About Bioline  All Journals  Testimonials  Membership  News  Donations

The Journal of Food Technology in Africa
Innovative Institutional Communications
ISSN: 1028-6098
Vol. 9, Num. 1, 2004, pp. 3-12

The Journal of Food Technology in Africa Vol. 9 No. 1, 2004, pp. 3-12

The Effects of Technological Modifications on the Fermentation of Borde, an Ethiopian Traditional Fermented Cereal Beverage

*Kebede Abegaz1,2, Thor Langsrud1, Fekadu Beyene2 and Judith A. Narvhus1

1 Department of Food Science, Agricultural University of Norway, P.O.Box 5036, N-1432 Ås, Norway.
Awassa College of Agriculture, Debub University, P.O.Box 5, Awassa, Ethiopia. *Corresponding author

Code Number: ft04001


Four independent experiments were carried out to study the effect of modifying some steps in the technology of the four-phase traditional borde fermentation using malt and a mixture of unmalted cereals. When maize flour was substituted for maize grits in Phase I fermentation, the titratable acidity was greater throughout this phase and decreased after 24 h. Substitution with flour resulted in a higher yield, improved acceptability and extended keeping quality of borde. In addition, the wet milling at the last stage of the process could be omitted. When Phase I was omitted from the process, the starting pH at Phase II was much higher than when fermented maize from Phase I was used. Although the pH by the end of Phase II was comparable in both treatments, the borde made using fermented maize from Phase I was superior in all sensory attributes. Unmalted ingredients were heat treated in various ways and all methods were found to produce acceptable borde. However, borde from uncooked ingredients was totally unacceptable. An investigation on the effect of merging some phases of the fermentation showed that it is possible to prepare an acceptable borde using a simplified method of production. There were no marked variations in microbial load of borde from all the above treatments. It was found possible to shorten the duration and simplify the technology of borde fermentation with some variations in acceptability.

Keywords: food processing; traditional fermentation; cereal beverage, borde; Ethiopia


Traditional fermentation processes are increasingly attracting the attention of scientists and policy makers as a vital part of food security strategies (van de Sande, 1997). Traditional methods and age-old techniques of food processing are still used in developing countries especially in communities with lowincome levels. These countries require food-processing technologies that are technologically appropriate, suitable for their regions and affordable in rural and urban economies. Household-level fermentation is one such indigenous technology that has been developed for a wide range of foods and beverages from an extensive range of raw materials. However, the transformation of home-based arts into modern industries necessitates acquisition of scientific knowledge of the raw materials and processes used (Novellie and De Schaepdrijver, 1986) so that the problems involved in scaling-up can be addressed. In many African countries, cereal-based traditional fermented products (Lorri, 1993; Steinkraus, 1996; Bvochora, 1999 et al; Mahgoub et al., 1999) are consumed both as beverages and foods. Ethiopian borde is one of those types of beverage among others. The recommended research priorities on traditional fermented foods are improving understanding of the fermentation process; refining the processes; increasing utilization of the process and developing local capabilities (BOSTID, 1992).

Borde is produced using traditional fermentation technology from a variety of locally available cereal ingredients. The unmalted cereals and the malt may be from one or a mixture of cereals. The amount, types and combinations of malt (maize, barley, wheat, finger millet) vary within and between households on the basis of availability and preferences of cereals regardless of localities (Abegaz et al., 2002).

There is only limited published information on the fermentation and microbiology of borde in southern and central Ethiopia (Ashenafi and Mehari, 1995; Bacha, 1997). Contrary to both of these reports, however, the traditional processing technology of borde in southern Ethiopia has been shown to have four major phases (Abegaz et al., 2002). This process is time-consuming and inefficient. Borde production involves grinding, fermentation, roasting, steam cooking, boiling, cooling, mashing, wetmilling and wet-sieving operations. To improve the fermentation of borde, it is necessary to undertake basic studies on its traditional processing technology and quality characteristics. To our knowledge, there is no published information on the effects of various process factors on fermentation and yield of borde.

The objective of this work was, therefore, to investigate the effect of modifying selected techniques used in the traditional production of borde in an attempt to simplify the technology and reduce loss of residues. The effects of substituting maize grits with flour, fermentation at Phase I, various methods of cooking, and merging some phases of the main fermentation at Phase II, III and IV on acceptability of the product were investigated.


