41IntroductionMango  (Manguifera  indicaL.)  is  a  well-knownand  highly  consumed  fruit,  as  fresh  fruit,  in  juices, marmaladesand     dry     snacks     (Villamizar and  Giraldo,  2010)  and  different  varieties   exist   like   Tommy   Atkins,   Haden,   Manzana,  Kent,  among  others.    In  100  g  of  this  fruits  there  are,  in  average,  0.7  g  of  pro-tein,  16.8  g  of  total  carbohydrates,  10  mg  of  calcium, 13 mg of phosphorus, 0.4 mg of iron, 7  mg  of  sodium,  189  mg  of  potassium,  4800  Ul  of  Vitamin  A,  0.05  mg  of  Thiamine,  0.05  mg of Riboflavin, 1.1 mg of Niacin, and 35 mg of ascorbic.Frying is a complex process widely used in the  food  industry.    During  this  process,  food  is immersed  in  oil  at  a  higher  temperaturethan  the  water  boiling  point  in  order  to  pro-duce a vapor contraflow (bubbles) with the oil in  the  food  surface  (Bouchon et al.,  2003).This  process  generates  products  with  the  or-ganoleptic  qualities  (color,  texture  and  taste)  demanded  and  appreciated  by  consumers;  however these products have high fat content that  is  not  suitable  for  human  consumption  (USDA,  2008).    Besides  the  effects  of  oil  and  saturated fat consumption on human health, due  to  the  high  temperatures  and  oxygenexposition  during  the  frying  process  other  effects that are not desirable are present such as  degradation  of  important  nutritious  com-ponents  and  generation  of  toxic  compounds  (acrylamide)  in  the  snack  or  in  the  frying  oil  (Fillion and Henry, 1998).  To reduce fat con-tent  without  losing sensorial  qualities,  nume-rous complementary processes or alternatives to  the  frying  process  have  been  proposed  (Mellema,  2003; Ziaiifar et  al., 2008).    One  of  them   consists   of   pressure   reduction   by   working  on  vacuum  to  reduce  water  boiling  point  in  the  food  and  eliminate  it  at  low  tem-peratures (Mir-Belet  al.,  2009).    It  has  been  proven that besides the reduction on final fat content  (Garayo  and  Moreira,  2002),  vacuumfrying  products  have  other  advantages,  like  very  low  acrylamide  content  (Granda et  al., 2004)  and  better  organoleptic  and  nutritious  quality (Shyu and Hwang, 2001; Da Silva andMoreira, 2008; Troncosoet al. 2009).Frying   process   implicates   simultaneous   mass and  heat  transfer  that  make  important  microstructural  changes,  both  in  the  surface  and in the mass of the product.  Heat transfer generates protein denaturation, starch gelati-nization,  water  vaporization,  crust  formation  and,  color  development,  which  are  typical phenomena  of  the  combination  effect  of  mul-tiple  chemical  reactions.    Mass  transfer  is  characterized   by   the   interchange   of   com-pounds like water and other soluble materials occluded  in  the  starch,  which  allows  oil  pe-netration in food (Mir-Bel et al., 2009). Oil absorption is not clearly explained due to  the  multiple  variables  involved,  such  as  initial product structure, diverse interchanges between  the  product  and  heating  media,  va-riation in oil products and properties, chemi-cal  reactions,  interactions  between  food  com-pounds and oxidized lipids, and hydrolysis of fats in the frying oil due to humidity (Velasco et  al.,  2008).      The  aim  of  this  study  was  to  observe  the  effect  of  vacuum  frying  on  the  quality  of  mango  snacks  (texture,  color,  hu-midity percentage, water activity and fat con-tent)  in  order  to  define  the  best  conditions  (temperature, time and pressure) of the frying process. Materials and methodsMango  (Manguifera  indica  L.)  of  the  cultivar  Tommy  Atkins  between  6  and  8  days  after  harvesting  were  used.    Mangos  have  14-15 °brix   measured   on   a   table   refractometer   according to AOAC 932.12, they were coming from  Tolima  and  were  bought  in  the  Arme-nia ?s  market  in  Quindio, 1480 MASL and at-mospheric pressure 640 mmHg. Paste characterization and preparationSelected  mangos  were  characterized  during  post-harvesting  by  physicochemical  analysis  of  soluble  solids  (°brix),  humidity  percentage  (bh), water  activity  (aw),  acidity  %,  starch content, color and texture.  With  the  preliminary  characterization  re-sults,  mangos  were  washed,  peeled  and  ma-nually  pulped  to  get  230  g  of  pulp  that  were  homogenized  on  a  blender  for  1  min  to  get  a  puree type of paste.  Next, 46 g of starch and wheat  flour  (4:1)  were  added  and  homoge-nized  with  the  mango  puree  to  fill  round  molds  (2  mm  thickness  ad  4  cm  diameter)  

