DURUM WHEAT AND PASTA

Effect of Process Variables on Spaghetti Quality'

A. DEBBOUZ 2' 3 and C. DOETKOTT 2

ABSTRACT Cereal Chem. 73(6):672-676

Five process variables (water absorption, barrel temperature, screw had a significant effect on pasta color and firmness, while water tem-
speed, mixing time, and water temperature) were investigated to deter- perature affected only firmness. Contour plots showed that spaghetti
mine their effects on pasta quality. Response surface methodology brightness and firmness improved at lower water absorption (30.5-31%)
(RSM) was used to identify the process variables levels yielding opti- and lower barrel temperature (35-45 0 C) and at intermediate screw speed
mum pasta color and cooking characteristics. Water absorption and barrel (25 rpm). Spaghetti yellowness showed similar trend, except that it in-
temperature had the greatest effect on the response variables studied creased at higher barrel temperature. Cooking loss appeared to be mini-
(pasta brightness, yellowness, firmness, cooked weight, and cooking mized at water absorptions and barrel temperatures ranging from 31.5 to
loss). Screw speed was a significant factor in the response surface mod- 32% and 45 to 500 C, respectively.
els for all the pasta quality parameters, except cooking loss. Mixing time

The production of top-quality pasta requires not only high qual- tions used in pasta certainly need to be fully investigated for their
ity semolina but also optimal processing conditions. Semolina char- optimization. Only two studies (Walsh et al 1971, Abecassis et al
acteristics and their effect on pasta quality have been fully doc- 1994) have reported the process parameter effects on pasta quality.
umented (Matveef 1966; Matsuo and Irvine 1970; Walsh et al 1971; The objectives of this research were to study the process vari-
Matsuo et al 1972; Seyam et al 1974; Dexter and Matsuo 1978, ables that had, according to the literature, the most effect on pasta
1979, 1980; Grzybowski and Donnelly 1979; Dick and Matsuo 1988), quality and to optimize these variables for the production of high
while little information has been published regarding the influence quality pasta.
of process variables on pasta quality. Among the processing stages,
pasta drying was perhaps the most studied. A review of the pro- MATERIALS AND METHODS
cessing conditions used for the production of pasta products showed
a wide variation in parameters utilized by different manufacturers. Raw Material
Hummel (1966) recommended mixing spaghetti dough 5-10 To conduct the experiment, 1,000 lbs. of commercial semolina
min at 30% moisture, using water temperatures of 40-60TC for the (Durakota 1) was obtained from the North Dakota State Mill. The
coarser grade semolina and lower temperatures for the finer gran- 10 bags (100 lbs. each) were uniformly blended together to pro-
ulation. He also reported that the dough should be extruded at duce a homogeneous semolina stock.
screw speeds between 20-40 rpm. Some of the advantages stated
for mixing warm water include color improvement, dough sof- Semolina Analysis
tening and ease of extrusion, and smoother macaroni surface. The moisture, protein, ash, and wet gluten were determined on
Similar pasta manufacturing conditions have also been used by the semolina according to standard methods (AACC 1995). Mix-
Matsuo and coworkers (1978). ograph evaluation of semolina was determined according to the
Manser (1981) found that the best pasta was obtained when the standard method with some modifications. The semolina sample
pasta dough was extruded at 31% absorption and at 40-50TC bar- (10 g) was mixed 8 min at constant water absorption of 5.8 ml,
rel temperature using drying temperature not higher than 80TC. using a spring setting of 8. The mixogram was scored by com-
According to Milatovic (1985), water temperatures between 36 and paring it to reference mixograms (Dick and Youngs 1988).
45TC should be used for the cold dough making and water tem- The number of specks in semolina was determined on a flat sur-
peratures between 45 and 650C should be used for the warm face by counting the visible specks (bran and black particles) in
system, where pasta is usually dried at high temperature. He also three different 1 in. sq. areas. The average of the three readings
suggested 15 min for mixing time and 29.5% for dough moisture. was converted to the number of specks per 10 sq. in.
Kobrehel and Abecassis (1985) used a temperature of 35TC to The commercial semolina was of intermediate gluten strength
extrude a spaghetti dough that had been mixed for 25 min, while (mixogram score = 6) and had 11.9% protein, 0.72% ash, 398 sec
Kim and coworkers (1986) used only 2 min mixing time. falling number, 32.3% wet gluten, and 16 specks/10 sq. in.
For experimental pasta extruders, Cubadda (1988) recom-
mended screw speeds between 45 and 50 rpm, and barrel tem- Experimental Design
perature and dough moisture not to exceed 50TC and 31%, respec- The process variables (independent variables) investigated were
tively. Whether all these processing conditions reported by the water absorption (30-32-34%), water temperature during mixing
different macaroni manufacturers and pasta research laboratories (35-45-55 0 C), mixing time (3-5-10 min), barrel temperature (35-
make the best pasta possible is not known. The processing condi- 45-55°C), and screw speed (20-25-30 rpm). The dependent vari-
ables or pasta quality factors measured were spaghetti yellowness
(b value), spaghetti brightness (L value), cooked weight, cooked