Four independent experiments were carried out in duplicate and repeated three times at room temperature (21- 25&$176;C) as described in sections 2.1-2.3 below. In all the experiments, borde was prepared (at the Awassa College of Agriculture) from malt and a mixture of unmalted cereal ingredients by an experienced brewer using a modified traditional (MT) recipe. The modifications were substitutions of: 1) earthenware pot with plastic jar; 2) maize fermented for 48-72 h at Phase I with 24-48 h; and 3) 13-18% malt with 3% (Abegaz et al., 2002; Unpublished results). In addition, substitutions of: 1) flour for maize grits; 2) non-fermented flour for 24-48 h fermentation at Phase I (omission of Phase I); 3) only one or two merged phases instead of three main fermentation phases (Phase II, III and IV); and 4) the combination of roasting, steaming and boiling with only one type of heating in the whole process were carried out in the present work. The main equipment used were plastic jars, metal plate and pan, grinding stone, bowls and sieve (1 mm pore size). Samples for analysis were collected at 6 h intervals and/or the beginning and end of each phase. The microbial load, pH, titratable acidity (TA) and acceptability of borde were evaluated.

Treatment and proportion of ingredients

Borde was prepared using unmalted maize (Zea mays), sorghum (Sorghum bicolor) and mixed flour from wheat (Triticum sativum), finger millet (Eleusine coracana), tef (Eragrostis tef) and malt flour. The proportions (w/w) of unmalted ingredients used in all the experiments were, 2 maize: 2 a mixture of wheat, finger millet and tef: 1 sorghum. The heat treatments of unmalted cereal ingredients used and the preparation of malt are described by Abegaz et al. (2002). The malt was prepared from barley (Hordeum vulgare) and maize. All the unmalted ingredients were cooked at 90-98°C and cooled to 23-25°C before blending with malt flour and/or the fermenting mash at the appropriate phases of borde fermentation. Whenever flour replaced grits, roasting of enkuro (granular mass) was substituted by baking of kita (flat bread) at Phase II except experiment 2.3A below. The malt required in all treatments was calculated against the weight of grits or flour used in Phase I. Then 80% of the total required malt was added at Phase II and the rest at Phase IV. Except where otherwise stated, 3% barley malt (w/w) was used.

Phases of borde fermentation

The four phases of borde fermentation (Abegaz et al., 2002) are briefly described as follow:

Phase I
Maize grits or flour was mixed with equal amount of water (w/v) and fermented for 24-48 h before apportioning into 2:2:1 (w/w) and using each part in Phase II, III and IV respectively.

Phase II
About 40% of 24 h fermented maize from Phase I was roasted on a hot griddle into enkuro (experiment 2.3A) or baked into kita (experiments 2.3B, C and D). After cooling, the enkuro or kita was thoroughly blended with water and 80% of the total malt into a thick brown mash called tinsis. The tinsis was left to ferment for 24 h.

Phase III
A further 40% of the fermented maize from Phase I (48 h) was gently roasted or baked, kneaded with more mixed fresh flour, moulded into dough balls and then steam cooked into gafuma. The gafuma was broken into pieces, cooled and blended with the fermented tinsis into a thick brown mash called difdif. The difdif was left to ferment for 18 h.

Phase IV
The remaining 20% of fermented maize from Phase I (48 h) was mixed with sorghum and boiled for 90 min with continuous stirring into a thick porridge. After cooling, the gelled porridge was added to the fermented difdif along with the remaining 20% malt, thoroughly mixed and sieved. When grits and sorghum grains were used (experiment 2.3A), residues were repeatedly wetmilled with grinding stones and wetsieved by slurrying with water.


A. Effect of substituting flour for grits
Equivalent weights of flour were used to replace grits or grains in the MT process. An equal mixture (w/w) of barley and maize malt (18%) as per the brewer's traditional judgment was used. Both the grits (G) and flour (F) used at Phase II were roasted to enkuro.

  1. G: maize grits, sorghum grains and mixed flour + malt flour
  2. F: all in the form of flour + malt flour.

The production of borde was carried out according to the traditional process except that it was unnecessary to wet mill the fermenting mash of F at Phase IV. The fermenting mash of G was wetsieved and repeatedly wet-milled, while F was wet-sieved only once. It was found that Phase IV took four instead of the normal six hours when F was used (result not shown). The production process was therefore timed so that both G and F borde were ready for consumption at the same time. The yield of borde and the residue of G and F were compared.

B. Effect of omitting Phase I in borde fermentation
This experiment was to investigate possibilities of producing borde without Phase I. All ingredients were used as flour with 3% barley malt (unpublished results).

  1. MT: fermented flour from Phase I + traditional cooking methods (4 phases)
  2. NF: non-fermented flour + traditional cooking methods (3 phases)
  3. NFP: non-fermented flour + all unmalted ingredients cooked as porridge (3 phases)

NF and NFP fermentations were initiated together with that of Phase II for the MT. The malt flour and cooked ingredients were blended with water at this step of each treatment. All the treatments were sieved once and the filtrate was left for 4 h fermentation at Phase IV.

