On-line version ISSN 1678-457X
Ciênc. Tecnol. Aliment. vol.19 n.1 Campinas Jan./Apr. 1999
Organic acid profile of commercial sour cassava starch1
Organic acids are present in sour cassava starch ("polvilho azedo") and contribute with organoleptic and physical characteristics like aroma, flavor and the exclusive baking property, that differentiate this product from the native cassava starch. Samples of commercial sour cassava starch collected in South and Southeast Brazil were prepared for high performance liquid chromatography (HPLC) analysis. The HPLC equipment had a Biorad Aminex HPX-87H column for organic acid analysis and a refractometric detector. Analysis was carried out with 0.005M sulfuric acid as mobile phase, 0.6ml/min flow rate and column temperature of 60° C. The acids quantified were lactic (0.036 to 0.813 g/100g), acetic (0 to 0.068 g/100g), propionic (0 to 0.013 g/100g) and butyric (0 to 0.057 g/100g), that are produced during the natural fermentation of cassava starch. Results showed large variation among samples, even within the same region. Some samples exhibited high acid levels, mainly lactic acid, but in these neither propionic nor butyric acids were detected. Absence of butyric acid was not expected because this is an important component of the sour cassava starch aroma, and the lack of this acid may suggest that such samples were produced without the natural fermentation step.
Keywords: fermentation, cassava sour starch, quality.
ANÁLISE DE ÁCIDOS ORGÂNICOS EM AMOSTRAS DE POLVILHO AZEDO COMERCIAL. A presença de ácidos orgânicos no polvilho azedo, além de contribuir com aspectos como sabor e aroma, tem, conforme a literatura indica, correlação com a propriedade de expansão, que é um fator determinante no uso alimentício. Amostras de polvilho azedo foram coletadas nas Regiões Sul e Sudeste diretamente nas empresas ou no comércio. Foram preparadas para análise em cromatografia líquida de alta eficiência (CLAE), sendo que o cromatógrafo estava equipado com coluna Biorad Aminex HPX-87H para análise de ácidos orgânicos e detector refratométrico. As condições de análise envolveram o emprego da fase móvel ácido sulfúrico 0,005M, fluxo de 0,6 ml/min e temperatura da coluna de 60oC. Os ácidos quantificados foram lático (0,036 a 0,813 g/100g), acético (0 a 0,068 g/100g), propiônico (0 a 0,013 g/100g) e butírico (0 a 0,057 g/100g), presentes na fermentação natural. Os resultados revelaram grande variação entre as amostras, com diferenças mesmo dentro das Regiões. Algumas amostras apresentaram quantidades elevadas de ácidos, especialmente do ácido lático, mas nestas amostras os ácidos propiônico e butírico não foram detectados. A ausência do ácido butírico não era esperada, uma vez que esse ácido está diretamente relacionado com o aroma característico do polvilho azedo. O fato pode sugerir que a obtenção de algumas das amostras estudadas pode ter ocorrido sem o processo fermentativo natural.
Palavras-chave: fermentação, polvilho azedo, qualidade.
Sour cassava starch (polvilho azedo) is traditionally produced by natural fermentation of cassava starch. The fermentative process was studied by several researchers and is predominantly lactic [7, 20, 21, 24]; starch is the medium for bacterial growth . Since 1973 , a number of studies about the microbiology of sour starch fermentation have been published. CEREDA  divided the fermentation in three steps, not always very distinct. Microorganisms responsible for establishing the appropriate conditions for fermentation in the first step include Achromobacter, Escherichia, Pseudomonas, Alcaligenes, Bacillus and Clostridium genera. In the second step, acidogenic microorganisms that have higher nutritional requirements appear. Finally at the third step saprophytic microorganisms become present and are responsible for the degradation of some fermentation products. Examples of microorganisms of this step are the genus Bacillus and some yeasts; at this step there is generation of compounds that contribute to the typical aroma and taste of sour starch. Three types of fermentation are identified: lactic, with the presence of Lactobacillus, Leuconostoc and Pediococcus genera, propionic, where Propionibacterium genus is found and butyric, with Clostridium genus. MARTINEZ , who showed the importance of microaerophilic microorganisms that are present in all fermentations, isolated mainly the following gram positive bacteria: sporulated Bacillus, non sporulated Lactobacillus, Streptococcus and also yeasts from the Saccharomyces genus. According to MESTRES & ROUAU , the lactic bacteria population increases during the first five days of fermentation and lactic acid production continues until the 20th day. At the end of fermentation, generally 30 days, total and lactic counts are relatively low, less than 109 CFU/g  and lactic acid content is around 10g/kg .
