International Journal of Food Science and Technology 2008, 43, 357–364 357

Original article
Effect of preservatives addition on the shelf-life extensions and
quality of flat bread as determined by near-infrared spectroscopy
and texture analysis

Mahmoud Abu-Ghoush,1* Thomas J Herald,2 Floyd Dowell,3 Feng Xie,3 Fadi M Aramouni2 & Ronald Madl4
1 Nutrition and Dietetics, The Hashemite University, Zarqa 13133, Jordan
2 Food Science Institute, Kansas State University, Manhattan, KS 66506, USA
3 United States Department of Agriculture, Agriculture Research Service, Grain Quality Research, Manhattan, KS 66502, USA
4 Department of Grain Science and Industries, Manhattan, KS 66506, USA
(Received 22 June 2006; Accepted in revised form 08 March 2007)

Summary Many people do not have refrigeration or frozen storage to inhibit mould growth and keep the Arabic bread
fresh for more than a few days. Therefore, shelf-life extension is necessary for this type of bread. The present
study hypothesised that the addition of preservatives may be done in Arabic flat bread (AFB) to extend shelf-
life. Thus, objectives of this study were to evaluate selected preservatives to inhibit mould growth and to
employ physical techniques, to monitor bread aging. Three preservatives, fumaric acid (0.2%, F), sodium
propionate (0.3%, P), and sodium propionate-fumaric acid mixture (PF) were used. Tensile tests, and near-
infrared reflectance spectroscopy (NIRS) were used to monitor bread ageing. The addition of PF in the AFB
formula significantly increased the time of tearing at 0 day. For all treatments, the NIRS results showed high
R2-values between the actual storage days and NIRS predictions. The NIRS and texture analysis are
valuable tools to detect the effect of the preservatives on AFB shelf-life and quality.
Keywords Antimicrobial agents, flat bread, shelf-life, near-infrared spectroscopy, texture analysis.

preservatives (acetic, fumaric, sorbic and propionic

Introduction
acids or their potassium and calcium salts).
Flat bread is the oldest and most popular bread in the The AFB is considered to be at optimum quality
world (Qarooni, 1990). Over 1.8 billion consume flat during the first few hours after baking. The bread starts
breads daily (Qarooni, 1996). Flat breads contribute to lose strength and become firm after 24 h (Quail, 1996).
approximately 85% of the caloric value to the diets of These changes represent the initial stages of the staling
Middle Eastern populations (Mousa et al., 1979). Ara- process. Bread staling is a complex process that involves
bic flat bread (AFB) not unlike other breads has a a set of physicochemical changes (moisture distribution,
limited shelf-life. The two major considerations of shelf- firmness, tearing properties, and opacity) in the crumb,
life are spoilage and staling. Many countries lack either not those resulting from the action of microorganisms
appropriate packaging or storage facilities to extend the (Zobel & Kulp, 1996; Toufeili et al., 1998)
shelf-life and quality of the AFB. The economic loss Several methods have been used to specifically evalu-
contributed to spoilage and waste is over 1 billion ate the firmness of AFB. These methods include instron
dollars per year (Baik & Chinachoti, 2000). The water universal testing machine (Gurjal & Gaur, 2002),
activity of AFB ranges between 0.9 and 0.96 (Quail differential scanning calorimetry (Sidhu et al., 1997),
et al., 1990) which facilitates mold growth. The most and dynamic rheological testing (Toufeili et al., 1994).
prominent types of moulds associated with AFB are These are very diverse analytical methods that range
Aspergillus, Rhizopus and Penicillium families (Grundy, from destructive to nondestructive, each with their
1996). Quail (1996) suggested several methods to inhibit inherent advantages and disadvantages.
microbial growth in AFB, including modified atmo- Xie et al. (2003) used NIRS to follow the progress of
spheric packaging (CO2, NO2), irradiation, freezing, and staling. Using the NIRS, the researchers detected
changes in white pan bread structure during storage.
*Correspondent: Fax: +962 5390 3350; They compared the NIRS results to texture analysis
e-mail: [email protected] (TA). They found that there was a high correlation

doi:10.1111/j.1365-2621.2007.01594.x
 2007 The Authors. Journal compilation  2007 Institute of Food Science and Technology Trust Fund
358 Effect of preservatives addition on the shelf-life extensions and quality of flat bread M. Abu-Ghoush et al.

