Engineering, Environment

 

Optimization of plantain drying process using response surface methodology

 

Uwem INYANG*, Innocent OBOH and Michael IBABENG

 

 Department of Chemical and Petroleum Engineering

Faculty of Engineering, University of Uyo, Uyo, Akwa Ibom State, Nigeria

E-mail(s): *uweminyang7@yahoo.com; innocentoboh@uniuyo.edu.ng; belmadeng@gmail.com

* Corresponding author, phone: 08023621934

 

Received: October 24, 2017 / Accepted: December 27, 2017 / Published: December 30, 2017

 

Abstract

Optimization of plantain drying process parameters to obtained high quality product is very vital as improper drying conditions can affect the composition of the plantain flour and thereby increases the risk of running into deterioration of the nutrient contents of the product and loss. Matured unripe plantains were processed by washing, peeling and cut into required size. The samples obtained were analysed in triplicate using different drying oven temperatures (60, 70 and 80oC) and time (187, 242, and 311 minutes) at pulp size of 3 mm by employing Design of experiment (DOE). Statistical analysis and Response surface methodology based on the D-Optimal were carried out. The significant factors on the experimental design response were identified from the analysis of variance. The optimum conditions for the processing of unripe plantain into dry product were obtained by using temperature of 80oC, time of 187 minutes and pulp size of 3 mm which resulted in final carbohydrate content of 87.4%. From these results obtained, the use of response surface methodology in optimizing plantain drying process can be recommended.

Keywords

Oven; Drying; Optimization; Response surface methodology; Plantain drying; Design of experiment; D-Optimal; Moisture content

 

 

 

Introduction

 

            In recent years, much attention has been drawn to the quality of dehydrated food products. Different drying methods have effect on the biochemical and functional properties of dehydrated products [1].

Heat treated plantain flour, can be used for many applications in food processing such as cake, biscuit and wafer flours, beadings, batter flour, coatings; soup, sauces, baby food and thickeners for the industry, it sometimes could be mixed with wheat flour for baking [2; 3]. Heat treated flour was first patented, with the process using a temperature range of 100oC to 115oC for a time of 60 min [4, 5]. Heat treated flour allows high ratio recipes to be developed which generate products with longer shelf life, finer texture, moist crumb and sweeter taste.

The mechanism by which heat treatment improves the flour is not fully understood, but it is known that during the heat treatment process, protein denaturation and partial gelatinization of the starch granules occurs, as well as an increase in batter viscosity [4].

            In the production of dried foodstuffs, minimizing drying cost is the most critical factor. However, the conditions which produce minimum costs are likely not to give the desired quality characteristics [6]. It has been reported that many industrial drying processes were found experimentally, but their validity has not been evaluated since. Optimization of the drying operation leads to an improvement in the quality of the dried product and a reduction in the cost of processing as well as optimizing the throughput [6, 7]. This is because the design and optimization of dryers used for food crops is dependent on the thermal and physical properties of the specific crops [6]. The capacity to preserve food is directly related to the level of technological development. The slow progress in upgrading traditional food processing and preservation techniques in West Africa contributes to food and nutrition insecurity in the sub-region [8].

            Optimization of the drying operation leads to an improvement in the quality of the dried product, a reduction in the cost of processing as well as optimizing the throughput [9]. This is because the design and optimization of dryers used for food crops is dependent on the thermal and physical properties of the specific crops [10]. An optimization problem requires the determination of the optimal (maximum and minimum) values of a given function called the objective function under a given set of constraints. One of the most widely used optimization techniques is the regression analysis [11].

            Therefore, it is important to optimize the process parameters with reference to the physicochemical and quality characteristics of plantain. Also, according to [6], drying temperature and drying time have significant effect on the drying kinetics and quality of the dried product. Therefore, the processing parameters considered in this investigation were the drying temperature and drying time.

The objective of this work is to optimize the parameters of plantain drying process using response surface methodology (RSM).