Other researchers attempting to identify organic acids present in sour starch found lactic, acetic, propionic and butyric acids [9, 11, 14]. Formic, succinic  and valeric  acids were detected in commercial sour starch samples, as well as in laboratory produced samples, but the methodology employed for analyzing such samples was different from that used in this paper. CÁRDENAS & De BUCKLE  found lactic acid representing 66 to 82% of the sour starch total acidity, whereas CARVALHO et al.  60%, and ASTÉ  95%; acetic, propionic and butyric acids were also detected in amounts of 2.5, 1.5 and 2%, respectively. MESTRES & ROUAU , when analyzing samples produced in Colombia and employing the same methodology of this work found ca. 0.6 g of lactic acid per 100g of starch. CARVALHO et al.  reported values of 0.40 to 0.53g/100g for total acidity, expressed as lactic acid. BANGOU  studied cassava starch fermentation at two different temperatures, 20 and 35° C, for a period of 15 days and determined lactic and butyric acid contents. The results showed that at 35° C, butyric acid was the major component of the total acidity (1.38mg/g), followed by lactic acid (0.93mg/g). At 20° C, on the other hand, the result was different, 0.13mg/g of butyric acid and 2.47 mg/g of lactic acid were produced. CEREDA & CATÂNEO  had already concluded in an earlier study that at lower fermentation temperatures, and consequently lower fermentation rates, there was higher generation of lactic acid, whereas higher temperatures or higher fermentation rates, resulted in higher production of butyric acid.
In a bibliography review, FIGUEROA  showed that lactic, acetic, butyric, propionic and valeric acids were found on cassava sour starch samples and that total acidity was highly correlated with the lactic acid content. Heterogeneous samples from Colombia also exhibited lactic, acetic, propionic and butyric acids . ESCOBAR & MOLINARI  detected lactic, acetic and butyric acids and found a positive correlation between acid content and the expansion and water binding capacity.
Quality differences observed among samples commercialized in the geographic area of this study [17, 18] called our attention to the importance of determining their profile of organic acids. Sour starch baking ability is its most interesting characteristic [12, 25] that makes sour starch technologically different from natural cassava starch. The denomination "sour" is due to the natural fermentation that occurs as explained above, with the formation of organic acids. This acidity is considered by the Brazilian legislation as a criterion for differentiating between sweet or sour cassava starch .
CEREDA  evaluated 12 commercial samples from three Brazilian States (São Paulo, Paraná and Minas Gerais) and concluded that it was not possible to correlate any physico-chemical property with the place of production. The author found great variation on ash content (CV = 65.7%) and titrable acidity (CV = 68.2%).
Nowadays there is a growing market for sour starch, mainly due to the increase in the number of small industries that manufacture cheese bread dough for both home and fast food store consumption. The fermentation step for the production of cassava sour starch is a time consuming process and the substitution of this naturally produced starch for chemically modified ones may appear attractive.
The aim of this study was to provide information about the organic acid profile of sour cassava starch samples collected in South and Southeast Brazil, compare results with published data and try to group samples according to this characteristic. It must be stressed that sample origin is difficult to be ascertain because sometimes a sour starch is produced in one region and packaged and sold in another, without identification of the production state.
2 MATERIALS AND METHODS
Sour cassava starch samples (n=29) were collected on South and Southeast Brazil, directly on factories or bought on market; one sample of natural cassava starch was a gift from a factory from São Paulo state and was analyzed for comparison. All reagents employed were compatible with chromatographic analysis and were prepared for sample and standard (lactic, acetic, propionic and butyric acid, all from Merck) injections.
The HPLC equipment had a Biorad Aminex HPX-87H column (300x7,8mm), oven for column temperature control and a refractometric detector. Chromatograms were generated and integrated by using a personal computer and the software Chroma .