between NIRS and TA measurements although NIRS After fermentation, the dough was scaled off into 24–
can follow the staling during storage better than TA. 60 g pieces. The pieces were rounded by hand into balls
The objectives of this study were to identify one or more and covered with plastic cover to prevent skin formation.
preservatives that may be included in an AFB formula- The dough rested for 10 min. The dough was dusted
tion to extend shelf-life. Secondly, to investigate if TA and lightly with flour and flattened by gentle hand pressure
NIRS could be used to detect the effect of preservatives after which the dough was passed through a two stages
addition on the changes associated with staling. roll sheeter. In the first stage, the gap between the rolls
was 10 mm, and in the second stage, the gap was set to
3 mm. The sheeted dough was transferred to a stainless
Materials and methods
board and covered with a piece of cloth to minimize
moisture loss. Then, the dough was transferred for final
Wheat collection and milling
proofing for 30 min at 86 F (30 C )and 70 ± 5% RH.
Jagger wheat from crop year June 2002 was collected The bread was baked at 400 C (752 F) for 90 s, on a
from a local farm in Manhattan KS. This Jagger wheat preheated aluminium tray (215–450 mm) that accom-
was used to obtain flour with approximately 74–75% modates two full size loaves. However, for consistent
extraction rate after the milling (milled in the Depart- results, one loaf was baked each time. Three batches of
ment of Grain Science and Industry, Kansas State bread, 24 loaves per batch were produced for each
University, Manhattan, KS, USA). The flour was treatment. After baking, the bread was cooled for
collected in 50 lb-bags (22.7 kg) and stored in a 10 min. Twelve loaves were stored in high density
)10 C freezer. Proximate analysis of the wheat and polyethylene bags at 23C ⁄ 50% RH and used for
the flour used in the bread production were conducted near-infrared spectroscopy at days 0, 1, 2, 3 for control
by the Analytical Laboratory of the Department of (C) and fumaric acid (F) treatments, and 0,1, 2, 3, and
Grain Science and Industry (Kansas State University, 7 days for propionate (P) and fumaric acid combined
Manhattan, KS, USA). Moisture content was deter- with propionate treatments (P–F) (These days were
mined by using AACC approved method 44–15 A. selected according to the mould appearance and bad
Crude protein was determined according to AACC bread quality which indicate ending loaves shelf-life).
approved method 46–16. Ash was determined according The other twelve loaves were stored in high-density
to AACC approved method 08–01. Falling number was polyethylene bags and used for texture, moisture, water
conducted according to AACC approved method 56– activity and colour evaluation for the same storage
81B (1995). Starch damage was determined according to periods and conditions and stored at 23 C ⁄ 50% RH.
approved method 76–30A using Chopin SD4. Flour After 2 h, the loaves were assessed with the quality
colour was determined using a colorimeter (Mini Scan parameters for day 0 quality evaluation.
Minolta CR-210, Minolta Corporation, Konica Minolta
Germany, Munich, Germany).
Arabic bread analysis
Shelf-life
Arabic bread production
Mould growth on the AFB is an indicator of length of
Sodium propionate (0.3%) and fumaric acid (0.2%) shelf-life. Every AFB surface was checked every day for
were used as preservatives in four different Arabic the appearance of mould.
bread formulas. These preservatives were either used
alone or in combination. A mixture containing 1000 g Moisture content, water activity and (pH)
flour (14% moisture basis) , 1% yeast and 1% salt The moisture content of AFB was measured using
were mixed with and without preservatives (sodium AACC method 44–15A (American Association of Cer-
propionate 0.3% and ⁄ or 0.2% fumaric acid). Water eal Chemist., 1995). The AFB moisture content was
(53%; for a consistency of 850 BU was used). This measured at 0 day for all the treatments and after 3 and
consistency was found by experimentation to give the 7 days of storage at 23C and 50% RH. Water activity
most reliable prediction of baking absorption ( Qaro- was determined using an AQUA LAB CX-2 (Decagon
oni et al., 1987; Quail et al., 1990) when mixed with the CO, Pullman, WA, USA) after calibration with stand-
dry ingredients at different speeds. The dough was ard salt (AOAC, 1980).The pH was measured with a pH
mixed at: low speed for 3 min. and then at medium meter (Accumet portable AP63, Fisher Scientific,
speed for another 3 min ( 1 min beyond development Denver, Colorado, USA) with automatic temperature
time) by using Hobart planetary mixer (model C-100, compensation following the AACC method 02–05
Hobart, Troy, OH, USA) . After mixing, the dough (American Association of Cereal Chemist., 1995)-Electr-
was transferred into a covered plastic bowel, placed ometric method. The pH meter was calibrated with
into a proofing cabinet, allowed to ferment for 60 min buffer solutions of 4 and 7. The pH measurements were
at 86 F (30 C) and 70 ± 5% RH. made at ambient temperature.