 

Material and method

 

Preparation of plantain samples      

 

            A bunch of French horn specie of matured unripe plantain produced by the Faculty of Agriculture, University of Uyo, Uyo, Akwa Ibom State was obtained and used for this study.

The unripe plantain bunch was collected and prepared the same day it was harvested. Plantain flour was produced from the plantain fingers using the traditional method of processing plantain into flour Plantain was immersed in a plastic bowl, washed with potable water to remove dirt and then peeled and sliced into thin thicknesses (3 mm) using a stainless kitchen knife, meter rule, and a cylindrical mould to achieve approximately equal surface area for uniform heat transfer. The thicknesses of sliced samples were measured using a Vernier slide Caliper to confirm the actual thickness of the pulps which were then oven-dried at different temperatures and later ground into flour.

 

Determination of moisture content

 

            A clean crucible was dried to a constant weight in air oven at 110°C, cooled in a desiccator and Weighed.

Two grams of finely ground sample was accurately weighed into the previously labelled crucible and reweighed. The crucible containing the sample was dried in an oven to constant Weight [12]. The percentage moisture content was calculated thus Eq. (1):

                                                                      (1)

Where: W1 - weight of empty crucible; W2 - weight of crucible plus sample before drying; W3 - weight of crucible plus sample after drying.

           

Determination of ash content by furnace method

 

The porcelain crucible was dried in an oven at 100°C for 10 min, cooled in a desiccator and Weighed. Two grams of the finely ground sample was placed into a previously weighed porcelain crucible and reweighed; it was first ignited and then transferred into a furnace which was set at 550°C. The sample was left in the furnace for eight hours to ensure proper ‘ashing’. The crucible containing the ash was then removed; cooled in a desiccator and weighed [12]. The percentage ash content was calculated as follows Eq. (2):

                                                                   (2)

Where: W1 - weight of empty crucible; W2 - weight of crucible plus sample before ‘ashing’; W3 - weight of crucible plus sample after ‘ashing’.

                                              

Determination of crude fibre content

 

The sample (2 g) was weighed into a round bottom flask, 100 cm3 of 0.25 M sulphuric acid solutions were added and the mixture boiled under reflux for 30 min.

The hot solution was quickly filtered under suction. The insoluble matter was washed several times with hot water until it was acid free. It was quantitatively transferred into the flask and 100 cm3 of hot 0.31 M Sodium Hydroxide solution was added, the mixture boiled under reflux for 30 min and was filtered under suction.

The residue was then washed with boiling water until it was base free, dried to constant weight in an oven at 100°C, cooled in a desiccator and weighed. The weighed sample was then incinerated in a muffle furnace at 550°C for 2 h, cooled in a desiccator and reweighed [13].

Calculation by Eq. (3) use the loss in weight on incineration (C1-C2):

                                      (3)    

Where: C1 - weight of desiccator before heating; C2 - weight of desiccator after heating

Determination of lipid content using Soxhlet extraction method

 

A clean, dried 500 cm3 round bottom flask containing few anti-bumping granules was Weighed with 300 cm3 petroleum ether (40 - 60°C) for extraction poured into the flask filled with Soxhlet extraction unit. The extractor thimble weighing twenty grams was fixed into the Soxhlet unit. The round bottom flask and a condenser were connected to the Soxhlet extractor and cold water circulation was connected and put on.

The heating mantle was switched on and the heating rate adjusted until the solvent was refluxing at a steady rate. Extraction was carried out for 6 h. The solvent was recovered and the oil dried in an oven set at 70°C for 1 h. The round bottom flask and oil was then Weighed.

The lipid content was calculated thus Eq. (4):

             (4)

 

Determination of nitrogen and crude protein content using Kjeldahl method

 

1 g of the sample was weighed into a standard 250ml Kjeldahl flask containing 1.5 g CuSO4 and 1.5 g of Na2SO4 as catalyst and 5ml concentrated H2SO4. The Kjeldahl flask was placed on a heating mantle and was heated gently to prevent frothing for some hours until a clear bluish solution was obtained. The digested solution was allowed to cool and transferred to 100 ml standard flask and made up with distilled water. 20 ml portion of the digest was pipetted into a semi micro Kjeldahl distillation apparatus and treated with equal volume of 40% Sodium hydroxide solution. The ammonia evolved was steam distilled into a 100 ml conical flask containing 10ml solution of saturated boric acid to which 2 drops Tashirus indicator (double indicator) had been added.