C. Effect of merging some phases of borde fermentation
This experiment was designed to investigate the effects of merging Phase III and IV (M1); Phase II, III and IV (M2 and M3); or Phase II and III (M4) using 24 and/or 48 h fermented maize flour and a mixture of fresh flour. These treatments were compared with the MT recipe. All ingredients were used as flour with 3% barley malt. The cooking methods and proportion of ingredients used in the merging phases were identical to each representative phase as used in the four-phase MT procedure.

  1. MT: 24 and 48 h fermented maize from Phase I + fresh mixed flour (4 phases)
  2. M1: 24 and 48 h fermented maize from Phase I + fresh mixed flour (3 phases)
  3. M2: 48 h fermented maize from Phase I + fresh mixed flour (2 phases)
  4. M3: 24 h fermented maize from Phase I + fresh mixed flour (2 phases)
  5. M4: 24 h fermented maize from Phase I + fresh mixed flour (3 phases)

Figures 1a, b, c and d show the flow diagram of MT, M1, M2/M3 and M4 respectively. Finally, each treatment was wet-sieved and then the filtrate was left for 4 h fermentation.

D. Effect of using different cooking methods in borde production
In order to select a cooking method that would ease the traditional technology of borde production, baking, steam cooking and boiling were used singly for the whole cooking process and compared with their combined use in the MT procedure. Traditionally, roasting, steam cooking and boiling were used at Phase II, III and IV respectively. However, baking substituted roasting in the MT procedure due to the use of flour. Noncooked ingredients were used as a control. Otherwise, unmalted flour and 3% barley malt were used in all treatments.

  1. MT: modified traditional cooking (baking, steam cooking and boiling),
  2. B: all ingredients baked into kita (flat bread),
  3. S: all ingredients steam cooked to gafuma (steamed dough balls),
  4. P: all ingredients boiled to a thick porridge,
  5. NC: ingredients not cooked.

After cooling, the ingredients were blended with malt and/or fermenting mash at the appropriate phases and finally, the filtrate was fermented for 4 h at Phase IV.


pH and titratable acidity (TA)
The pH was measured using a digital pH meter (ORION 420A, Boston, USA) after calibration at 25°C using buffers of pH 4 and 7 (Merck KGaA, 64271 Darmstadt, Germany). TA was determined on 5 ml samples by titration using 0.1 N NaOH and 0.1 ml 0.5% phenolphthalein as indicator. The amount of acid produced was expressed as percent lactic acid.

Microbiological analysis
Ten g sample was transferred aseptically to a Stomacher bag (Lab-Blender 400, Seward Medical, London, England) with 90 mL sterile 0.1% peptone water (Merck) and homogenized for 30 s at 'normal' speed. The homogenate was then serially diluted and aliquots of 0.1 mL from appropriate dilutions were spread-plated in duplicate on pre-dried plates of violet red bile dextrose (VRBD), plate count agar (PCA) and yeast extract glucose chloramphenicol (YGC) agar. The VRBD and PCA were from Merck. The YGC consisted of (gram L-1): yeast extract, 5.0; glucose, 20.0; chloramphenicol, 0.1; bromophenol blue, 0.01; agar, 15; pH, 6.0 to 6.2. Purple-red colonies on VRBD agar plates were counted as Enterobacteriaceae (EB) after incubation at 30°C for 24 h. The total aerobic mesophilic count (AMC) was enumerated on PCA plates after incubation at 30°C for 48 h. Yeast and mould colonies were counted on YGC plates after incubation at 28°C for 3 to 5 days. The numbers of EB, AMC or yeast on duplicate countable plates are reported as log CFU g-1 calculated from the mean of three replicates.

Sensory evaluation
A consumer-oriented panel of judges was used to assess the quality of borde produced during this study. Five judges who regularly consume borde and who are similar to the target population of consumers were selected from 21 volunteers. The judges were selected on the basis of perception of the sensory attributes of borde developed from analysis of earlier survey data from six localities (Abegaz et al., 2002). The sensory attributes are: appearance (uniformly turbid), foaming activity, thickness, aroma (degree of freshness), taste (degree of sweet-sourness) and smoothness (mouth feel). Samples (50 mL) of borde were coded with random numbers and presented in 100 mL beakers to the judge. The panelists were asked to evaluate their acceptance of each borde and to score from 1 (least-) to 5 (most- acceptable).