EFFECT OF VACUUM FRYING PROCESS ON THE QUALITY OF A SNACK OF MANGO (MANGUIFERA INDICA L.)    42that were refrigerated at 7 °C and HR 21% for 48 h. Frying processOnce  the  paste  was  extracted  from  the  mold,  it  was  fried  in  a  machine  adapted  for  that  purpose, a metal frame with a 250 ml beaker for  oil,  a  pulley  system  for  immersion  and  a  basket  to  carry  samples  was  placed  on  a  va-cuum  oven  and  the  sample  was  immersedusing magnets.  For each test, palm oil on the oven was at process temperatures (100, 110 and 120 °C), then  the  system  was  adjusted  to  the  proper  vacuum pressure (0.4, 0.5 and 0.6 bar) before the  paste  immersion  and  frying  was  done  during  specific  times  (30,  45,  60,  75  and  90  s),   kinetic   that   was   used   to   determine   changes in quality characteristics.  Snack was taken  out  and  superficial  oil  was  extracted  with a paper towel taking into account a ratio paste:oil of 1:15 weight: volume. Treatments  were  replicated   three   times   and  samples  were  analyzed  by  triplicate  for  humidity  percentage  (%  bh), water  activity, color, texture (crunchiness) and fat (%).Physico-chemical analysisHumidity contents of mango, paste and snack were  determined  in  a  vacumm  dry  oven  (J.P  Selecta  S.A.)  according  to  the  method  AOAC  20,013, 1980 (A.O.A.C.,1980), in the following way: %????????(??) =??????????  ??−?????????????  ????????????  ??×100Water   activity   (aw)   determinations   are   measured  directly  at  room  temperature  (bet-ween  20  and  30  °C)  using  a  Decagon dew point  hygrometer  (AquaLab  3  TE)  with  0.001  sensibility.    Color  determination  was  done  with  a  colorimeter  (Minolta  CR  –  10)  to  eva-luate  changes  in  fruit,  paste  and  snack  color  by  the  Cielab  system,  with  D65  illuminant  and 10° for the observer.  Taking  the  values  for  the  coordinates  L* (dark-light), a*  (green-red), b*   (blue-yellow) and calculating the color difference in compa-rison   to   the   paste,   represented   by   ?E, according to the following equation??=?(?∗−?0∗)2+(?∗−?0∗)2+(?∗−?0∗)2where:  L* =  L value for the treatment sample.  L0* =  L value for the standard sample. a* =  a value for the treatment sample.  a0* =  a value standard sample.  b* =  b value treatment sample.  b0* =  b value standard sample.  To  determine  the  fracturability  point  for  the   mango,   paste   and   snack,   a   texture   analyzer    (TA.    XT.    Plus)    was    used    in    compression  mode  with  5  g  strenght,  3  mm  distance and 10 mm/seg speed (fracturability of tortilla chips).Fat   content   in   paste   and   snack   were   determined using a fat detector  (DET  – GRAS P Selecta) by the DG – 01 (without hydrolisis) method  reported  on  the  equipment  mannual  according to the equation: %???=????????  ????? ???? −?????????????  ????????????  ??×100Statistical analysisData  for  the  analysis  was  organized  using  Statgraphics  Plus  5.1  software  (Statgraphics,  2001)  to  evaluate  the  effect  of  the  different  treatments   on   the   snack   characteristics.      These effects  were  guided  using  ANOVA  ana-lysis for one significance (P < 0.05).  The expe-rimental   design   used   was   an   unbalanced   factorial design 5x32, with five time levels (30, 45, 60, 75 and 90 s), three temperature levels (100,  110  and  120  °C)  and  three  vaccumm  pressure levels (0.4, 0.5 and 0.6 bar) with the following model: ????=?+??+??+??+?????+?????+?????+????????+????where: ?:  general mean. ??:  ith time level in seconds.??:  jth temperature level in °C. ??:  sth vacuum pressure level in bar. ?????:  first  order  interaction  between  time  and temperature. ?????:  first  order  interaction  between  tem-perature and pressure. 