International Journal of Food Science and Technology 2008  2007 The Authors. Journal compilation  2007 Institute of Food Science and Technology Trust Fund
 Effect of preservatives addition on the shelf-life extensions and quality of flat bread M. Abu-Ghoush et al. 359

Texture analysis. A validation model was used to predict the staling

The change in the texture of AFB due to staling was among treatments compared to control. The control
measured using the tensile test. Measurements of time data was used to predict the change in staling for 0, 1, 2,
for tearing in tension (hold until time) was carried out. and 3 days for all treatments. The control was used only
Uniform AFB strips shaped like a bar bell (taken from for the first 3 days because shelf-life expired after 3 days
the center of the top 115 mm ‘long’, 20 mm wide at the because of mould growth. Staling at day 7 for PF
ends, and 6 mm wide at the centre) were gripped with a treatments was predicted using a model developed using
special clamp, one end was attached to a TA.XT2 all P data. The NIRS prediction of storage days for all
Texture Analyzer plate form, and the other end was treatments correlated with actual storage time. The AFB
attached to the texture analyzer arm. The distance staling rate was analysed by plotting NIRS measured
between the two arms was set to 85 mm. A tensile grip storage time against actual storage time. The predicted
probe was used with the following settings: pretest storage time and the raw spectra absorbance were used
speed: 1 mm ⁄ s, test speed: 1.7 mm ⁄ s, posttest speed: to characterize the staling of the AFB. These raw spectra
10 mm ⁄ s, 40% strain force: 10 g, trigger: mode auto were prepared by averaging the scanned spectra of all 24
with acquisition rate 250 points per second (pps). loaves of the three replicates for all of the treatments.
The RPD (Ratio of performance to deviation, and is
Near-infrared spectroscopy – spectra measurements calculated by dividing the standard deviation of the
A diode-array NIR spectrometer (DA7000 Perten reference data by the standard error of performance.
Instruments, Springfield, IL, USA) was used to collect The RPD is an indicator of the usefulness of a
spectra. The wavelength range was 400–1700 nm. Data calibration used in the validation or prediction by the
were recorded as (1 ⁄ R), where R is the relative reflect- standard error of performance. The RPD is used as an
ance. A reference standard Spectralon was used to indicator of the usefulness of a calibration (Williams,
collect the baseline. Each spectrum was recorded as an 1997).
average of 15 scans ⁄ s.
Twelve loaves from each treatment were measured
Statistical analysis
daily until the bread reached end of shelf-life as
determined by the presence of mould. NIRS Measure- The statistical model was a two-way factorial classifica-
ments was taken on the top and the bottom of each loaf. tion in complete randomised design. Data were analysed
The average of 15 scans were taken on each section using statistical analysis software (version 8.2, SAS
every second and was recorded as one spectrum. Twenty Institute Inc., Cary, NC, USA). Analysis of variance
four NIR spectra were taken for each treatment ⁄ day. (anova) and means separations were calculated by the
general linear model procedure (Proc GLM). Compar-
Data analysis isons among treatments were analysed using Fisher least
A cross validation method was employed to determine significant difference (LSD). Treatment means were
the applicability of NIRS to distinguish staling in AFB considered significant at P £ 0.05.
within each treatment. Spectra were analysed with
partial least squares (PLS) regression (Martens and
Results and discussion
Naes, 1989) using a commercial software (PLS Plus ⁄ IQ
for Grama ⁄ 32, Galactic Industries, Salem, NH, USA) at
Moisture content, water activity and pH
550–1700 nm wave length range. The optimum number
of PLS factors was determined by cross validation. The Moisture content did not differ significantly among
number of PLS factors for the treatments, C, F, P, and treatments at 0 (Fig. 1), although, there was a signifi-
PF was 12, 7, 10, and 10, respectively. These factors cant decrease in the moisture content within treatments
were judged by the following criteria: beta coefficient from day 0 to day 3. Zobel & Kulp (1996) reported that
(the spectra with the least amount of noise), higher R2, water is an effective plasticizer in bread. They noted that
and the minimum partial residual error sum of squares after 3 days, the water content significantly decreased
(PRESS). Cross validation was employed to create the and the bread was firmer. Therefore, they concluded
calibration models. The critical wavelengths identified that less water was available to act as a plasticizer to
(those with the maximum absorbance) were used to keep the bread soft. Quail (1996) reported that staling is
measure the beta coefficients for the calibration models. due, in part, to moisture loss but is mainly because of a
The cross validation involved removing one sample chemical change to the starch–protein interaction. It was
from the data set (n) and performing the calibration with found that the baked breads at low water levels (during
the rest of the data set (n)1). All samples in the dataset moisture redistribution) bread begins of mechanical
were left out and measured once in turn (n)a, n)b) . The firming, and losses of softness. This could be due to the
first step in creating of the calibration model was to starch recrystallisation and starch–protein interactions
determine the PRESS value. which are the major component of bread firming.

