Ind. Eng. Chem. ProceSS Des. DeV. 1983, 22, 31-36
31
o = weighting factor in eq 12
T = sampling interval, s
To = feed temperature, O C TR = reactor bed temperature, O C T, = wall temperature, O C To, = average temperature of oil in cooling coils, O C ub = bubble rise velocity, cm/s ud = superficial velocity at minimum fluidization, cm/s uH2 = feed rate of hydrogen at STP, cm3/min UC, = feed rate of butane at STP, cm3/min u, = superficial gas velocity, cm/s V = total volume of the bed, cm3 V , = average bubble volume, cm3 V , = volume of emulsion phase, cm3 uo = total volumetric flow rate of inlet gas, cm3/s W = weight of catalyst, g z = 2-transform variable 2 = time history vector for the three temperatures, TR, T,,
Subscripts b = bubble phase e = emulsion phase 0 = feed Superscripts i = the ith component set = set point value measd = measured value Registry No. Butane, 106-97-8.
Literature Cited Dahlin, E. 8. Inst. Control Syst. 1966, 47, 77. Jutan, A.; MacGregor, J. F.; Wright, J. D. A I C M . J . 1977, 23, 742-750. Kato, K.; Wen, C. Y. Chem. Eng. Sci. 1969, 24, 1351. Kobayaski, H.; Arai, F.; Sunagawat, T. Chem. Eng. Tokyo 1965, 2 9 , 858. Orcutt, J. C.; Davldeon, J. E.; Word, R. L. Chem. Eng. - Rag. Symn. Ser. NO. 58, 1962, 1-15. Orlkas, A.; Hoffman, T. W.; Shaw, I. D.; Reilly. P. M. Can. J . Chem. €ng. 1972. 50. 628-636. Shew, I. D. W.D. The&, Chemlcal EnginesringDepartment. McMaster Unlversitv. 1974. Shaw, i.’D.; Hoffman, T. W.; Orllkas, A.; Reilly, P. M. Can. J . Chem. €ng. 1972. 50. 637-643. Shaw. 1: D.;’Hoffman, T. W.; Reilly, P. M. AICM. Symp. Ser. 70, 1973, 41-52. Shinskey, F. G. “Rocess Control Systems”; McGraw-HIII: New York, 1979. Smith, C. L. “Olgltal Computer Process Control”; Intext Ed. Publ.: Scranton, PA, 1972.
Toil
2 = the estimate of 2 from the model
Greek Letters bed void fraction, dimensionless tS3, = propane selectivity error at time t zTRI-= temperature error at time t pg - average molal density of gas, g-mol/cm3 pp = average particle density, g/cm3 T = process time constant 0 = process dead time X = desired closed loop time constant (a tuning parameter in Dahlin’s algorithm) 6 = desired closed loop dead time 9 = proportionality constant in eq 8 z =
Received for review May 26, 1981 Revised manuscript received May 24,1982 Accepted June 17, 1982
Production of Base-Catalyzed Phenolic Resins in Bubble Columns Patrkia M. Frontlnl and Roberto J. J. Wlliams’ Lbpamwnt of Chemical Engineerlng, University of Mer del Plats, (7600) Mer del Plats, Argentina
The performance of a bubble column as a gas-liquid contacting device for the production of resols is shown. A gaseous mixture of formaldehyde and nitrogen was bubbled in an alkaline phenol solution held at 80 OC. The formaldehyde absorption rate was controlled by mass transfer in the gas phase. The volumetric mass transfer coefficient was satisfactorily predicted wlth equations describing transient diffusion in spherical geometry. Values of holdup and Sauter mean bubble diameter are reported as a function of the superficial gas velocity. Several characteristics of the resulting resols were analyzed using chemical, IR, DSC, and viscosimetric techniques. The process showed several advantages when compared with the usual batch process.