ACTA AGRONÖMICA. 61 (1) 2012, p 40-49 43?????:  first  order  interaction  between  tem-perature and pressure. ????????:   second    interaction    between    the    three factors.????:  experimental  error  associated  with  the three factors.Additionally,  for  comparison,  Tukey ?s  test  at 95%  was done. Results and discussionFruit characterizationAnalysis  showed  that  mango  fruits  used  in  this study had water content in fresh pulp of 87.45  ±  2.15%,  water  activity  of  0.986  ±  0.003, °brix 14.26 ± 0.38, cutting force of 3.93 ±   0.060   kgf,   a   fat   content   of   0.00169   ±   0.00018%, results that are similar to the ones found by Stafford (1983).  Snack characterizationANOVA  results  for  all  the  snack  variables  studied are presented on Table 1.  It is obser-ved that the relation of the factors with those variables  is  significant,  meaning  that  from  this  analysis  there  is  not  an  ideal  treatment  that  can  be  define  as  the  best.    For  this  rea-son,  each  one  of  the  snack  physico-chemical properties was analyzed separately in relation to  the  change  in  vacuum  pressure  done  in  each treatment.  Humidity contentResults  showed  that  pressure,  temperatureand  time  of  each  treatment  affected  snack  humidity  content  (P  <  0.05)  in  the  different  vacuum pressures evaluated (Figure 1 a, b, c).  For  each  one  of  the  treatments,  humidity  content  was  reduced  when  the  vacuum  pre-ssure  was  increased  (a,  b,  c),  the  same  ha-ppened  when  temperature  and  frying  time  were increased.  Initially, water loss celerity was high due to its content  on  the  snack  surface,  but  it  was  accelerated  with  vacuum  pressure  increase  due  to  the  rise  on  strength.    However,  with  longer  frying  times  humidity  loss  was  slower  due  to,  in  part,  the  microstructural  changes  that  happen  at  the  beginning  of  pressuriza-tion  that  affect  the  water  release  from  the  surface.    Additionally,  because  of  the  initial  dryness  (refrigeration)  a  low  concentration  of  available water is presented and crust forma-tion  makes  a  higher  resistance  to  the  water  scape (Mariscal and Bouchon, 2008). Water activity (aw)   In  table  1  is  shown  the  effect  of  the  vacuum pressure  and  temperature  on  the  aw  (P  <  0.05),  whereas  the  time  effect  is  low  (P  >Table 1. Analysis of variance for the treatment effect on some properties of mango snacks. Source of variation (treatments)Fat content(%)Humidity content(%bh)?ECut point(kgf) awModel0.00020.00000010.00010.00590.0000001Pressure0.00000030.00000020.00000020.00050.0000001Temperature0.00000020.00000010.00000010.00490.00000005Time0.00020.01190.10080.01090.1329Pressurex  temp.0.00000010.00000010.00000010.00000030.00000002Pressure x time0.02310.00000010.00030.00320.0000001Temp. x time0.00210.16490.07160.00220.0000001Pressure  x temp. x time0.14730.13090.00740.0240.00000002Best treatment0.6-110-600.6-120-750,5-110-450.5-110-900.5-110-90Worst treatment0.5-100-3004-100-300.6-120-900.4-110-900.4-100-30P < 0.05 significant, reliability of 95%.P < 0.01 highly significant, reliability of 99%. 

EFFECT OF VACUUM FRYING PROCESS ON THE QUALITY OF A SNACK OF MANGO (MANGUIFERA INDICA L.)    440.05).  In Figure 2 is shown the water activity behavior of the fried snack in each one of the treatments.  The observed levels are relatively low  which  favors  product  preservation  impe-ding  oxidation  and  proliferation  of  harmful  microorganisms,  also  these  levels decrease with temperature increases and tend to stabi-lize  with  the  time.    Similarly,  even  if  water  activity reaches low levels changes in vacuum pressure do not have an effect in its variation, which  could  be  due  to  the  fact  that  in  the  vacuum frying process the boiling water tem-perature   decreases   (Mir-Bel et   al.,   2009)   eliminating   high   volumes   of   free   water.Figure 2.  Water activity in mango Tommy Atkins snacks under three vacuum pressures and times x variable temperatures. 

ACTA AGRONÖMICA. 61 (1) 2012, p 40-49 45ColorSnack  color  was  affected  by  temperature  and  vacuum pressure (P < 0.05); but this was not due to the kinetics originated by temperature and time  (P  >  0.05)  (Table1).    Color  changes  (?E) in comparison to the ones of the paste for each   treatment   are   shown   in   Figure   3.      Changes  are  represented  by  ?E  increments that  have  a  slight  variation  due  to  relatively  low   working   temperatures   in   the   vacuum   pressure  treatments  that  reduce  the  Amadori  products  which  favors  melanoidins.    In  fried  food  under  atmospheric  pressure,  darkening  is  notorious  since  it  requires  temperatures  >  150  °C  (Pokorny,  1999),  showing  chemicalFigure 3.Color changes in mango Tommy Atkins snacks under three vacuum pressures and times x variable temperatures. 

