
signicance of the relationships, data from the two harvests was
combined. The strength of the relationship between variables (i.e.,
subjective versus quantitative) was measured using the Pearson
correlation coef cient (r ). The coef cient of determination (r 2) and
the signicance of the relationship between variables was
expressed by probability levels ( p = 0.05). A linear regression
model was applied to the data collected and was used to describe
the relationship between variables.
3. Results and discussion
3.1. Correlation between subjective color and L *, hue angle,
anthocyanins, chlorophyll and ascorbic acid contents
Subjective color was signicantly correlated to L* values for all
the FFVs evaluated, regardless of the storage conditions, except for
strawberry stored under different RH regimes (Table 8; Fig. 1). In
general, as color ratings decreased, L* values also decreased,
meaning that the surface color of the FFVs became darker.
Similarly, with the exception of blueberry, there was a signicant
correlation between color ratings and hue angle for all the fruit
evaluated (Table 8; Fig. 1). As color ratings decreased, hue angle
values also decreased; for avocado and pepper, decreased in hue
angle corresponded to changes in color from bright dark lime green
to a brownish-green color (Tables 3 and 5; Fig. 1). In pepper,
decreased in hue angle corresponded to changes in color from
bright dark-green to dull green (Table 3).In strawberry and tomato,
decreased in hue angles corresponded to changes in the surface
color from bright light-red to a deep dull red (Tables 6 and 7; Fig.1).
Table 8
Pearson correlation coef cient (r ), coef cient of determination (r 2), linear regression equation, and signicance of the relationship ( p) between subjective color and:
instrumental color coordinates (L* value and hue angle), anthocyanins, chlorophyll and ascorbic acid content contents for avocado, blueberry, pepper, strawberry and tomato
stored at different temperature or humidity regimes.
Commodity (treatment) Color vs L* value Color vs hue angle Color
vs anthocyanins
(g kg1)
Color
vs chlorophyll
(g kg1)
Color vs ascorbic acid
(g kg1)
Avocado
(RH) r = 0.602 r = 0.906 NM p > 0.05 (NS) NM
r 2 = 0.363 r 2= 0.821
y = 4.409 x + 68.321 y =13.870 x + 56.882
p
8/18/2019 Do Nascimento, 2015
7/12
Nunes et al. (2013) also showed that there was a highly positive
correlation between banana visual color and hue angle values and
suggested that color can be a reliable way to estimate visual color
changes. Visual evaluation of browning in fresh-cut nectarines was
also found to be highly correlated to color parameters, particularly
with b* and chroma (Pace et al., 2011). In addition, color changes
have also been shown to be strongly correlated with sugar content
in banana (Nunes et al., 2013). Therefore, the use of color charts
(e.g., banana industry) when used by trained individual(s) can be
an easy and fast way to estimate FFVs quality attributes that are
associated with changes in surface color.
Color ratings were also signicantly correlated with anthocya-
nin contents, particularly in strawberry (Table 8; Fig. 2). As color
ratings decreased, anthocyanin content also decreased, meaning
that fruit with a deep purplish-blue or purplish-red color had
lower anthocyanin contents than fruit showing a bright light blue
or red color. Although in avocado the relationship between color
and chlorophyll content was not very strong, there was a
signicant correlation between color ratings and chlorophyll
content in avocado stored at different temperatures (Table 8).
Therefore, as color ratings decreased (color changed from green to
brown) chlorophyll content also decreased. Proulx et al. (2010) also
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
L * v a
l u e
30
35
40
45
50
55
60
65
70
Avocado (RH)
Tomato (RH)
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
5.0
H u e a n g
l e
60
80
100
120
140
160
Avocado (RH)
Avocado (T)
Pepper (RH)
Color rating (1-5)
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
H u e a n g
l e
0
20
40
60
80
100
120
Strawberry (RH)
Strawberry (T)
Fig. 1. Scatter plots and linear regression lines showing the relationship between
subjective color (1 = very poor; 5 = excellent) and L* value for avocado and cluster
tomato and the relationship between subjective color and hue angle for avocado,
pepper and strawberry stored at different temperature and RH regimes. RH =
relative humidity; T = temperature.
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
T o
t a l a n
t h o c y a n
i n s
( g k g
- 1 )
0
1
2
3
4
5
6
7
8
Strawberry (RH)
Strawberry (T)
Blueberry (RH)
Color rating (1-5)
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
T o
t a l a s c o
r b i c a c
i d ( g k g
- 1 )
0
2
4
6
8
10
12
14
Strawberry (RH)
Strawberry (T)
Pepper (RH)
Fig. 2. Scatter plots and linear regression lines showing the relationship between
subjective color (1 = very poor; 5 = excellent) and total anthocyanins content for
strawberry and blueberry and the relatioship between subjective color and total
ascorbic acid content for strawberry and pepper stored at different temperature and
RH
regimes.