 2007 The Authors. Journal compilation  2007 Institute of Food Science and Technology Trust Fund International Journal of Food Science and Technology 2008
360 Effect of preservatives addition on the shelf-life extensions and quality of flat bread M. Abu-Ghoush et al.

40 Addition of P to the bread formula did not change the

0 day 3 days 7 days pH (pH 5.60), but the shelf-life of AFB was extended to
8 days. Luck & Jager (1997) reported that the propion-
ate inhibitory action may be as a result of the accumu-
a a lation of the preservative in the cells of the mould.
a
a b Within the cells of mould, sodium propionate dissoci-
b b b
Moisture %

b b ates in water to form sodium ions and propionate ions.

30 The propionate ions react with water to form propionic
acid (three carbon skeleton). The toxicity of propionic
acid is related to inability of the moulds to metabolize
the three carbon skeleton (Fennema, 1996). This action
inhibits the growth of the mould.
Adding sodium propionate combined with fumaric
acid to the bread formula changed the pH ( pH 4.77).
Essentially, the fumaric acid increased the effectiveness
20
C P F PF of sodium propionate. These results are consistent with
those obtained by Wijaya (2003) in tortilla bread.
Treatments

Figure 1 The moisture changes in Arabic flat bread produced with

Tearing force
different preservatives [control (C ), fumaric acid (F), sodium-
propionate (P), and sodium propionate-fumaric acid combination Just because preservatives can inhibit microbial growth
(PF)] at 0, 3, and 7 days storage. does not mean that the quality of the product is
extended or improved.
Rogers et al. (1988) reported that moisture content was Figure 2 shows that no significant difference in tearing
found to be inversely proportional to the rate of staling. forces among the treatments at day 0. After 1 day, P and
Hallberg & Chinachoti (2002) found that standard white PF treatments exhibited the same tearing force as day 0,
bread showed a more rapid increase in firmness during indicating that there was no change in bread quality.
storage mainly because of the loss of moisture to the However, a significantly lower tearing force was
crust and surrounding environment. observed for the control and the F treatment at day 1.
The water activity for AFB formulated with the At day 2, both the P and PF treatments showed a
preservatives ranged between 0.945 and 0.955 and was significant decrease in tearing force. At day 3, there was
not significantly different among treatments or during no significant difference between the control and PF
storage. These results are consistent with the results treatment in tearing force. These treatments showed
obtained by Czuchajowska et al. (1989) who found that significantly higher tearing force compared to F and P
there was little change in the aw between the fresh crumb treatments, suggesting that adding PF treatment to the
and the older crumb. flour did not affect the AFB tearing force after 3 days.
The pH of the bread for the control, F, P, and P–F However, adding F or P separately affected the AFB
were 5.51, 4.27, 5.60, and 4.77, respectively. These tearing force after 3 days. These results may indicate
differences in the pH play an important role in deter- that PF treatment slowed staling; the AFB was not as
mining the taste, colour, and texture of the finished brittle as for either the F or P treatment. Thus, the
product (Maselli & Pomper, 1960). preservatives do inhibit mould growth but do not help
Shelf-life of AFB formulated with preservatives with maintaining freshness.
After 3 days, the AFB control and F treatments Overall, the force required to tear the bread decreased
exhibited mould growth (Rhizopus nigricans - bread significantly with storage time for all the treatments,
mould) indicating end of shelf-life (Fraizer & Westhoff, suggesting that the bread quality of different treatments
1978). The P treatment extended shelf-life of the AFB an decreased over time. These results are consistent with
additional 5 days compared to the control and F the results obtained by Rao et al. (1986) in Chapati
treatments. A 16 day shelf-life was obtained for AFB bread. They found that the tearing strength (indicative
formulated through the synergetic action of PF in the of brittleness) decreased markedly during storage. Tou-
combination preservative treatment. feili et al. (1998) reported that the force required to
The addition of fumaric acid to the bread formula initiate tearing in AFB and probing extensibility
increased acidity (pH 4.27) but did not extend the shelf- decreased significantly as the bread aged. The sample
life of AFB compared to other preservatives. The lost resilience and became brittle after 3 days of storage.
moulds can grow at this pH range (Fraizer & Westhoff, These changes may be as a result of the retrogradation
1978), but fumaric acid increased the effectiveness of of amylopectin fractions, protein–starch interactions,
preservative agents (Maselli & Pomper, 1960). and the redistribution of moisture between starch and