1
Published with the approval of the Director of the Agricultural Experiment
firmness, and cooking loss. A 35 factorial design was used with
Station, North Dakota State University, Fargo, ND. two replicates collected at each of the 243 combinations.
2
Department of Cereal Science and Information Technology Services, respec-
3
tively, North Dakota State University, Fargo, ND 58105. Pasta Processing and Drying
Corresponding author. Fax: 701/231-7723. The spaghetti samples were produced using 900 g of semolina
Publication no. C-1996-0925-03R. samples, as described by Walsh et al (1971). The process variables
C 1996 American Association of Cereal Chemists, Inc. were changed to fit the conditions of the experimental design. The
672 CEREAL CHEMISTRY
extruded spaghetti samples were dried using a high-temperature Statistical Analysis
drying cycle (Debbouz et al 1994). The data were analyzed using the response surface regression
analysis procedure (RSREG) of the Statistical Analysis System
Spaghetti Analysis (SAS Institute Inc., Cary, NC) and the fit model platform in the
Spaghetti L and b values were determined by light reflectance Analyze menu of JMP (SAS Institute). A second-order response
using a Minolta color difference meter (model CR310, Minolta surface model was initially fit to each of the five pasta quality
Camera Co., Japan) according to the standard method (AACC factors using the five process variables as predictors. The full
1995). The spaghetti samples were cooked according to the second-order response surface models included 20 terms (5 linear,
method of Dick et al (1974). Cooked weight was determined by 5 quadratic, and 10 cross products) plus an intercept. Fitted reduced
weighing the rinsed spaghetti and reporting the results in grams. models were also used in an attempt to obtain a more parsimoni-
Cooking loss of the cooking and rinse waters collected from each ous description of the data. The reduced models included all the
sample was determined by evaporating to dryness in an air oven at significant (P < 0.05) linear, cross product, and quadratic terms.
115'C. The residue was weighed and reported as percentage of the Linear terms that were not significant, but had significant cross
original spaghetti sample. product or quadratic terms, were also retained in the model.
Spaghetti firmness was measured by shearing two cooked spa- Because of the small differences observed between the full and
ghetti strands with a specially designed plexiglass tooth (Oh et al reduced models, the simpler reduced models were used to gener-
(1983) attached to a texture analyzer (Texture Technologies Corp. ate the response surface contour plots. The responses evaluated
Scarsdale, New York). were spaghetti brightness, yellowness, cooked firmness, cooked
weight and cooking loss. Equations were calculated describing the
response of each variable studied as a function of the experimental
factors selected. Contour plots were generated for each pasta
TABLE I quality parameter using all pairs of process variables that were
T-Ratios for the Quality Parameters in the Response Surface Regression included in the models, with the remaining process variables held
Color Cooking Characteristics at their center points. These plots help identify trends in the prod-
uct quality at different process variable levels.
Independent Brightness Yellowness Cooked Cooking
Variablesa L b Firmness Weight Loss
RESULTS AND DISCUSSION
Linearb
Xl -14.49**c 5.80** 2.28* -6.53** -7.45**
-28.19**
Table I shows the process variables that had a significant effect
X2 5.72** 7.20** 3.94** -10.36**
X3 -2.83** 3.54** 3.33** 6.21** nsd on spaghetti color and cooking quality. Water absorption and ex-
X4 ns ns -2.71** ns ns truder barrel temperature affected all of the spaghetti quality param-
X5 ns 7.97** -2.36* ns ns eters tested. The extruder screw speed had an effect on spaghetti
Quadraticb color and cooking properties (firmness, cooked weight). Mixing
X12 10.77** -4.94** -2.94** 8.13** 7.14** time affected mostly spaghetti yellowness and firmness. Pasta
X22 10.07** ns -6.01** -6.60* 7.72**
X32 3.99** ns -5.93** ns 2.08*
yellowness (b value) was higher when shorter mixing time and
X52 ns ns ns -3.08** ns higher barrel temperatures were used (contour plots not shown).
Cross productb Since most carotenoid pigment destruction by lipoxygenase
XlX2 30.43** -6.13** -7.58** 3.72** 5.88** occurs during the mixing stage (Tsen and Hlynka 1962, Linstroth
XX 3 ns -4.18** ns -8.66** ns 1981), a shorter mixing time would be expected to reduce pigment
X1X4 ns ns 2.56* ns ns oxidation by lipoxygenase. Higher barrel temperatures may also
X1X5 ns -9.28** ns ns ns
X2X3 ns -2.00* 3.03** 0.97** 3.23**
have decreased lipoxygenase activity and destruction of the pig-
X2X5 ns -2.03* ns -2.17* 2.43* ments during extrusion. This enzyme is heat-labile with optimum
X3X5 2.15* ns ns ns ns and inactivation temperatures of 30 and 68°C, respectively (Fox
a XI = water absorption, X2 = barrel temperature, X3 = screw speed, X4 = and Mulvihill 1982).
water temperature, X5 = mixing time. The regression equations that were used to generate the contour
b Included only if they had a significant effect on one or more quality plots are presented in Table II. High R2 values and model signifi-
parameters. cance were obtained for most of the quality factors tested, except
c **= P < 0.01, * =P<0.05.
for cooking loss (R2 = 0.34). The R2 represents the proportion of
d Not significant.
variability in the data explained or accounted for by the model