Statistical analysis
The average values of pH, TA, sensory scores and microbial load of the samples from triplicates of independent experiments were reported after performing ANOVA using Minitab Release 13.1 (Minitab Inc.) statistical Programme at the 5% level of significance. The five-judge sensory scores of each attribute were compared between treatments in each experiment.


A. Effect of substituting flour for grits in borde production
The TA in Phase I before cooking was found to develop differently according to whether maize G or F was used (Fig. 2). When F was used, the TA was higher but declined after 24 h. The TA was low and still slowly increasing after 24 h in case of G. The pH of F (4.0±0.03) was significantly lower (p<0.05) than G fermentation (4.2±0.07) after 24 h at Phase I just before cooking at Phase II. This difference could be due to the fact that flour has larger surface area that would provide more available soluble substrate thereby accelerating the fermentation. The rate of hydrolysis of starch is dependent on the accessible surface area (Aggarwal and Dollimore, 1998). In a study on kenkey (Nche et al., 1996), dry-milled maize was reported to have a fast water uptake and high endogenous enzyme activity at 4 and 25°C and then a rapid swelling of the maize starch component on heating to 95°C after 24 h fermentation. In our laboratory, higher amounts of soluble carbohydrates were observed during flour fermentation than grits (data not shown). Higher concentrations of total soluble and fermentable sugars were reported in fine grits than coarse grits (Lasekan et al., 1995).

When cooled enkuro was blended with malt flour in Phase II (Fig. 2), the starting pH (4.3±0.1) of F fermentation was thereby lower (p<0.05) than the G (4.6±0.2). During Phase II, the decrease in pH of F fermentation was not as rapid as when G were used. This lower initial pH and extensive utilization of carbohydrates in Phase I are probably the cause of the retarded pH reduction during F fermentation in Phase II. The lower pH would result in slow malt amylase activity, thus producing less fermentable carbohydrates for the microorganisms. Carbohydrate degrading enzymes in malted and unmalted finger millet have an optimal range of pH 4.6- 6 (Nirmala et al., 2000). However, slower hydrolysis of starch occurred at low pH and then production of reducing sugar decreased gradually (Syu and Chen, 1997). The fermentation of G progressed to a lower pH 3.8 compared to pH 4.0 in that of the F at the end of Phase II. Although the grits would be expected to have less soluble carbohydrates in Phase I, more substrate may become available after cooking in Phase II. The fermentation of F showed the highest pH from 18 h at Phase II to the end of Phase IV and the highest TA throughout borde fermentation. The reduction of TA after 24 h F fermentation in Phase I and the low TA in G contrary to its low pH after 18 h in Phase II onwards need further study. The high pH at the start of G in Phase II (Fig. 2) and M3/M4 in Phase III (Fig. 4), owing to less acid production in Phase I, resulted in a faster reduction of pH than that of F and M2 respectively. However, the carried-over effect of acid from preceding to the succeeding phases would be possible to optimise by monitoring the progress of pH and TA in each phase.

The temperature of fermenting mash increased from 23.2±0.2 to 29±1.2°C with no significant difference between F and G fermentation. Both reached maximum at about 12 h in Phase III.

Both F and G resulted in actively fermenting acceptable borde. However, borde made from the F was preferred and had a significantly better (p<0.05) aroma and taste (Table 1A). It was also observed that the keeping quality of borde from F was at least 4 h longer than borde from G, which soon developed a vinegary aroma (results not shown). The repeated wet milling and wet sieving of intact grits at Phase IV could create additional surface area of the substrate and also allow microbial contamination. This may cause secondary fermentation and result in production of acetic and butyric acids that are detrimental to flavour (van der Merwe et al., 1964/ 65). Acetic acid bacteria commonly occurred with the most visible characteristics of vinegary flavour and off-odour in deteriorating traditional fermented cereal beverages (Sanni et al., 1999). However, the mean counts (log CFU mL-1) of AMC and yeast in borde from F (10.6±0.3 and 8.5±0.5) were not significantly different (NSD) from that of in G borde (10.3±0.5 and 7.9±0.6). EB were not detected in both cases, which would be due to the acidic fermentation. The new ingredients added at different phases may reduce the acidic stress and also replenish substrates for microbial growth in the fermenting mash of borde.

The substitution of grits with flour considerably improved the efficiency of borde production since the problems of unhygienic and tedious wet-milling, large loss of residue and low profit were resolved. The spent residue from G was 3.7±0.0.16 kg compared to 0.3±0.05 kg in F fermentation. Thus, the net yield (recovery) of borde increased from 70% to 97%. Improvements in aroma, taste and keeping quality were observed in borde made from flour.