International Journal of Food Science and Technology 2008  2007 The Authors. Journal compilation  2007 Institute of Food Science and Technology Trust Fund
 Effect of preservatives addition on the shelf-life extensions and quality of flat bread M. Abu-Ghoush et al. 361

C P F PF
70 a
65 a a a a a
60 a b b b b
b b
55 b b b
c
50 c
45

Force (g)
40
35
30
25
20
Figure 2 The tearing force of Arabic flat 15
bread produced with different preservatives 10
(control (C), fumaric acid (F), Sodium-Prop- 5
ionate (P), and Sodium propionate–fumaric 0
acid combination (PF)) at 0 , 3, and 7 days 0 1 2 3 7
storage. Days

protein systems (Toufeili et al., 1993; Sidhu et al., 1997). consistent than the tearing force results. These results
These changes as well as the differences in the pH were consistent with the results obtained by Toufeili
between different treatments may be responsible for the et al. (1998) who found that probing extensibility
changes in the firming of AFB during storage. (indicating the time of tearing) decreased significantly
(P < 0.05) as aging progressed. These changes could be
attributed to the retrogradation theories mentioned
Tearing time
above (Toufeili et al., 1993; Sidhu et al., 1997). Bread
The AFB-PF treatment exhibited a significantly higher staling is a complex process that involves a set of
tearing time compared to the other treatments at day 0 physiochemical chemical changes (moisture distribution,
(Fig. 3). After day 1, the PF treatment exhibited a 67% firmness, tearing properties). These changes could be
reduction in tearing time compared to day 0. Addition- responsible for the significant reduction in the bread
ally, all other treatments exhibited a significant decline tearing force and time during storage.
in tearing time from day 0 to day 1. The reduction in The R2 for the tearing force ranged between 0.61 and
tearing time may indicate loss of freshness. At day 3, 0.98, whereas with standard error of determinations
there was no significant difference between the control ranged between 5.1 and 9.1. However, the R2 for the
and PF treatment in tearing time, although there was a tearing time ranged between 0.84 and 0.99, with
significant difference between the control, PF from one standard error of determination between 0.28 and
side and F, and P from another, suggesting that adding 0.65. These results indicate the importance of using the
the PF combination to the flour did not affect on the tensile test by using texture analyzer to study the
AFB tearing time after storage compared to the control changes in Arabic bread over time.
while adding F and P separately affected the AFB
tearing time after storage.
Staling evaluation of Arabic flat bread using NIRS
The measure of tearing time vs. tearing force is
ostensibly a more sensitive indicator of a change in The cross-validation model was used to evaluate the
freshness. The tearing time results obtained were more effectiveness of the NIRS to measure staling within each

8
a C P F PF
7
6 b b
5 c
second

d
4 d d
e e f e e
f e
3 f f f f
2
Figure 3 The tearing time of Arabic flat bread
1
produced with different preservatives (control
(C), fumaric acid (F), Sodium-Propionate (P), 0
and Sodium propionate–fumaric acid combi- 0 1 2 3 7
nation (PF)) at 0, 3, and 7 days storage. Days

 2007 The Authors. Journal compilation  2007 Institute of Food Science and Technology Trust Fund International Journal of Food Science and Technology 2008
362 Effect of preservatives addition on the shelf-life extensions and quality of flat bread M. Abu-Ghoush et al.