TABLE II
Regression Equations Describing the Response of Each Quality Parameter as a Function of the Significant Effects Retained in the Model
Quality
Parameter Equation R2
Brightness (L) 216 - 7.47 XI - 1.37 X2 - 0.274 X3 - 0.121 X5 + 0.0341 X1 X2 + 0.00267 X3 X5 + 0.0853 X12 + 0.00319 X22 + 0.00506 X3 2 0.90
Yellowness (b) - 18.1 + 2.38 XI + 0.234 X2 + 0.284 X3 +0.742 X5 - 0.0056 X1 X2 - 0.0076 X1X3 - 0.0235 XX 5 - 0.000731 X2 X3 0.80
- 0.00103 X2X5 - 0.0319 X12
Firmness - 7.65 + 0.854 XI + 0.235 X2 + 0.233 X3 - 0.063 X4 - 0.000522 X5 - 0.00597 XX 2 + 0.00201 XX 4 + 0.000954 X2 X3 0.92
- 0.0164X12 - 0.00134 X2 2 -0.00528 X3 2
Cooked weight 88.9 - 5.68 XI + 0.325 X2 + 1.20 X3 + 0.313 X5 + 0.00685 X1X2 - 0.0319 XX 3 - 0.00587 X2X3 - 0.00222 X2X5 + 0.106 X12 0.85
- 0.00344 X22 - 0.0165 X5 2
Cooking loss 63.7-2.97 XI - 0.367 X2 - 0.144 X3 - 0.0291 X5 + 0.00498 XX 2 + 0.00109 X2 X3 + 0.00114 X2 X5 + 0.0427 X12 0.34
+ 0.00185 X~2 + 0.00199 X32
a XI = water absorption, X2 = barrel temperature, X3 = screw speed, X4 = water temperature, X5 = mixing time.