The tip of the condenser was rinsed with distilled water and then titrated with 0.1 M Hydrochloric acid until a purple-pink end point was observed. The crude protein was obtained by multiplying the %Nitrogen content by a factor of 6.25 [13]. Calculations by Eq. (5):

Crude %Protein = %Nitrogen x factor

 X 6.25                                                       (5)                                          

Determination of carbohydrate

 

The carbohydrate was determined by difference i.e. the sum of the percentage moisture, ash, crude lipid, crude protein and crude fibre was subtracted from the total dry matter [14]. Calculation by Eq. (6):

             (6)

 

Oven drying process

 

Drying experiments were carried out in the oven which was preheated for 90 minutes to reach the steady state set drying air temperature conditions of 60oC, 70oC, and 80oC for the study. The plantain samples with pulp size of 3 mm were placed in the oven (model DHG 9101 S.N) tray. The sliced plantain samples were allowed to dry to attain the time required to reach the recommended microbiologically shelf-stable product of 0.10 (kg water/kg dry matter) [15]. All experiments were carried out in triplicates.

 

Stages of plantain drying process

 

Matured plantain bunch was harvested, washed, peeled and cut into required sizes before it was later dried at various oven temperatures as can be seen in Figure 1.

Figure 1. Flow chart for drying of plantain

 

 

                        Data analysis

 

Response Surface Methodology (RSM) techniques were employed for analysis using Design Expert software. The dependent variable was the carbohydrate content (%) of the flour and the independent variables were the drying temperature (0C) and the drying time (min). According to [16-17], Response Surface Methodology (RSM) is one of the experimental designing methods which can surmount the limitations of conventional methods collectively and has an advantage that it reduces the number of experimental trials needed to evaluate multiple parameters and their interactions.

Analysis of variance (ANOVA) was employed to determine the significance of the relationship between the independent and dependent variables, for which was judged by the coefficient of determination (R-squared and adjusted R-squared) and p-value [18]. The model and its coefficients was considered significant when p<0.05 which indicated that the model was statistically significant at α-level of 0.05 [18]. The model was valid when the lack of fit was not significant (p>0.05) which indicated a good fit, a low coefficient of variance (C.V. % <10) which meant that the experiments perform was highly reliable, and adequate precision (Adeq Precision > 4) which indicate that the model as fitted was adequate for predicting [6].

 

Experimental design for optimization of the formulation

 

D-Optimality was used in the design of the experiment to establish the effect of the process parameters on the response [16]. 

 

Results and discussion

 

The proximate composition of the plantain sample studied at different temperatures of 60, 70 and 80oC were obtained as presented in Table 1.

 

Table 1. Proximate composition of plantain sample at the different drying temperatures 

Nutrient contents (%)

Temperature (0C)

Sample A (600C)

Sample B (700C)

Sample C (800C)

Protein

4.75%

3.70%

3.56%

Fibre

8.70%

10.10%

10.34%

Fat

0.89%

0.92%

1.56%

Ash

1.20%

4.60%

5.50%

Carbohydrate

67.85%

68.20%

71.20%

 

From Table 1, it can be seen that carbohydrate has the highest percent in all the samples and the amount of carbohydrate increases as the temperature increases. Also, the amount of fat, fibre, and ash increases as the temperature increases which is in accordance with the results in literature [19]. On the other hand, the protein and moisture content of the plantain flour decreases with increase in the drying temperatures as shown in the Table 1 above which also agrees with the results in literature [20].