2.2.1 - Sample preparation
Samples weighing 15 g (DM) were suspended in 30 ml of a 0.005M sulfuric acid solution for a period of contact of 24 hours at room temperature and with occasional agitation. After this period of contact, the suspensions were centrifuged twice. Firstly at 3000 rpm for 10 minutes and then at 10000 rpm for 15 minutes.
2.2.2 - HPLC analysis
The 0.005M sulfuric acid mobile phase was filtered through a 0.22m m membrane. The flow rate was set at 0.6 ml/min and the column temperature kept at 60° C. At injection, the samples were also filtered to 0.22m m and the volume injected was of 20m l (BRABET ; BRABET & DUFOUR ).
3 RESULTS AND DISCUSSION
The total organic acid content is highly correlated with lactic acid in cassava sour starch . This fact was evident when a linear regression was made with these two values; the r2 obtained for the 30 samples of cassava starch was 0.9779, as shown on Figure 1. It is possible to conclude that lactic acid is the most important component of sour starch acidity. The presence of lactic acid on cassava natural starch (sample 11) may be explained, once it is known that the production technology allows an incipient fermentation process .
FIGURE 1. Linear regression between lactic acid content and total acidity of the cassava starch samples.
When considering lactic acid as a percentage of total acidity, it may be observed that these values varied from 38.2 to 100%. From the 30 samples, including sample 11 (natural starch), 10 samples had lactic acid as the only component of the total acidity and in 7 other samples this acid was responsible for more than 90% of the total acidity. In the other samples, lactic acid contents were from 60.5 to 89.0% and in one sample (number 22] this acid represented only 38.2% of the total acids, whereas acetic accounted for 45.2 and butyric acid for 16.6%. This sample did not contain propionic acid.
Table 1 illustrates differences among sour cassava starch samples according to their organic acid profile. When comparing with previously published data it was possible to conclude that not all samples presented the typical organic acid profile. Propionic and butyric acids were detected in only a few samples. It must be observed with attention the absence of butyric acid in some of the samples. This acid is produced during the cassava starch natural fermentation and was expected to be present [9,11,14]. Employing a different methodology, CEREDA , analyzed sour cassava starch samples from Paraná and Santa Catarina states and found butyric acid as the main organic acid, followed by lactic and acetic acids.
The predominantly lactic fermentation  seems to be initially heterofermentative, followed by a homofermentative process . MESTRES & ROUAU , analyzing samples of sour cassava starch produced in a pilot plant in Colombia and employing the same methodology of the present work, were able to detect only lactic acid and traces of acetic acid. There was one difference between those authors organic acid extraction procedure and the one employed here. While MESTRES & ROUAU  made a suspension of 6.25% (DM) sour starch, in this work a more concentrated sour starch suspension was made (33.3% DM), which made it possible to detect acids present in lower levels.
Some of the samples studied exhibited several peculiarities such as the absence of typical aroma, very low granulometry and whiter color than usual, resembling the native cassava starch. Coincidentally these samples contained only lactic acid or, in some cases, lactic plus a small amount of acetic acid.
Analysis of the organic acid profiles allowed us to make further considerations:
In nine samples lactic acid was the only component of the total acidity, in amounts from 0.112 to 0.813g/100g.
Six samples presented lactic and acetic acids as components of the total acidity. In this group lactic acid contents varied from 0.145 to 0.627 g/100g and acetic acid from 0.018 to 0.028g/100g. On a percentage basis, lactic acid was responsible for 88.9 to 96.6% and acetic acid for 3.4 to 11.1% of the total.
Nine samples did not have propionic acid as a component of their acidity and their total acid contents varied from 0.095 to 0.677 g/100g (38.2 to 93.7% for lactic, 3.9 to 45.2% for acetic and 2.4 to 16.6% for butyric acid).
Five samples presented all the acids generally found on sour starch produced by natural fermentation and the total acid contents varied from 0.173 to 0.684g/100g. On percentage basis, these values corresponded to 60.5 to 91.1% for lactic acid, 3.7 to 18.7% for acetic acid, 0.5 to 3.9% for propionic acid and 4.7 to 19.0% for butyric acid.