Vol. 73, No. 6,1996 673

(Montgomery 1984). To generate the contour plots, a combination creasing trend at higher barrel temperature and screw speed
of two independent variables were plotted at a time for the five (contour plot not shown). Abecassis et al (1994) also reported an
response variables. Selected contour plots, including the process increase in the spaghetti yellow index as the open surface of the
variables that had the most effect on spaghetti color and cooking die or extruding speed increased.
quality, are presented in Figures 1-8. Contour plots for spaghetti Although color and overall appearance of spaghetti are impor-
brightness and yellowness, as functions of water absorption and tant quality characteristics to the consumer, the texture or the
screw speed, are shown in Figures 1 and 2 respectively. Both quality firmness of the cooked product and its solid loss to the cooking
parameters increased as water absorption decreased. water are generally accepted as the ultimate quality parameters
Oh et al (1985) also observed an increase in noodle brightness at that greatly influence consumer acceptance.
lower water absorption. This is perhaps due to reduced enzymatic The contour plot for spaghetti firmness as function of water
browning (polyphenol oxidase) and carotenoid pigment oxidation absorption and barrel temperature is shown in Figure 3. It appears
at lower water absorption (Nicolas 1978). Walsh et al (1971) using that pasta firmness increased as water absorption and barrel
linear programming also found that high water absorption had an temperature decreased. This is in agreement with the results
adverse effect on pasta color. Spaghetti yellowness showed an in- reported by Walsh and coworkers (1971).

Color L (Brightness) Firmness

55 55

-
50 0
50

s5 I-.-
0, CU
CD
at
CD a)
45 a) 45
E a)

co
mi
m 40 40

35 35
30 31 32 33 34 30 31 32 33 34

Water Absorption (%) Water Absorption (%)

Fig. 1. Contour plot of spaghetti brightness (L value) as a function of Fig. 3. Contour plot of spaghetti firmness as a function of water
water absorption and barrel temperature. absorption and barrel temperature.

Color b (Yellowness) Firmness

55 55 - - -

15.. A
i.3
,
L, 50 - 0 50 - / =- DI5i.5
0 0,
L-

:3
0)
i5
0-
0,
0. 5i.7
CD
45 ED 45 -
CD
E
E
5i.9

m 40 40 -
m
6.1

35 35 I L Ii
I I
I

30 31 32 33 34 20 22 24 26 28 30

Water Absorption (%) Screw Speed (RPM)

Fig. 2. Contour plot of spaghetti yellowness (b value) as a function of Fig. 4. Contour plot of spaghetti firmness as a function of screw speed
water absorption and barrel temperature. and barrel temperature.
674 CEREAL CHEMISTRY
 Screw speeds close to the midpoint (25 rpm) gave the best firm- increase in cooking loss of 250% when the barrel temperature was
ness (Figs. 4 and 5). This may be attributed to the gluten rheo- raised from 35 to 700C. However, their study failed to show the
logical properties. Gluten was perhaps not sufficiently developed at critical limits or the optima for the process variables used.
lower screw speed and overworked at higher extrusion speed. As pre- Temperatures above 55 0C have been reported to denature gluten
viously reported (Matveef 1966, Matsuo and Irvine 1970, Feillet and to adversely affect pasta quality (Milatovic and Mondelli
1977), gluten has a very critical influence on pasta cooking quality. 1991). Manser (1981) also found that the best pasta products were
Cooked weight had a tendency to increase as water absorption extruded in the temperature range of 40-500 C. The shaded area in
and barrel temperature increased (Figs. 6 and 7). This result con- Figure 9 shows the predicted optimal process variable values that
curs with the results obtained by Abecassis et al (1994). As shown will produce spaghetti with bright yellow color, high firmness,
in Figure 8, the lower cooking losses were obtained at barrel tem- and low cooking loss. These values ranged between 30.5 and 31%
perature and water absorption that ranged between 45 and 500C for water absorption, and between 45 and 53 0C for barrel tem-
and between 31.5 and 32%. Abecassis et al (1994) reported an perature.

Firmness Cooked Weight

55

0-..
50 -
a- a)
V CD
a)
D- 45
C) E
a,
C)
C.) ad
m 40

35
30 31 32 33 34 20 22 24 26 28 30

Water Absorption (%) Screw Speed (RPM)

Fig. 5. Contour plot of spaghetti firmness as a function of water Fig. 7. Contour plot of cooked weight as a function of screw speed and
absorption and screw speed. barrel temperature.