EFFECT OF VACUUM FRYING PROCESS ON THE QUALITY OF A SNACK OF MANGO (MANGUIFERA INDICA L.)    46changes that generate acrylamide (Fillion and Henry,  1998).    Vacuum  frying  process  pre-vents formation of acrylamide and compounds that  causes  dark  colors  following  Maillard ?s  reaction,   because   of   the   reduce   oxidation   during  the  process  (Da  Silva  and Moreira, 2008). Texture  This characteristic was affected by the applied treatments (P < 0.05) (Figure 4).  As result of changes in vacuum pressure, texture showed a  high  variability  in  the  different  treatments,  this difficult the determination of its influence in  the  reduction  of  the  strength  needed  to  break the snack showing variation when timeFigure 4.Texture changes in mango Tommy Atkins snacks under three vacuum pressures and times x variable temperatures. 

ACTA AGRONÖMICA. 61 (1) 2012, p 40-49 47and  temperature  are  increased.    This  is  the  result of fast gelatinization of starch granules in contact with hot oil and the transformation of its superficial structure in a crunchy crust (Pokorny, 1999).Fat contentVacuum   pressure,   temperature   and   time   during the frying process affect fat content (P <  0.05).    In  Figure  5  is  shown  that  while  va-cuum  pressure  and  temperature  increment,  fat  content  reduces  and  stabilizes  during  the  process.    Fat  content  behavior  in  the  snack  has two phases.  In the first one, frying, high temperatures  causes  a  partial  evaporation  of  water  removed  from  the  snack  interior  being  partially  replaced  by  oil;  this  evaporation  isFigure 5.  Fat contentsin mango Tommy Atkins snacks under three vacuum pressures and times x variable temperatures. 

EFFECT OF VACUUM FRYING PROCESS ON THE QUALITY OF A SNACK OF MANGO (MANGUIFERA INDICA L.)    48intense  since  the  vacuum  pressure  increases  the speed of  mass and heat transfer resulting on  a  water  boiling  temperature  decrease  (76  to  86  °C).    The  second  phase,  pressurization,  happens when the snack is removed from the oil  increasing  quickly  pressure  and  tempera-ture in the pores; this generates oil adherence to the snack surface and oil penetrance in the snack (sponge effect) until the pressure in the pores equals the atmospheric pressure (Tron-coso et al., 2009).Best treatmentBest treatment combinations for each vacuum pressure related      to      the      evaluated characteristics are included in Figure 6.  The best  results  for  texture,  water  activity  and  color  happen  at  vacuum  pressure  of  0.5  bar  (0.5   bar),   whereas   for   fat   and   humidity   content  is  at  0.6  bar  (-0.6  bar);  this  means  that a vacuum pressure of 0.4 bar is too low for  getting  a  snack  with  good  characteristics.    The  fat  content  difference  at  0.5  bar  vs.  0.6  Figure 6.Best treatment combinations for mango Tommy Atkins snacksunder three vacuum pressures and times x variable temperatures. 