According to [20], increase in the macro nutrients i.e. fat, fibre, ash, and carbohydrate may be attributed to the application of heat. It can be seen that the apparent increase in carbohydrate, ash, fat, and fibre contents of the plantain flour at the different drying temperatures in this study could be as a result of the removal of moisture which tends to increase the amount of the individual nutrients. Application of heat can be both beneficial and detrimental to nutrients. Heat improves the digestibility of food, promotes palatability and also improves the keeping quality of food, making them safe to eat [21]. However heating process may result in nutrients’ losses by inducing biochemical and nutritional variation in food composition [21].

Table 2 shows the drying times required to reach the recommended microbiologically shelf-stable product of 0.10 (kg water/kg dry matter).

 

Table 2. Drying times of samples required to reach shelf-stable

Sample

A

B

C

Temperature (oC)

60

70

80

Drying time (minutes)

311

242

187

 

We observed in the Table 2 which is the time it took the moisture content of the dried plantain to reduce to approximately 10% of the initial content were 311, 242 and 187 minutes for 60, 70 and 800C respectively. The pulp size of 3 mm was used which is in line with previous researchers who reported that as thickness increases, the diffusion path, becomes longer and vice versa [22, 23, 24].

 

 

                        Optimization of the process parameters with carbohydrate content

 

The D-Optimal design which was employed to correlate the experimental relationship between the response and the independent variables is seen in Table 3 below.

 

Table 3. Design of experiment

Ctr. No.

FACTOR 1

A: Temperature (0C)

FACTOR 2

B: Time (min.)

RESPONSE 1

Carbohydrate (%db)

1

60.0

187.0

82.9

2

70.7

187.0

85.7

3

60.0

311.0

83.5

4

60.0

311.0

84.9

5

80.0

244.8

84.3

6

60.0

187.0

82.0

7

80,0

244.8

86.9

8

69.1

255.2

84.7

9

75.2

278.1

82.7

10

80.0

311.0

71.1

11

64.0

223.8

82.0

12

80.0

187.0

84.2

13

60.1

263.9

81.0

14

80.0

311.0

81.1

15

68.5

311.0

84.0

16

80.0

187.0

85.7

17

74.1

223.8

85.6

 

The G Efficiency = 78.1% indicated that the experimental design was efficient for analysis.  The Table 3 shows the results of the dependent variable (carbohydrate) which were obtained at thickness of 3 mm and at the shown drying times and temperatures.

In the Table 4a we have the ANOVA response surface model.

 

Table 4a. RESPONSE 1 carbohydrate. ANOVA for response surface 2FI model - Analysis of variance table (Partial sum of squares – Type III)

Source

Sum of

squares

df

Mean square

F - value

p – value

Prob >F

Block

22.82

1

22.82

 

 

 

Model

82.62

3

28.87

3.93

0.0363

Significant

A- temperature

0.30

1

0.30

0.041

0.8433

 

B- Time

24.61

1

24.61

3.35

0.0920

 

AB

61.72

1

61.72

8.41

0.0133

 

Residual

88.08

12

7.34

 

 

 

Lack of fit

32.47

7

4.64

0.42

0.0073

not significant

Pure Error

55.61

5

11.12

 

 

 

Cor Total

197.52

16

 

 

 

 

 

In table 4b we have the ANOVA response surface 2FI model.

Table 4b. ANOVA for response surface 2FI model

Indicator

Value

R-Square

0.4958

C.V. %

3.26

Adj R-Square

0.3698

Adeq Precision

5.808

 

From Tables 4a-b, it was observed that the model for carbohydrate content was significant (R-Square = 0.4958 and p<0.05). The model terms (x and y) were also significant which implied that the process parameters had significant effects on the carbohydrate content of the plantain flour. Adequate precision measures the signal to noise ratio. A ratio greater than 4 is desirable. Adequate precision = 1.808 greater than 4, indicated that the model as fitted was adequate for predicting and can be used to navigate the design space.

The lack of fit on the other hand was not significant (Prob>F) implying that the model was valid. A low coefficient of variance (C.V. % = 3.26) indicated that the experiments were highly reliable.

Figure 2 above is the three dimensional graph that visualizes the graphical relationship between the various process parameters and the response.