Cooked Weight Cooking Loss

55

50
CD
CU
L-

0~
i5

a)
45
E

m 40

35
30 31 32 33 34 30 31 32 33 34

Water Absorption (%) Water Absorption (%)

Fig. 6. Contour plot of cooked weight as a function of water absorption Fig. 8. Contour plot of cooking loss as a function of water absorption and
and barrel temperature. barrel temperature.
Vol. 73, No. 6,1996 675
 DEBBOUZ, A., PITZ, W. J., MOORE, W. R., and D'APPOLONIA, B. L.
1994. Effect of bleaching on durum wheat and spaghetti quality.
Cereal Chem. 72:128.
DEXTER, J. E., and MATSUO, R. R. 1978. The effect of gluten protein
fractions on pasta dough rheology and spaghetti making quality. Cereal
Chem. 55:44.
O-%
DEXTER, J. E., and MATSUO, R. R. 1979. Effect of starch on pasta
0
Color b= 28 dough rheology and spaghetti cooking quality. Cereal Chem. 56:190.
0 DEXTER, J. E., and MATSUO, R. R. 1980. Relationship between durum
wheat protein properties and pasta dough rheology and spaghetti
cooking quality. J. Agric. Food Chem. 26:899.
DICK, J. W., WALSH, D. E., and GILLES, K. A. 1974. The effect of
ED field sprouting on the quality of durum wheat. Cereal Chem. 51:180.
DICK, J. W., and MATSUO, R. R. 1988. Durum wheat and pasta prod-
ucts. Pages 507-547 in: Wheat Chemistry and Technology, Vol. II. Y.
L. Pomeranz, ed. Am. Assoc. Cereal Chem.: St. Paul, MN.
0
Cooking Loss =5.7 DICK, J. W., and YOUNGS, V. L. 1988. Evaluation of durum wheat,
semolina, and pasta in the United States. Pages 237-248 in: Durum
Color L =55 Chemistry and Technology. G. Fabriani and C. Lintas, eds. Am. Assoc.
Firmness = 6.0 Cereal Chem., St. Paul, MN.
___________________________________________ FEILLET, P 1977. La qualit6 des pates alimentaires. Ann. Nutr. Diet. 12:299.
l lI l [ FOX, P. F., and MULVIHILL, D. M. 1982. Enzymes in wheat, flour, and
30 31 32 33 34 bread. Pages 107-152 in: Advances in Cereal Science and Technology,
Vol. V.Y.Pomeranz, ed. Am. Assoc. Cereal Chem.: St. Paul, MN.
Water Absorption (%) GRZYBOWSKI, R. A., and DONNELLY, B. J. 1979. Cooking properties
of spaghetti. Factors affecting cooking quality. J. Agric. Food Chem.
Fig. 9. Opti mum region (shaded area) obtained by superimposing contour 27:380.
plots of spaaghetti quality parameters as a function of water absorption HUMMEL, C. 1966. Macaroni Products: Manufacture, Processing and
and barrel tiemperature. Packing. C. Hummel, ed. Food Trade Press: London.
KIM, H. I., SEIB, P A., POSNER, E., DEYOE, C. W., and YANG, H. C.
CONCLUSIONS 1986. Milling HRW to farina: Comparison of cooking quality and
color of farina and semolina spaghetti. Cereal Foods World 31:810.
KOBREHEL, K., and ABECASSIS, J. 1985. Effect of high temperature
Water altsorption and barrel temperature had the most influence drying of pasta products on the composition and activity of peroxidase
on spagheitti quality, since they had a significant effect on all of in relation with the color of spaghetti. Lebensm. Wiss. Technol. 18:277.
the respon [se variables investigated. Screw speed had an effect on LINSTROTH, J. 1981. What's new? Color in semolina and macaroni.
four (brigiitness, yellowness, cooked weight, and firmness) of the Macaroni J. 63:16.
five qualit, y parameters studied. Mixing time showed a significant MANSER, J. 1981. Optimale Parameter fur die Teigwarenherstellung am
effect on pPasta yellowness and firmness; water temperature affected Beispiel von Langwaren Getreide Mehl Brot 35:75.