B. Effect of omitting Phase I in borde fermentation
The development of pH and TA during Phase I fermentation of MT is shown in Fig. 3. The fermentation of NF and NFP were started together with Phase II of the MT. Regardless of the different cooking methods used, the fermentation of NF and NFP showed the same development in pH. When nonfermented flour was used at the start of NF and NFP fermentation, the initial pH was higher (p<0.05), but was significantly lower than MT at the end of Phase II. The same pattern was observed for Phase III, but the difference was much less pronounced. This may be due to greater amylase activity at higher pH and thereby a greater concentration of available substrate for fermenting organisms in NF and NFP than in the MT. The lower initial pH and retarded fermentation of the MT can be considered as the carried over effects of acidic maize from Phase I. Conversely, NF and NFP would possibly benefited from unexploited substrate (non-fermented maize) within the range of optimal pH 4.6-6 for carbohydrate degrading enzymes (Nirmala et al, 2000) and pH 5.5 for most of bacterial a-amylases (Aguilar et al, 2000). Syu and Chen (1997) reported that the rate of starch hydrolysis is governed by initial substrate concentration and enzymatic activity and also showed that low pH slows down the hydrolysis of starch and accumulation of reducing sugars for microbial growth. The amount of fermented maize was small as compared to the amount of fresh flour added in the MT at Phase IV and then the pH of borde was identical for the three treatments.

No significant differences (p<0.05) were found in microbial load, pH and TA of borde regardless of the methods of cooking and whether fermented or nonfermented flour was used. The AMC was 10-10.5, while yeast increased to 8.2- 8.5 (log CFU mL-1) in all treatments. All sensory attributes achieved significantly lower scores when maize was not fermented in Phase I before cooking (Table 1B). When borde produced using the two methods of ingredient cooking were compared (NF and NFP), NFP borde achieved significantly lower scores (p<0.05) for foaming and consistency. This could be a compounded effect of using non-fermented flour and cooking into porridge. The results show that the fermentation at Phase I is important for the sensory quality of borde. In the traditional fermentation, Phase II and IV are initiated by the addition of malt, fermented and non-fermented ingredients to the pot. The results show the magnitude of the effect of adding fermented ingredients on the initial pH. This is particularly marked at the start of Phase II, where the initial pH differed by 1.6 units between treatments. It may be expected that this difference in pH would have a far-reaching effect on the development of the microbial flora at this stage. At the starting pH 4.4, as observed for the MT, only aciduric organisms such as lactic acid bacteria and yeasts would be able to grow. In Phase III and IV, the drop in pH is very small since the initial pH is already so low that few microorganisms would be able to produce additional acid. The effect of omitting Phase I was much greater than using different cooking methods on the fermentation and quality of borde. However, borde produced by all the three treatments was acceptable.

C. Effect of merging some phases in borde fermentation C.1 Merging of Phase III and IV
After 24-48 h fermentation of maize in Phase I, a significant decrease in pH and increase in TA are shown in Fig. 4. The TA of 24 and 48 h fermented maize (FM) was NSD (p<0.05) one from another in contrast to their pH. After cooking the FM (pH 4.07±0.06) from Phase I and blending it with malt, the MT and M1 treatments showed low initial pH 4.2±0.13 and slow development of pH in Phase II. The addition of 48 h FM (pH 3.66±0.04) from Phase I to MT and M1 resulted in a significantly lower (p<0.05) initial pH 4.7 as compare to pH 5.0 in treatments of M3 and M4 where 24 h FM (pH 4.07±0.06) was added at Phase III. The low pH achieved in FM from Phase I affected further production of acid in the subsequent phases of borde fermentation. The acid and high number of acid tolerant microorganisms from Phase II (tinsis) (unpublished results) could be responsible for the low pH at Phase III and IV fermentation of MT and M1 compared to other treatments.

C.2 Merging of Phase II, III and IV or II and III
The fermentation of maize for 24 h in Phase I showed a significant decrease in pH and increase in TA (Fig. 4). This FM (pH 4.07±0.06) was used with mixed fresh flour in treatments of M3 and M4 (Figs. 1c and d). Conversely, 48 h FM (pH 3.66±0.04) was used for M2. After cooking the unmalted ingredients and blending with malt, M3 and M4 showed higher (p<0.05) initial pH 5 than M2 (pH 4.7). However, the main fermentation of M2, M3 and M4 (Phase II) was started concurrently with Phase III of MT and M1. M2 had the slowest development of pH among all treatments under investigation. After 18 h fermentation, M2 showed higher pH (p<0.05) than MT and M1. This is not surprising since M2 is started with high quantities of acidic FM (48 h) from Phase I and unmalted new ingredients added together in the absence of Phase II for starter building up (tinsis). Thus, too small amount of malt inoculum comparing to the cooked ingredients with low initial pH may explain the slowest fermentation of M2. The same amount of malt and absence of Phase II also occurred in M3 and M4. However, the M3 and M4 started with higher pH due to less acidic FM from Phase I. Thus, M3 and M4 could be benefited from positive effects of higher initial pH on amylase activity that would produce greater amount of sugars for the growth of fermenting organisms. Increasing the amount of malt (inoculum) may also accelerate the fermentation of borde in M2, M3 and M4.