Table 1 Summary of the NIRS for different treatments using cross

NIR (predicted days)

top and bottom
validation 9 C F P PF
7 a
Treatments Loaf section R2 a
SEEb RPDc b
5 c c
3 d e ed
d ed
Control Top + bottom 0.82 0.25 5.1 f f e
1 g f hg
Fumaric acidd Top + bottom 0.9 0.33 3.9 g
–1
Sodium propionatee Top + bottom 0.93 0.6 2.5 0 1 2 3 7
Sodium propionate + fumaric acidf Top + bottom 0.93 0.54 2.39 Actual storage days
a
All correlation are with actual storage time. Figure 4 The NIRS results of the staling of Arabic flat bread with
b
SEE refers to standard error of estimation. different preservatives(control (C), fumaric acid (F), sodium-propion-
c
RPD = ratio of standard error of performance to the standard error ate (P), and sodium propionate–fumaric acid combination (PF)) using
deviation of the reference data. validation. Bars with the same letters are not significantly different at
d
0.2% Fumaric acid. P < 0.5.
e
0.3% Sodium propionate.
f
0.2% Fumaric acid + 0.3% sodium propionate.
treatments using other preservatives at day 0, 1, 2, and 3
of storage. There was no significant difference in the
treatment. The top-bottom models of the AFB were freshness loss between the control and PF after 3 days of
evaluated. The R2, SEE, RPD were shown in Table 1. storage. So the highest freshness loss was in bread
These high correlations and good range of RPD indicate produced using F alone. This may be as a result of the
the importance of using NIRS as a tool to study the active double bond within the fumaric acid structure.
staling of AFB produced by different treatments. It Sidhu et al. (1980) reported that fumaric acid has a
means that they can be used to predict the effect of reducing effect. This mechanism appears to be a free
different additives on the staling. These results were radical that stabilizes the sulphydryl protein groups,
consistent with the results obtained by Xie (2002) in resulting in a dough structure that has fewer disulphide
studying the staling of white pan bread by using NIR. cross-links. So the final product was S-Succinyl-l-
Researcher reported that RPD for different batches cystein with additional two carboxyl groups. This
ranged between 3.7 and 5.0. compound may promote interactions between the amy-
The validation model was used to evaluate the lose and amylopectin fractions or may result in an
effectiveness of the NIRS to measure staling amongst increased interaction between protein and the starch
treatments. The top-bottom models of the AFB were during aging, which would directly affect the staling
evaluated. The R2 range between 0.88 and 0.92 (Table 2) rate. As discussed previously, F decreased the pH (4.27)
with SEE range between 0.25 and 0.52 for all the of the bread when compared to the control (5.52). Thus,
different treatments. These results indicate the import- fumaric acid may change the ionic status of the gluten
ance of using NIRS validation to distinguish between and subsequent aggregation (Friend et al., 1995). The
staling of different treatments. lowest freshness loss was observed for the bread
The top and bottom sections of the AFB-F treatment produced with sodium propionate alone (pH 5.62).
exhibited a significantly higher freshness loss than However, Luck & Jager (1997) reported that propionic
treatments using other preservatives at day 0, 1, 2, and acid may interact with some amino acids in the gluten
3 of storage (Fig 4). However, the top and bottom AFB- chain and affect protein functionality.
P treatment exhibited a significantly lower staling than The absorption spectral differences offer another way
to evaluate freshness loss. These spectral differences are
shown in Fig. 5. The absorbance spectral differences
Table 2 Summary of the NIRS for different treatments using valid- were highest between C and F treatments and the
ation absorbance spectral differences (Fig. 5) were lowest
Treatments Loaf section R2 a
SEEb
between C and PF treatments . The absorbance for the
control was lower than the absorbance for the bread
Control Top + bottom 0.95 0.25 with either sodium propionate or the combination of
Fumaric acidc Top + bottom 0.88 0.69 sodium propionate and fumaric acid. These results were
Sodium propionated Top + bottom 0.97 0.48 consistent with the results of freshness loss mentioned
Sodium propionate + fumaric acide Top + bottom 0.92 0.52 above. The bread with F had highest freshness loss with
a
All correlation are with actual storage time.
lowest absorption while the bread formulated with the
b
SEE refers to standard error of estimation. PF combination had the same freshness loss as control
c
0.2% Fumaric acid. and nearly the same absorption. These results may be
d
0.3% Sodium propionate. explained by the starch retrogradation and crystallinity.
e
0.2% Fumaric acid + 0.3% Sodium propionate. Starch reverts to an insoluble state during staling (Zobel