Figure 2. The effect of the process parameter on the response

It entails the level of interaction which existed between the process parameters i.e. B: time and A: temperature and their possible effect on the response. 3D plot of AB interaction show that 2FIS allow for twisting of the plane, but did not allow for hills and depressions [25].

Validation of the optimum process parameters

           

Since our main objective here was to maximize the carbohydrate content of the plantain flour and minimize the drying time and temperature, the criteria was to keep “in range” the drying temperature and time while the carbohydrate content was “maximized” in our targeted goal as shown in Table 5.

 

Table 5. Criteria for optimization of the process parameters (constraint)

Name

Goal

Lower limit

Upper limit

x: Time

Is in range

187 min

311 min

y: Temperature

Is in range

60oC

80oC

z: Carbohydrate

Maximize

71.13

100

 

From the result in Table 5, it shows the criteria for the process parameters where time (lower limit was 187 min and upper limit was 311 min); Temperature (lower limit was 60oC and upper limit was 80oC) and carbohydrate (lower limit was 71.13 and upper limit was 100).

 

Table 6. Solution for optimization of the process parameters

Number

Temperature

(oC)

Time

(min)

Carbohydrate

(%)

Desirability

1

80.000

187.000

87.374

0.563

2

79.666

187.000

87.277

0.559

3

80.000

189.033

87.234

0.558

4

80.000

192.790

86.976

0.549

5

74.770

187.000

85.854

0.510

6

60.000

311.000

83.983

0.445

 

From the optimizations shown in Table 6, it can be seen that the optimum processing parameters for plantain flour were: drying temperature (800C), drying time (187 min) carbohydrate content (87.374 %), pulp thickness (3 mm).

 

Conclusion

 

            This study has analysed some nutritional compositions of unripe plantain at different drying temperatures. The plantain sample studied had comparable proximate compositions to what is reported in the literature. From the proximate analysis, moisture content ranged from 7.84 - 16.60%, ash content from 1.2 - 5.5%, fibre content from 8.70 - 10.34%, fat content from 0.89 - 1.56%, protein content from 3.56 - 4.75%, and carbohydrate content from 67.85 -71.20%.  The dried plantain may find application in food industry due to low moisture content as ingredient for foods that require good shelf life.

            Response surface methodology was successfully used to investigate the effects of temperature and time on the final carbohydrate content of the plantain using oven drying method. Response Surface Methodology (RSM) was used to generate the design and optimize the drying conditions for plantain. The generated experimental design was considered to be efficiently applicable when the G Efficiency was greater than 60%. The optimum drying conditions were obtained at temperature of 80oC, time of 187 minutes. The optimum carbohydrate content was 87.4% db. Also, the model and its coefficients were found to be significant through the response analysis.

 

Acknowledgement

 

We thank the university of Uyo management and authority Department of Chemical and Petroleum Engineering for giving us the opportunity to carry out this research.

 

References

 

1.   Ogunlakin G.O., Oke M.O., Babarinde G.O., Olatunde D.G., Effect of drying  methods on proximate composition and  physic-chemical properties of Cocoyam Flour. American Journal of Food Technology, 2012, 7, p. 245-250.

2. Kiin-Kabari D.B., Banigo E.B., Quality characteristics of cakes prepared from wheat and unripe plantain flour blends enriched with Bambara groundnut protein concentrate,  European Journal of Food Science and Technology, 2015, 3 (3), p. 1-10.

3.   Ibeanu V., Onyechi U., Ani P., Ohia C., Composition and sensory properties of plantain cake, African Journal of Food Science, 2016, 10 (2), p. 25 - 32.

4.   Neill G., Al-Muhtaseb A.H., Magee T.R.A., Optimisation of time/temperature treatment, for heat treated soft wheat flour, Journal of Food Engineering, 2012.

5.   Russo J.V., Doe C.A., Heat treatment of flours as an alternative to chlorination, Journal of Food Engineering, 1970, 5, p. 363 – 374.

6    Kawongolo J.B., Muranga F.I., Optimization of processing technology for commercial drying of bananas (Matooke), 2013.