The effect of high inoculum from Phase II on the main fermentation in the consecutive phases was pronounced. This may explain the leading progress of MT and M1 in comparison to that of M2, M3 and M4 fermentation. The malt and porridge added to MT and M4 at the final phase resulted in higher pH as compared to M1 and M3 respectively. After sieving and then fermenting the filtrate for 4 h, the pH in M3 (3.97±0.02) borde was NSD (p<0.05) from MT (3.94±0.08) in contrast to M1 (3.77±0.1) and M4 (4.09±0.04). The pH of M2 was 4.06±0.06. The TA of MT (0.44±0.1), M1 (0.44±0.14), M2 (0.6±0.16), M3 (0.59±0.29) and M4 (0.57±0.15) borde were NSD (p<0.05). In all the treatments, AMC and yeasts were in the ranges of 9.9-10.5 and 7.4-8.4 (log CFU mL-1), respectively. Despite the low pH (3.97-4.09) in "young" borde, the incidence of EB (<1-2.7 log CFU mL-1) in 66% of M3 and M4 samples may raise the issue of safety. It is possible that a slightly longer fermentation would improve the microbial safety and ripening of borde.

The results of the sensory assessment showed that the MT and M1 treatments resulted in borde that were NSD except from a reduced foaming in M1 (Table 1C). This indicates that the addition of a small amount of malt and more ingredients (porridge) at Phase IV is responsible for the short active fermentation that revitalized the basic sensory attributes of borde. The MT borde was also judged as sweet-sour compared to sourer taste of M1. The other treatments produced a significantly inferior borde, although all products were acceptable (score >3). The foaming, aroma and taste of M2, M3 and M4 borde were judged to be markedly inferior to MT and M1. However, these products were described by the judges to be having an acceptable taste but were "too young". After extra 3-4 h fermentation, it was observed that M3 and M4 achieved all the sensory attributes of borde in contrast to M2. The low pH of 48 h FM from Phase I negatively affected the quality of M2 borde. The M3 was significantly better (p<0.05) than M4 in foaming and texture. It may be possible to improve the quality M3 and M4 borde by slightly extending the final fermentation stage and/or increasing the amount of malt inoculum. M3 and M4 had a total fermentation time of 46 h compared to 70 h for other three treatments (Figs. 1a, b, c, d). Among all these options, M3 could be suggested as a simplified technology for production of acceptable borde that may help for modernization. Further work should be done to improve these methods. Optimisation of the pH in Phase I, amount of inoculum and duration of the main fermentation in subsequent phases appeared to be vital for production of quality borde using these methods.

These results show that it is possible to shorten and simplify the traditional production technology of borde by merging some of the phases, and produce an acceptable borde. The 96±2 h (4 phases) traditional fermentation (unpublished results) could be reduced to 46 and 70 h (2-4 phases) as shown in Figs. 1a, b, c, d. Scientific improvements of traditional product processing have often led to changes in quality characteristics especially organoleptic (Demuyakor and Ohta, 1993). However, a possible method to overcome changes in product quality is to understand the process and formulation required to improve the traditional product using participatory approach as was tried in this work.

D. Effect of using different cooking methods in borde production
As shown in Fig. 5, the pH was consistently lower in NC (non-cooked) than the treatments of MT, B, S or P with cooked ingredients. After 24 h in Phase II, NC showed significantly lower pH than all other treatments. The NC was NSD from other treatments at Phase III, but it was significantly lower than MT at the end of Phase IV. The TA in all cooked treatments was about 0.5±0.04, which was lower than 0.84±0.07 in the NC. The number (log CFU mL-1) of yeasts (8.3±0.15 to 8.6±0.22) and AMC (9.8±0.19 to 10.2±0.4) were lower in treatments with cooked ingredients than the yeasts (9.2±0.31) and AMC (10.6±0.03) in the NC. All treatments with cooked ingredients rely mainly on the malt enzymes and micro flora. The occurrence of low pH and high microbial load in NC could be the effects of aciduric organisms in non-cooked maize from Phase I, diversity of endogenous enzymes and micro flora from mixed fresh flour in addition to the malt.