only pastal firmness. Spaghetti yellowness showed a increasing MATSUO, R. R., and IRVINE, G. N. 1970. Effect of gluten on the qual-
trend at lo wer water absorption and at higher barrel temperature. ity on the cooking quality of spaghetti. Cereal Chem. 47:173.
absorption and barrel temperature had a positive MATSUO, R. R., BRADLEY, J. W., and IRVINE, G. N. 1972. Effect of pro-
Lower wa ater britn and firel temperature s appeared tein content on the cooking quality of spaghetti. Cereal Chem. 49:707.
effect on F)asta brightness and firmness. Cooking losses appeared MATSUO, R. R., DEXTER, J. E., and DRONZEK, B. L. 1978. Scanning
to be mini mized at water absorptions and barrel temperatures be- electron microscopy study of spaghetti processing. Cereal Chem. 55:744.
tween 31.5i and 32% and between 45 and 50TC, respectively. MATVEEF, M. 1966. Influence du gluten des bles durs sur la valeur des
From theC surface regression contour plots, it appears that spaghetti pates alimentaires. Bull. ENSMIC 213:116.
color and (cooking quality were optimized at =31% absorption, 25 MILATOVIC, L. J. 1985. Produzione di lasagna ricce con farina di grano
rpm screw speed, and 45-530 C barrel temperature. tenero ed essiccazione ad alta temperature. Tec. Molitoria 5:419.
MILATOVIC, L. J., and MONDELLI, G. 1991. Processes and unit op-
erations in industrial pasta manufacturing. Pages 69-95 in: Pasta Tech-
LITERATURE CITED nology Today. Chiriotti-Pinerolo: Italy.
MONTGOMERY, D. C. 1984. Design and Analysis of Experiments. John
ABECASSI IS, J., ABBOU, R., MOREL, M.-H., and VERNOUX, P. Wiley and Sons: Singapore.
1994. Inffluence of extrusion conditions on extrusion speed, tempera- NICOLAS, J. 1978. Effects de different paramrtres sur la destruction des
ture, and pressure in the extruder and on pasta quality. Cereal Chem. pigments carotenoides de la farine de bld tendre au cours du petris-
71:247. sage. Ann. Technol. Agric. 27:695.
AMERICA]N ASSOCIATION OF CEREAL CHEMISTS. 1995. OH, N. H., SEIB, P. A., DEYOE, C. W., and WARD, A. B. 1983.
Approve(d Methods of the AACC, 9th ed. Method 08-01, approved Noodles. I. Measuring the textural characteristics of cooked noodles.
April 19461, revised October 1981 and October 1986; Method 14-22, Cereal Chem. 60:433.
approved1 October 1976, reviewed October 1982 and October 1994; OH, N. H., SEIB, P. A., and CHUNG, D. S. 1985. Noodles. III. Effect of
Method 38-12, approved November 1995; Method 44-1SA, approved processing variables on quality characteristics of dry noodles. Cereal
October 1975, revised October 1981 and October 1994; Method 46- Chem. 62:437.
IhA, app)roved October 1976, revised October 1982 and September SEYAM, A., SHUEY, W. C., MANEVAL, R. D., and WALSH, D. E.
1985, retviewed October 1994; Method 54-40A, approved April 1961, 1974. Effect of particle size on processing and quality of pasta prod-
revised October
( 1988 and October 1994, final approval November ucts. Assoc. Oper. Millers Bull. December: 3497.
1995. Thie Association, St. Paul, MN. TSEN, C. C. and HLYNKA, I. 1962. The role of lipids in oxidation of
CUBADDAI, R. 1988. Evaluation of durum wheat, semolina, and pasta in dough. Cereal Chem. 39:209.
Europe. Pages 217-228 in: Durum Chemistry and Technology. G. WALSH, D. E., EBELING, K. A., and DICK, J. W. 1971. A linear pro-
Fabriani and C. Lintas, eds. Am. Assoc. Cereal Chem., St. Paul, MN. gramming approach to spaghetti processing. Cereal Sci. Today 16:385.

[Received February 6, 1996. Accepted July 17, 1996.]

676 CEREAL CHEMISTRY