7.   Sturm B., Werner C. H. and Oliver H., Optimizing the drying parameters for hot-air-dried apples. Drying Technology. An International Journal, 2012, 30, (14), 1570-1582.

8.   Aworh O.C., The role of traditional food processing technologies in national development: the West African experience, Using Food Science and Technology to Improve Nutrition and Promote National Development. Robertson, G. L. & Lupien, J. R. (Eds). International Union of Food Science & Technology, 2008, p. 1 – 18.

9.   Madamba P.S., Driscoll R.H., Buckle K.A. Shrinkage, drying and porosity of garlic during drying, Journal of Food Engineering, 1994, 23, 309-319.

10. Rasuoli M., Seiiedlou S., Ghasemzadeh H. R., Nalbandi H., Influence of drying conditions on the effective moisture of diffusivity and energy of activation during hot air drying of garlic, Australian Journal of Agricultural Engineering, 2011.2(4):96-101.

11. Olaoye J.O., Oyewole S.N., Optimization of some “poundo” yam production parameters, Agric Eng Int: CIGR Journal, 2012, 14 (2), p. 58-67.

12. AOAC, Official methods of analysis, (15th edn) Association of Official Analytical Chemists. Arlington, V. A. USA, 1980.

13. AACC, Approved methods of the AACC, American Association of Cereal Chemists, (10th ed.). St. Paul, MN: The Associations, 2000.

14. Muller H.G., Tobin G., Nutrition and food processing, Croom Helm, London, 1980.

15. Kumari S., Kumari N.K., Jyothi J., Lavanya J.L., Swarnalatha I., Study on drying behaviour of Sapota (Manilkara Achras) in solar tray dryer and hot air cabinet dryer, IOSR Journal of Environmental Science, Toxicology and Food Technology (IOSR-JESTFT), 2016, 10 (40), p. 40 - 64 .

16. Noordin M.Y., Venkatesh V.C., Sharif S., Elting S., Abdullah A., Application of response surface methodology in describing the performance of coated carbide tools when turning AISI 1045 steel, Journal of Materials Processing Technology, 2004, 145: 46-58.

17. Ramakrishna G., Susmita M., Application of response surface methodology for optimization of Cr (111) and Cr (V1) adsorption on commercial activated carbons, Research Journal of chemical sciences, 2012, 2 (2), p. 40 – 48

18. Box G.E.P., Draper N.R., Empirical model-building and response surfaces, Wiley, New York, 1987.

19. Inana M.E, Adindu M.N, Nigeria preliminary quality evaluation of selected plantain flour (Musa Paradisiaca) sold in Port Harcourt markets, Nigeria, 2015.

20. Yarkwan B., Uvir R.H., Effects of drying methods on the nutritional composition of unripe plantain flour, Food Science and Quality Management, 2015, 41:5-10.

21. Agoreyo B.O., Akpiroroh O.A., Orukpe Osaweren O.R., Owabor C.N., The effects of various drying methods on the nutritional composition of Musa paradisiaca, Dioscorea rotundata and Colocasia esculenta,  Asian Journal of Biochemistry, 2011, p. 1-7.

22. Krokida M.K., Kiranoudis C.T., Maroulis Z.B., Marinos-Kouris D., Drying related properties of apple, Drying Technology, 2000, 18 (6):1251 – 1267.

23. Nguyen M.H., Price W.E., Air-drying of banana: Influence of experimental parameters, slab thickness, banana maturity and harvesting season, Journal of Food Engineering, 2007, 79, p. 200 – 207.

24. Islam M.S., Haque M.A., Islam M.N., Effect of drying parameters on dehydration of green banana (Musa sepientum) and its use in potato (Solanum tuberosum) chips formulation, The Agriculturists, 2012, 10 (1), p. 87 - 97.

25. Tan I. A.W., Ahmad A.C., Hameed B.H., Optimization of preparation condition for activated carbons from coconut husk using RSM, School of Chemical Engineering, University Science Malaysia, Engineering Campus, Chemical Engineering Journal, 2011.