RH
=
relative
humidity;
T
=
temperature.
M.C.N. Nunes / Postharvest Biology and Technology 107 (2015) 43–54 49
8/18/2019 Do Nascimento, 2015
8/12
showed that in snap beans there was a signicant correlation
between visual color scores and chlorophyll content, that is,
chlorophyll content decreased as the color turned from bright
green to yellowish-green. In avocado, the reason for the weak
relationship between visual color and chlorophyll content might
be related to the fact that color ratings were based on the color of
the avocado skin whereas chlorophyll content was measured in the
pulp of the fruit.
Ascorbic acid content was signicantly correlated with color
ratings particularly for pepper and strawberry (Table 8; Fig. 2).
Therefore, as color ratings decreased, ascorbic acid content also
decreased. After harvest, decline in ascorbic acid content of FFVs
usually occurs rapidly and it normally parallels the increase in
water loss, particularly when FFVs are exposed to adverse
environmental conditions. Excessive loss of water causes tissue
damage and results in deterioration of the overall quality; with cell
wall disruption promoting the release of ascorbate oxidase and
subsequent oxidation of ascorbic acid (Proulx et al., 2010).
3.2. Correlation between subjective shriveling and weight loss
For avocado, blueberry, pepper and strawberry, shriveling was
negatively correlated with weight loss whereas in tomato weight
loss was negatively correlated with shriveling and particularly
with stem freshness, regardless of the storage conditions (Table 9;
Figs. 3 and 4). Therefore, as ratings for shriveling or stem freshness
decreased weight loss signicantly increased. Nunes and Emond
(2007) evaluated several fruits and vegetables and also found a
highly signicant correlation between weight loss and subjective
color, rmness and shriveling/wilting. Thus, as weight loss
increased during storage, rmness decreased, and wilting,
shriveling or browning increased. Lownds et al. (1994) also
reported that accidity in peppers appeared to be directly
associated with water loss whereas color ratings paralleled
differences in water loss rates suggesting a direct relationship
between these variables.
3.3.
Correlation
between
subjective
rmness
and
instrumental
texture
Subjective rmness was signicantly correlated with instru-
mental texture, regardless of the fruit and the storage regimes used
(Table 9; Fig. 5). As subjective rmness ratings decreased, values
for quantitative texture also decreased. In a previous study, a
signicant correlation was found between banana subjective
rmness (1–5 rating scale) and analytical texture, suggesting that
rmness on touch can be used to estimate softening (Nunes et al.,
2013). White et al. (2005) produced softening curves for avocado
based on hand rmness scores versus quantitative texture
Table 9
Pearson correlation coef cient (r ), coef cient of determination (r 2), linear regression equation, and signicance of the relationship ( p) between
subjective shriveling or steam freshness and weight loss and between subjective rmness and instrumental texture for avocado, blueberry, pepper,
strawberry and tomato stored at different temperature or humidity regimes.
Commodity (treatment) Shriveling/stem freshnessavs weight loss (%) Firmness vs texture (N )
Avocado (RH) r = 0.938 r = 0.725
r 2= 0.880 r 2= 0.523
y= 3.807 x + 21.751 y = 28.002 x 48.454
p
8/18/2019 Do Nascimento, 2015
9/12
measured with a penetrometer. The curves showed that as hand
rmness decreases from very hard (rating of 0) to very soft (rating
of 7) the force required to puncture the fruit also decreased.
3.4. Data tting
A linear regression model was applied to all datasets, and thus
the following equation was used to summarize the relationship
between variables: [ y = ax + b]; where y = quantitative attribute and
x = subjective quality attribute score (1–5) (Tables 8 and 9). Graphic
representations were shown only for those relationships in which r
or r 2 0.6 and p
8/18/2019 Do Nascimento, 2015
10/12
Fig. 6. Color chart for strawberry showing photographs of visual quality deterioration with subjective quality ratings (1 = very poor; 3 = acceptable; 5 = excellent) and
correspondent descriptors.
52 M.C.N. Nunes / Postharvest Biology and Technology 107 (2015) 43–54
8/18/2019 Do Nascimento, 2015
11/12
http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0135http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0135http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0135http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0130http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0130http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0130http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0125http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0125http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0125http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0120http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0120http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0120http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0115http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0115http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0115http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0110http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0110http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0110http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0105http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0105http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0100http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0100http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0100http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0095http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0095http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0090http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0090http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0085http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0085http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0085http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0085http://ucce.ucdavis.edu/files/datastore/234-49.pdfhttp://ucce.ucdavis.edu/files/datastore/234-49.pdfhttp://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0075http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0075http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0075http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0070http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0070http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0065http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0065http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0060http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0060http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0060http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0055http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0055http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0055http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0050http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0050http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0050http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0040http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0040http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0040http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0035http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0035http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0035http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0030http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0030http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0030http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0030http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0025http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0025http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0025http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0025http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0020http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0015http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0015http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0015http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0010http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0010http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0010http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0005http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0005http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0005http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0005
8/18/2019 Do Nascimento, 2015
12/12
Nunes, M.C.N., Delgado, A., Emond, J.P., 2012. Quality curves for green bell pepper(Capsicum annum L.) stored at low and recommended relative humidity levels.Acta Hort. 945, 71–78.