Table 2 summarizes sample separation and shows the region (state) where they were produced.
Observing Table 2 it was possible to conclude that (a) from the eight samples from Paraná State, only one did not present butyric acid in its composition, (b) from the 11 samples from Santa Catarina State, 6 presented the typical sour cassava starch acids, including butyric acid, whereas 5 had lactic and acetic or only lactic acid, (c) out of the 6 samples from Minas Gerais State, 5 presented lactic and acetic or only lactic acid and only one sample had butyric acid in its composition. This sample, on the other hand, had the lowest total acid content among all 29 sour starch samples studied. Total acid content of this sample was 0.095%, which represented 8.5 times less than the sample with the highest acid content, coincidentally also from Minas Gerais State and (d) all 4 samples from São Paulo State presented both lactic and acetic or only lactic acid in their compositions.
The organic acids detected and quantified were: lactic (0.013 to 0.813 g/100g), present in all 30 samples studied, acetic (0 to 0.068 g/100g) in 20 of the samples, propionic (0 to 0.013 g/100g), in five samples and butyric (0 to 0.057 g/100g), in 14 samples.
Analysis of organic acids by the methodology employed in this work may prove to be an important tool for cassava sour starch quality evaluation.
The samples of sour cassava starch commercialized in the regions considered in this study exhibited a heterogeneous organic acid profile, which suggests the necessity of improving the production process with the purpose of quality standardization.
 ASTÉ,M. Evolution des proprietés physico-chimiques de lamidon de manioc (Mcol 1522) au cours de la fermentation / Mise en évidence du rôle de lacide lactique e de lexposition au soleil sur lacquisition du pouvoir de panification de lamidon aigre. Programme CEE/STD3 "Valorisation du manioc en Amérique Latine". CIRAD-SAR/CIAT, Cali, 1994. [ Links ]
 BANGOU,O. Etude des modifications physico-chimiques et des proprietés fonctionelles de produits amylaces (mais, manioc) au cours des procedes de fermentation traditionnels. Montpellier, DESS Nutrition dans les pays en développement, CIRAD-CA-T/CER 95/2, Septembre 1995. 33p. [ Links ]
 BRABET,C. Etude des mécanismes physico-chimiques et biologiques responsables du pouvoir de panification de lamidon fermenté de manioc. Montpellier, 1994. (Ph.D. Thesis, Université de Montpellier II). [ Links ]
 BRABET,C.; DUFOUR,D. El almidón agrio de yuca. Producción y estudio de las propriedades fisicoquímicas. Simposio en carbohidratos. Escuela Politécnica Nacional de Quito, 4-6 de octubre, Quito, 1993. [ Links ]
 BRASIL. Decreto n° 12.486 de 20 de outubro de 1978. Diário Oficial do Estado de São Paulo, p. 3-25, 21/10/1978 (Normas técnicas especiais relativas a alimentos e bebidas). [ Links ]
 CÁRDENAS,O.S.; De BUCKLE, T.S. Sour cassava starch production: a preliminary study. Journal of Food Science, v. 45, 1980. p. 1509-28. [ Links ]
 CARVALHO,E.P.de; CANHOS,V.P.; VILELA,E.R.; CARVALHO,H.P.de. Polvilho azedo: aspectos físicos, químicos e microbiológicos. Pesquisa Agropecuária Brasileira, v. 31, n. 2, p. 129-37, 1996. [ Links ]
 CEREDA,M.P. Alguns aspectos sobre a fermentação da fécula de mandioca, 1973. (Tese de Doutorado). Faculdade de Ciências Médicas e Biológicas de Botucatu. [ Links ]
 CEREDA,M.P. Avaliação da qualidade de duas amostras de fécula fermentada de mandioca (polvilho azedo). Bol. SBCTA, v. 17, n. 3, p. 305-20, 1983a. [ Links ]