International Journal of Food Science and Technology 2008  2007 The Authors. Journal compilation  2007 Institute of Food Science and Technology Trust Fund
 Effect of preservatives addition on the shelf-life extensions and quality of flat bread M. Abu-Ghoush et al. 363

1.6
Control (C)
Fumaric Acid (F)
1.4 Propionate (P) Propionate
Fumaric Acid + Propionate (FP) (FP)
Control
1.2

Fumaric Acid

Absorption
1

0.8

0.6

Figure 5 Raw Spectral absorption of Arabic 0.4

Peak Absorption
flat bread with different preservatives (control Wavelength = 1450 nm
(C), fumaric acid (F), sodium-propionate (P),
0.2
and sodium propionate–fumaric acid combi- 400 600 800 1000 1200 1400 1600 1800
nation (PF)) after 3 days storage. Wavelength (nm)

& Kulp, 1996). This crystal form of starch may scatter storage with high correlations and low standard
more light back to the sensor, which is seen as lower deviations.
absorption in the bread with fumaric acid and higher
absorption in bread with sodium propionate. Although
References
there was no change in absorbance spectral differences
for all of the treatments were observed from 550 to American Association of Cereal Chemists (AACC). (1995). Approved
1150 nm except the Control-Fumaric spectral differ- Methods of the AACC, 9th ed. St. Paul, MN: American Association
of Cereal Chemists.
ences. The greatest difference in absorbance spectra was AOAC. (1980). Official Methods of Analysis of Analytical Chemists.
observed from 1430 to 1495. This peak might relate to 13th ed. Washington, DC: The Association of Analytical Chemists.
the second overtone of C–H, C–H2, CONH2, ROH, and Baik, M. & Chinachoti, P. (2000). Moisture redistribution and phase
H2O. transition during bread staling, Cereal Chemistry, 77, 484–488.
Czuchajowska, Z., Pomeranz, Y. & Jeffers, H.C. (1989). Water activity
This means that moisture loss, starch retrogradation, and moisture content of dough and bread, Cereal Chemistry, 66,
and protein interaction might affect bread staling . Xie 128–132.
(2002) reported that two important wavelengths 550 and Fennema, R. (1996). Food Chemistry. New York, NY: Marcel Dekker
1465 nm, were key for the NIRS to successfully classify Inc.
the starch–starch and starch–protein bread. However, Fraizer, W.C. & Westhoff, D.C. (1978). Food Microbiology. New
York, NY: McGraw-Hill, Inc.
three wavelengths 970, 1155 and 1395 nm, were associ- Friend, C.P., Ross, R.G., Waniska, R.D. & Rooney, L.W. (1995).
ated with amylopectin retrogradation. Wilson et al. Effect of additives in wheat flour tortillas. Cereal Foods World, 40,
(1991) reported that the maximum absorption for bread 494–497.
occurred at 1414 and 1465. These wavelengths were used Grundy, J.G. (1996). Preservatives. In: Baked Goods Freshness (edited
by R.E. Hebeda & H.F. Zobel). Pp. 189–203. New York, NY:
to study the states of water in the foods with respect to Marcel Dekker Inc.
hydrogen bonding. Gurjal, H.S. & Gaur, S. (2002). Effects of barley flour, wet gluten and
There was a good relationship between NIRS and liquid shortening on the texture and storage characteristics of
texture analysis data for the control and PF treatments. Chapati. Journal of Textural Studies, 33, 461–469.
In both methods, the control and PF treatments showed Hallberg, L.M. & Chinachoti, P. (2002). A fresh perspective on staling:
The significance of starch recrystalization on the firming of bread,
no differences in the freshness loss. Food Engineering and Physical Properties, 67, 1092–1096.
Luck, E. & Jager, M. (1997). Antimicrobial Food Additives. Pp. 1–260.
Berlin: Springer.
Conclusions Martens, H. & Naes, T. (1989). Multivariate Calibration. Chichester:
John Wiley and Sons.
The F–P inhibited mould growth 320% longer com- Maselli, J.A. & Pomper, S. (1960). The significance of pH in baking.
pared to the control. Tearing time proved to be a Bakers Digest, 10, 66–88.
better indicator of loss of freshness compared to the Mousa, E.I., Ibrahim, R.H., Shuey, W.C. & Maneval, R.D. (1979).
tearing force. NIRS proved to be very valuable tool in Influence of wheat classes, flour extraction, and baking methods on
indicating the changes in freshness of AFB during the Egyptian balady bread, Cereal Chemistry, 56, 563–566.