The NC ingredients resulted in totally unacceptable product (Table 1D). However, regardless of using only one form of cooking or their combination, the aroma and taste of borde were similar to MT. The aroma, taste and also foaming are the major quality attributes of borde. The borde made using only baking, steaming or boiling of ingredients was highly acceptable since all products attained a score >4, except for foaming, which was inferior to MT. The results illustrate that although it is important for the quality of borde that the ingredients are cooked, using only one form of heat treatment could be applied for a simplification of the process. Thus, boiling of the ingredients is the simplest technological process to be suggested for borde production. It has the advantage of: (1) less charring effect and smaller loss, (2) less tedious postcooking operations, (3) being easier to blend and sieve and (4) less or no addition of water and this also reduces postcooking microbial contamination.

When all the products were heated for 30 min at 80°C water bath, the NC was baked into stiff paste, while others were remained liquid (data not shown). This could indicate that cooking is a critical parameter in the production of acceptable borde due to the fact that gelatinisation of the starch improves its degradation by endogenous enzymes mainly from the malt. All the methods of cooking used in this work (90-98°C) would gelatinise cereal starch and inactivate enzymes and vegetative cells. However, the addition of malt at Phase II serves as sources of amylases and fermenting microbes for production of acceptable borde (unpublished results).

From this study and other basic works described by Abegaz et al. (2002; unpublished results), the traditional technology of borde production from gelatinised main ingredients and malt inoculum appeared to be developed consciously in an attempt to control the safety of borde using an acidic fermentation. The rationale of this technology reveals that: (1) the acid produced in Phase I creates acidic environment to the main fermentation that inactivates unwanted micro flora and selects for aciduric organisms. (2) The cooking of unmalted ingredients gelatinises the starch and eases its saccharification with malt amylases. (3) The use of only 16% of the total unmalted ingredients and 80% of the total amount of malt required indicates the attempt to select and attain high number of fermenting microorganisms mainly from the malt in Phase II. (4) Blending the "bulk starters" from Phase II (tinsis) and 56% of the total unmalted ingredients without malt at Phase III (difdif) indicates the major acidic fermentation of borde. (5) Addition of the remaining malt and 28% of the total unmalted ingredients in the form of porridge into the sourer difdif at the start of Phase IV eases the homogenisation and sieving processes. Thus, the malt liquefies the thick mash, reduces the bulk density and sweetens the end product (borde). It seems that the complexity of traditional technology of borde production is aimed at a step-by-step controlling of the bio-physico-chemical process parameters so as to get benefit from the merits of fermentation on safety and quality of the product.


In conclusion, the traditional method for production of borde is technologically complex, which does not readily lend itself to larger production volumes or to the introduction of production equipment. From the results of the experiments reported in this study, the opportunity for the simplification of the process has been observed: (1) Flour may be substituted for grits and whole grains. This modification resulted in a well-accepted borde, a less laborious process without wet milling and a reduction in residues from the final sieving that increased the net yield of borde. In addition, the fermentation proceeded faster than when grits were used and this may necessitate a slight adjustment of the times of the various phases. (2) Merging of the phases of borde process resulted in products that were acceptable, although organoleptically slightly inferior, which may require optimisation of the pH in Phase I, amount of inoculum and time of the main fermentation. A simplification of the process entailing a reduction of the number of fermentation phases would be advantageous. However, since products with a shortened total fermentation time showed the presence of EB, it may be possible to produce microbiologically safe borde by a shortened process only if all the ingredients are heat-treated and starter cultures are added. (3) Cooking of unmalted ingredients is a limiting step than omitting Phase I fermentation in the production of borde. Acceptable borde could be prepared by boiling the unmalted ingredients added after Phase I. This method could be readily adapted to small-scale production units.


The authors acknowledge the Norwegian Universities Committee for Development Research and Education (NUFU) for sponsoring this work. This work was carried out in the context of a North-South-South (NSS) program 26/96 "Research and development of indigenous fermented foods for small scale commercial processing in East and southern Africa" in collaboration with Agricultural University of Norway, Norway and Awassa College of Agriculture, Debub University, Ethiopia. We would like to thank Mrs. Warite Alambo for the traditional brewing of borde. The assistance of Wendosen Tadesse is highly appreciated.