ACTA AGRONÖMICA. 61 (1) 2012, p 40-49 49bar  was  not  significant,  the  same  happened  with  the  humidity  content  at  0.5  bar  vs.  0.6  bar,  meaning  that  the  best  vacuum  pressure  treatment  corresponds  to  0.5  bar,  110  °C  temperature and inmersion time of 90 sec.Conclusions•Vacuum   pressure   positively   affects   the   quality  characteristics  of  fried  snacks  of  mango   resulting   crunching,   with   good   color and similar to the fruit, with low fat content  and  adequate   humidity   for   its   conservation.•The best characteristics for the fried snack were  color  (?E  =  19  ±  1.5635),  texture  (fracturability strength   =   0.256367   ±   0.005736),  fat  content  (%fat  =  9.4995  ±  0.8744),  humidity  content  (%humidity  bh  =  1.25  ±  0.3037)  and water  activity  (aw  = 0.342  ±  0.0014),  obtained  with  a  vacuum  pressure of 0.5 bar, temperature of 110 °C and immersion time of 90 s.•Snack  characteristics  are  better  than  the  traditional   snack   characteristics   which   have humidity content lower than 4% and fat  content  lower  than  30%  (Rodríguez et al., 1999)ReferencesA.O.  A.C.    1980.  Método  oficial  20.013.  Humedad  en  plantas.  Métodos  oficiales  de  análisis  AOAC.  A Internacional.Bouchon,  P.;  Aguilera,  J.  M.;  and Pyle,  D.  L.  2003.  Structure    oil-absorptionrelationships    during    deep-fat  frying.  Journal.  Food  Science.  68:2711 - 2716.Da  Silva,  P.  F.;  and Moreira,  R.  G.  2008.    Vacuum frying  of  high-quality  fruit  and  vegetable-based snacks. LWT -  Food Science Technology. 41:1758 - 1767. Fillion, L. and Henry, C. J. 1998. Nutrient losses and gains    during    frying:    A    review.    InternationalJournal of  Food Science and  Nutrition, 49(2):157 -  168.Garayo,  J.  and  Moreira,  R.  2002.  Vacuum  frying  of  potato    chips.    Journal    of    Food    Engineering. 55(2):181 - 191.Granda,  C.;  Moreira,  R. G.;  and  Tichy,  S. E.,  2004.  Reduction of acrylamide formation in potato  chips  by   low-temperature   vacuum   frying.   Journal   ofFood Science, 69 (8):E405 -  E411.Mariscal,  M.  and Bouchon,  P.  2008.  Comparison between  atmospheric  and  vacuum  frying  of  apple  slices. Food Chemistry. 107:1561 -  1569.Mellema,  M.  2003.  Mechanism  and  reduction  of  fatuptake  in  deep-fat  fried  foods.  Food  Sci.  Technol. 14:364 -  373.Mir-Bel,    J.;    Oria, R.;    and Salvador,    M.    L.2009.Influence of the vacuum break conditions on oil   uptake   during   potato   post-frying cooling. Journal.  Food  Engineering.  Laboratory  of  Vegetal Food,  University  of  Zaragoza,  Miguel  Servet  177,  50013    Zaragoza,    Spain.    Journal    homepage:    www.elsevier.com/locate/jfoodengPokorny,  J.  1999.  Changes  of  nutrients  at  frying  temperatures. En: Boskou, D. y  Elmadfa, I.  (eds.). Frying    of    FoodTechnomic    Publishing    Co.    Lancaster. p. 69 -  103.Rodriguez,   C.;   Vitrac,   O;   and Dufour,   D.   1999.   Caracterización  del  proceso  de  fritura  de  chips  de  yuca   (Manihot    esculenta Crantz)   Estudio   de   algunas  variables  de  proceso.  GeoTropica  Revista  del área de recursos naturales 4: 68 – 79.Shyu, S. and Hwang, L. S. 2001. Effects of processing conditions  on  the  quality  of  vacuum  fried  apple  chips. Food Research International. 34: 133 – 142.Stafford, A.  E.  1983.  Mango.  handbook  of  tropical  foods. En:  H.T.  Chan,  Jr..  (ed.).  Marcel  Dekker,  Inc., Nueva York. p. 399 -  431.Statgraphics.           2001.           Disponible           en:           http://www.statgraphics.comTroncoso,  E.;  Pedreschi,  F.;  and  Zuñiga,  R. N.  2009.  Comparative    study    of    physical    and    sensory    properties   of   pre-treated   potato   slices   during   vacuum   and   atmospheric   frying.   Lebensmittel-Wissenschaft  und-Technologie.  Food  Science  andTechnology. 42:187 – 195.USDA  (U.S.  Department  of  Agriculture,  Agricultural  Research  Service).  2008.  USDA  national  nutrient  database   for   standard   reference.   Release   21.   Nutrient      data laboratory      home.      Page,      http://www.ars.usda.gov/ba/bhnrc/ndl.Velasco, J.; Marmesat, S.; and Dobarganes, C.; 2008. Chemistry  of  frying.  En:  Sahin,  S.  y  Sumnu,  S.G.  (eds.).  Advances in Deep-fat  Frying  of  Foods.  CRC Press, Boca Raton, Fla. p. 33 - 56.Villamizar,  R.  and  Giraldo,  G.  2010.  Obtención  y  caracterización  de  un  pasabocas  a  partir  de  una  pasta   a   base   de   mango   mediante   fritura   por   inmersión. Rev. Tumbaga UT 5:149 -  164.Ziaiifar,  A.  M.;  Achir,  N.;  Courtois,  F.;  Trezzani,  I.; and Trystram,  G.,  2008.  Review  of mechanisms, conditions,  and  factors  involved  in  the  oil  uptake  phenomenon during  the  deep-fat  frying  process.  International Journal    of    Food    Science    andTechnology. 43(8):1410 -  1423.