 2007 The Authors. Journal compilation  2007 Institute of Food Science and Technology Trust Fund International Journal of Food Science and Technology 2008
364 Effect of preservatives addition on the shelf-life extensions and quality of flat bread M. Abu-Ghoush et al.

Qarooni, J. (1990). Flat breads. AIB Research Department Technical Toufeili, I., Sleiman, G., Abu Salman, R. & Brockway, B. (1994).
Bulletin, 12, 1. Effect of staling on viscoelastic properties of pastes prepared from
Qarooni, J. (1996). Flat Bread Production. Pp. 1–141. New York, NY: Arabic bread, Journal of the Science of Food and Agriculture, 64,
Chapman and Hall. 271–273.
Qarooni, J., Moss, H.J., Orth, R.A & Wootton, M. (1987). A test Toufeili, I., Chammas, H. & Shadarevian, S. (1998). Identification of
baking technique for Arabic bread quality, Journal of Cereal key textural attributes of Arabic bread and their application to the
Science, 6, 69–80. assessment of storage and quality, Journal of Textural Studies, 29,
Quail, K.J. (1996). Arabic Bread Production. Pp. 1–129, St. Paul, MN: 57–66.
American Association of Cereal Chemists. Wijaya, N. (2003). Shelf Life and Rollability of Hot Press Wheat Flour
Quail, K.J., McMaster, G.J & Tomlinson, J.D. (1990). Effect of baking Tortillas. M.Sc. Thesis. Manhattan, KS: Kansas State University.
temp ⁄ time conditions and dough thickness on Arabic bread quality, Williams, P.C. (1997). Recent Advances in Near-Infrared Applications
Journal of the Science of Food and Agriculture, 53, 527–540. for the Agriculture and Food Industries. Manhattan, KS: Proc.
Rao, G.V., Leelavathi, K., Rao, P.H. & Shurpalekar, S.R. (1986). International Wheat Quality Conference.
Changes in the quality characteristics of chapati during storage, Wilson, R.H., Goodfellow, B.J., Belton, P.S., Osborn, B.G., Oliver, G.
Cereal Chemistry, 63, 131–135. & Russell, P.L. (1991). Comparison of Fourier transform mid
Rogers, D.E., Zeleznak, K.J, Lai, C. S. & Hoseney, R.C. (1988). Effect infrared spectroscopy and near infrared reflectance spectroscopy
of native lipids, shortening, and bread moisture on bread firming, with differential scanning calorimetry for the study of the staling of
Cereal Chemistry, 65, 398–401. bread, Journal of the Science of Food and Agriculture, 54, 471–483.
Sidhu, J.S., Nordin, P. & Hoseney, R.C. (1980). Mixograph studies III. Xie, F. (2002). The Study of Bread Staling Using Visible and Near Infra
Reaction of fumaric acid with gluten proteins during dough mixing, Red Reflectance Spectroscopy. PhD Thesis. Manhattan, KS: Kansas
Cereal Chemistry, 57, 159–163. State University.
Sidhu, J.S., Caceres, P.G. & Behbehani, M. (1997). Measurement of Xie, F., Dowell, F.E. & Sun, X.S. (2003). Comparison of near-infrared
starch properties during staling of Arabic bread, Starch ⁄ Stärke, 49, reflectance spectroscopy and texture analyzer for measuring bread
180–186. changes in storage, Cereal Chemistry, 80, 25–28.
Toufeili, I., Sleiman, G. & Abu Salman, R. (1993). Correlation of Zobel, H.F. & Kulp, K. (1996). Staling mechanism. In: Baked Good
selected chemical and physico-chemical parameters and sensory Freshness (edited by R. E. Hebeda & H. F Zobel). Pp. 1–65. New
perception of staling in Arabic bread, Lebensmittel Wissenschaft York, NY: Marcel Dekker Inc.
Technologie (LWT), 26, 34–37.

International Journal of Food Science and Technology 2008  2007 The Authors. Journal compilation  2007 Institute of Food Science and Technology Trust Fund