  • Abegaz, K., Beyene, F., Langsrud, T., Narvhus, J. A. (2002). Indigenous processing methods and raw materials of borde, an Ethiopian traditional fermented beverage. J. Food Technol. Africa. 7, in press.
  • Aggarwal, P. and Dollimore, D. (1998). A thermal analysis investigation of partially hydrolyzed starch. Thermochimica Acta 319, 17- 25.
  • Aguilar, G., Morlon-Guyot, J., Trejo-Aguilar, B., Guyot, J.P. (2000). Purification and characterization of an extracellular a-amylase produced by Lactobacillus manihotivorans LMG 18010T, an amylolytic lactic acid bacterium. Enzyme Microb. Technol. 27, 406- 413.
  • Ashenafi, M. and Mehari, T. (1995). Some microbiological and nutritional properties of Borde and Shamita, traditional Ethiopian fermented beverages. Ethiop. J. Health Dev. 9 (1): 105-110.
  • Bacha, B. (1997). The Microbial Dynamics of Borde Fermentation, a Traditional Ethiopian Fermented Beverage. SINET: Ethiop. J. Sci. 21 (2) 195-205.
  • BOSTID (1992). Research priorities in traditional fermented foods. In: Application of biotechnology to traditional fermented foods (Gaden, E.L., Bokanga, M., Harlander, S., Hesseltine, C.W., Steinkraus, K.H., eds.). Report of an Ad Hoc Panel of the Board on Science and Technology for International Development, National Research Council. National Academy Press. Washington D.C., pp 1-7.
  • Bvochora, J.M, Reed, J.D., Read, J.S., Zvauya, R (1999). Effect of fermentation processes on proanthocyanidins in sorghum during preparation of Mahewu, a non-alcoholic beverage. Process Biochem. 35, 21-25.
  • Demuyakor, B. and Ohta, Y. (1993). Characteristics of single and mixed culture fermentation of Pito Beer. J. Sci. Food Agric. 62, 401-408.
  • Lasekan, O.O., Idowu, M.A., Lasekan, W. (1995). Effect of germination and degree of grind (coarse/fine) on the extract and sugar contents of surghum malt. Food Chem. 53, 125-128.
  • Lorri, W.S.M. (1993). Nutritional and microbiological evaluation of fermented cereal weaning foods. Ph.D. Thesis. Department of Food Science, Chalmers University of Technology, Goteborg, Sweden.
  • Mahgoub, S. E.O., Ahmed, B. M., Ahmed, M. M. O., El Agib, E. A. A. (1999). Effect of traditional Sudanese processing of kisra bread and hulu-mur drink on their thiamine, riboflavin and mineral contents. Food Chem. 67, 129-133.
  • Nche, P.F., Odamtten, G.T., Nout, M.J.R. and Rombouts, F.M. (1996). Soaking of maize determines the quality of aflata for kenkey production. J. Cereal Sci. 24, 291-297.
  • Nirmala, M., Subba Rao, M.V.S.S.T., Muralikrishna, G. (2000). Carbohydrates and their degrading enzymes from native and malted finger millet (Ragi, Eleusine coracana, Indaf-15). Food Chem. 69, 175- 180.
  • Novellie, L. and De Schaeporijver, P. (1986). Modern developments in traditional African beers. Progress in Indust. Microbiol. 23, 74-157.
  • Sanni, A.I., Onilude, A.A., Fadahunsi, I.F., Afolabi, R.O. (1999). Microbial deterioration of traditional alcoholic beverages in Nigeria. Food Res. Int. 32, 163-167.
  • Steinkraus, K. H. (1996). Introduction to indigenous fermented foods. In: Handbook of indigenous fermented foods (K. H. Steinkraus, ed.). Marcel Dekker, New York, pp 1-5.
  • Syu, Mei-J. and Chen, Yu-H. (1997). A study on the a-amylase fermentation performed by Bacillus amyloliquefaciens. Chem. Eng. J. 65, 237- 247.
  • Van de Sande, T. (1997). Socio-economic pitfalls of enhancing indigenous capabilities in household fermentation. Food Control 8, 303- 310.
  • Van der Merwe, A le R., F. Schweigart and V.A. Cachia 1964 (June) and 1965 (January). In: Steinkraus, K.H. (Ed), Handbook of indigenous fermented foods, 2nd ed., Marcel Dekker, New York, pp 227-232.

Copyright 2004 - The Journal of Food Technology in Africa, Nairobi

The following images related to this document are available:

Photo images

[ft04001f1c.jpg] [ft04001f1a.jpg] [ft04001f2.jpg] [ft04001f1d.jpg] [ft04001f5.jpg] [ft04001f3.jpg] [ft04001f4.jpg] [ft04001f1b.jpg] [ft04001t1.jpg]
Home Faq Resources Email Bioline
© Bioline International, 1989 - 2018, Site last up-dated on 11-Dec-2018.
Site created and maintained by the Reference Center on Environmental Information, CRIA, Brazil
System hosted by the Internet Data Center of Rede Nacional de Ensino e Pesquisa, RNP, Brazil