Nunes, M.C.N., Yagiz, Y., Emond, J.P., 2013. Inuence of environmental conditions onthe quality attributes and shelf life of ‘Goldnger’ bananas. Postharvest Biol.Technol. 86, 309–320.
Nunes, M.C.N., 2008. Color Atlas of Postharvest Quality of Fruits and Vegetables.Blackwell Publishing, Iowa, USA.
Pace,
B.,
Cefola,
M., Renna,
F.,
Attolico,
G.,
2011.
Relationship
between
visualappearance and browning as evaluated by image analysis and chemical traits infresh-cut nectarines. Postharvest Biol. Technol. 61, 178–183.
Pelletier,
W.,
Brecht,
J.K.,
Nunes,
M.C.N.,
Emond,
J.P.,
2011.
Quality
of
strawberriesshipped by truck from California to Florida as inuenced by postharvesttemperature
management
practices.
HortTechnology
21
(4),
482–493.Perkins-Veazie, P., 2014. Blueberry. In: Gross, K.C., Wang, C.Y., Saltveit, M. (Eds.), The
Commercial Storage of Fruits, Vegetables, and Florist and Nursery Stocks.Agriculture Handbook 66. USDA-ARS http://www.ba.ars.usda.gov/hb66/blueberry.pdf (accessed 01.29.15.).
Proulx,
E.,
Nunes,
M.C.N.,
Emond,
J.P.,
Brecht,
J.K.,
2005.
Quality
attributes
limitingpapaya postharvest life at chilling and non-chilling temperatures. Proc. Fla.State Hort. Soc. 118, 389–395.
Proulx, E., Yagiz, Y., Nunes, M.C.N., Emond, J.P., 2010. Quality attributes limiting snapbean (Phaseolus vulgaris L.) postharvest life at chilling and non-chillingtemperatures. HortScience 45, 1238–1249.
Ressureccion, A.V.A., Shewfelt, R.L., 1985. Relationships between sensory attributesand objective measurements of postharvest quality of tomatoes. J. Food Sci. 50,1242–1245.
Safner, R., Polashock, J., Ehlenfeldt, M., Vinyard, B., 2008. Instrumental and sensoryquality characteristics of blueberry fruit from twelve cultivars. Postharvest Biol.Technol. 49, 19–26.
Sanford, K.A., Lidster, P.D., McRae, K.B., Jackson, E.D., Lawrence, R.A., Stark, R.,Prange, R.K., 1991. Lowbush blueberry quality changes in response tomechanical
damage
and
storage
temperature.
J.
Am. Soc.
Hort.
Sci.
116,
47–51.Sargent, S.A., Moretti, C.L., 2014. Tomato. In: Gross, K.C., Wang, C.Y., Saltveit, M.
(Eds.), The Commercial Storage of Fruits, Vegetables, and Florist and Nursery
Stocks.
Agriculture
Handbook
66.
USDA-ARS
http://www.ba.ars.usda.gov/hb66/tomato.pdf (accessed 01.29.15.).
Whitaker,
V.M.,
Chandler,
C.K.,
Santos,
B.M.,
Peres,
N.,
Nunes,
M.C.N.,
Plotto,
A.,
Sims,C.A., 2012. Florida MedallionTM (‘FL 05-1070) Strawberry. HortScience 47,296–298.
White, A., Woolf, A., Ofman, P., Arpaia, M.L., 2005. The International Avocado QualityManual, HortResearch, The Horticultural and Food Research Institute of NewZealand
Limited.
Auckland,
New
Zealand.Woolf, A.B., White, A., Arpaia, M.L., Gross, K.C., 2014. Avocado. In: Gross, K.C., Wang,
C.Y., Saltveit, M. (Eds.), The Commercial Storage of Fruits, Vegetables, and Floristand Nursery Stocks. Agriculture Handbook 66. USDA-ARS http://www.ba.ars.usda.gov/hb66/avocado.pdf (accessed 01.29.15.).
54 M.C.N. Nunes / Postharvest Biology and Technology 107 (2015) 43–54
http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0140http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0145http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0150http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0155http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0160http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0165http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0170http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0175http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0180http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0180http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0180http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0180http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0180http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0180http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0180http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0180http://refhub.elsevier.com/S0925-5214(15)30011-9/sbref0180http://refhub.elsevier.com/S