 CEREDA,M.P. Avaliação da qualidade da fécula fermentada comercial de amndioca. II. Características físico-químicas e absorção de água. Revista Brasileira de Mandioca, Cruz das Almas, v. 3, n. 2, p. 15-20, 1985. [ Links ]
 CEREDA,M.P. Estudos físico-químicos e microbianos da esterilização e da fermentação da fécula da mandioca, 1981. 155p. (Tese de Livre-Docência). Faculdade de Ciências Agronômicas, Campus de Botucatu, Universidade Estadual Paulista [ Links ]
 CEREDA,M.P. Padronização para ensaios de qualidade de fécula de mandioca fermentada (polvilho azedo). I. Formulação e preparo de biscoitos. Bol. SBCTA, v. 17, n. 3, p. 287-95, 1983b. [ Links ]
 CEREDA,M.P.; CATÂNEO,A. Avaliação de parâmetros de qualidade da fécula fermentada de mandioca. Revista Brasileira da Mandioca, v. 5, n. 1, p. 55-62, 1986. [ Links ]
 CEREDA,M.P.; LIMA,U.A. Aspectos sobre a fermentação da fécula de mandioca. III. Determinação dos ácidos orgânicos. Turrialba, v. 35, n. 1, p. 19-24, 1985. [ Links ]
 CHROMA: Système dacquisition et dintegration de chromatogrammes multivoie. Couternon, Biosystèmes, 199?. 54p. (Manuel dutilisation v.3.0). [ Links ]
 CHUZEL, G. Amidon aigre Colombie. In: Amelioration de la qualité des aliments fermentés à base de manioc. Montpellier, CIRAD/CEEMAT, ORSTOM, 1992. [ Links ]
 DEMIATE,I.M.; SENGER,S.A.; VOGLER,Z.; CEREDA,M.P.; WOSIACKI,G. Características de qualidade de amostras de polvilho azedo produzido ou comercializado no Estado do Paraná. Arq. Biol. Tecnol., v. 40, n. 2, p. 331-49, 1997a. [ Links ]
 DEMIATE,I.M.; SOUZA,T.O.de; PUGSLEY,S.; CEREDA,M.P.; WOSIACKI,G. Características de qualidade de amostras de polvilho azedo. Parte 2. Santa Catarina. Agropec. Catarinense, v. 10, n. 4, p. 51-6, 1997b. [ Links ]
 ESCOBAR,C.A.; MOLINARI,J. Obtención de parâmetros para la evaluación de la calidad de un almidón agrio de yuca. Cali, 1990. (Tesis de grado). Universidad del Valle. [ Links ]
 FIGUEROA,C. Etude bibliographique sur la fermentation de lamidon aigre et les characteristiques du produit obtenu. Paris, Institut National Agronomique Paris-Grignon, 1993. (Projet CEE STD3 "Valorisation des produits, sous produits et déchets de la petite et moyenne industrie du manioc en Amérique Latine" - operation 3 Bioconversions des farines et amidons de manioc. CIRAD-SAR). [ Links ]
 FIGUEROA,C; DAVILA,A.M.; POURQUIE,J. Lactic acid bacteria of the sour cassava starch fermentation. Letters in Applied Microbiology, v. 21, p. 126-30, 1995. [ Links ]
 FIGUEROA,C; DAVILA,A.M.; POURQUIE,J. Original properties of ropy strains of Lactobacillus plantarum isolated from the sour cassava starch fermentation. Journal of Applied Microbiology, v. 82, p. 68-72, 1997. [ Links ]
 MARTINEZ,A. Aislamiento y caracterización de la flora presente en la fermentación del almidón de yuca y estudio del processo, Doc. IIT, Bogotá, 1988. [ Links ]
 MESTRES,C.; ROUAU,X. Influence of natural fermentation and drying conditions on the physicochemical characteristics of cassava starch. J. Sci. Food Agric., v. 74, p. 147-55, 1997. [ Links ]
 WETSBY,A.; CEREDA,M.P. Production of fermented cassava starch (polvilho azedo) in Brazil. Tropical Science., v. 34, 1994. [ Links ]
1 Recebido para publicação em 15/09/98. Aceito para publicação em 31/03/99.
3 Universidade Estadual Paulista /CERAT, Campus de Botucatu. C.P. 237, Botucatu SP Brasil, CEP 18603-970.
* to whom correspondence may be addressed.