Introduction The base-catalyzed phenolic resins,i.e., resols, are widely used in several industrial applications such as particle boards, plywoods, fiberboards, copper-clad laminates, abrasive papers, etc. Resols are usually produced in batch reactors. This is due to economical reasons such as the multitude of resin specifications demanded by the market as well as technical arguments such as the formation of cured resin coats on hot surfaces, which render the continuous process impracticable. There is, however, some limited information regarding the existence of continuous industrial processes (Strelcow et al., 1972;Brunnmiiller, 1975). On the other hand, the batch production has some complications, the main one being the limitation of the vessel volume by the high exothermic heat. This makes 0196-4305/83/1122-0031$01.50/0
it necessary to process diluted solutions in order to exhaust the reaction heat by evaporation of water. In this way the reaction is kept under control but the reactor output is considerably decreased. An effective way of avoiding many of the processing difficulties, as well as decreasing the operation cost, is to carry out the reaction in a gas-liquid device. In this way, the gaseous stream leaving a methanol oxidation reactor can be contacted with an alkaline solution of phenol to produce a resol. Some of the advantages of this operation are as follows: (a) the reaction heat evolved per unit time is limited by the amount of formaldehyde reacting at any time; (b) there is no need of producing formalin (the commercial aqueous formaldehyde solution) as a previous step, resulting in a decrease of the operation cost and avoiding the consumption of demineralized water in the 0 1982 American Chemical Society
32 Ind. Eng. Chem. Process Des. Dev., Vol. 22, No. 1, 1983 Table I. Operating Conditions of the Bubble Column ~
run
T,"C
2, cm
A B C D E F
80 80 80 80 80 80 60
41.7 43.0 41.8 39.2 40.7 44.4 42.7
G
y, 0.29 0.29 0.24 0.10 0.17 0.42 0.24
formaldehyde absorption; (c) the reactor output can be raised and the amount of effluent reduced. Strupinskaya et al. (1968) reported the operation of a pilot plant absorbing the formaldehyde from the gaseous stream directly into phenol, using a countercurrent packed-column with an output of 5 tons/day. The resulting product was subsequently used for the batch production of phenolic resins. The aim of this study is to analyze the performance of a bubble column as a gas-liquid contacting device for the production of resols. Some of the advantages offered by bubble columns are the high heat transfer rates, low cost, minimum maintenance, high interfacial areas, and overall mass transfer coefficients (Mashelkar, 1970). Experimental Section Materials. The alkaline phenol solution had the following composition by weight: PhOH = 52.6, NaOH = 3.5, and H 2 0 = 43.9. Both PhOH and NaOH were technical grade reagents. The density and surface tension, expressed as ratios with corresponding water values, were (pip,) = 1.075 and (u/uw) = 0.62. Paraformaldehyde was a technical grade reagent with 92% formaldehyde, as reported by the supplier. This proportion was verified by titration with the chromotropic acid method (Urbanski et al., 1977). Nitrogen gas had a 99.99% purity. Equipment. A flowsheet of the experimental device is shown in Figure 1. Formaldehyde arising from the thermal decomposition of paraformaldehyde in an electrically heated copper vessel was added to the nitrogen stream (Ingraham and Fraser, 1969; Crusos, 1978). Temperature was set in the range 130-150 "C, where a significant decomposition rate was shown by differential scanning calorimetry (Du Pont 990 Thermal Analyzer). The gas line from the decomposition vessel to the bubble column was a copper tube electrically heated to avoid formaldehyde repolymerization. The glass bubble column (di = 5.1 cm, #o = 6 cm) was jacketed with a transparent acrylic tube (& = 8 cm, cb0 = 9 cm), where a water stream at a constant temperature was recirculated. Bubbling was produced by a single-orifice gas sparger (& = 0.25 cm). Attempts to use a glass porous septum as gas sparger failed because of the continuous blocking of pores by formaldehyde repolymerization. Liquid samples were obtained from an outlet valve placed a t the column base. Operation. The bubble column was operated as a semicontinuous reactor. The alkaline phenol solution was fed to the column and nitrogen bubbled at a constant flow rate in the range 1-4 L/min, at standard conditions. The solution was heated to 80 "C, unless otherwise stated. The temperature of the paraformaldehyde decomposition reactor was set in agreement with the required decomposition rate, using a calibration curve. The paraformaldehyde amount actually decomposed was checked by weighing the copper vessel before and after each run. The formaldehyde concentration in the exit gases was determined by absorption in water followed by titration with the chromotropic acid method (Urbanski et al., 1977). This determination was performed several times during each run. The polymerization extent was followed by measuring the viscosity increase (Rotovisco RV3 Haake). When it at-
y*
us, cmls
5.1 X 8.4 X 2.5 X 2.5 X 5.5 x 10-6 7.0 X 3.1 X
3.15 3.15 3.09 2.90 4.70 1.46 2.87
~
~~
~~~
Reo 1820 1810 1760 1540 2590 880 1790
Figure 1. Experimental apparatus: (1) nitrogen tube; (2) capillary flowmeter; (3) pressure gage; (4)paraformaldehyde decomposition veasel; (5)control device actionating an electric heater; (6) heated line; (7) jacketed bubble column; (8) inlet valve; (9) outlet sampling valve; (10)recirculating thermostat; (11) exit gases to analysis.
tained a level of several hundred centipoises at 30 "C, the heater of the decomposition vessel was turned off. When temperature had gone down to 80 OC, the resol was discharged. The gas holdup, bubble diameter distribution, and temperature of the polymerizing fluid were measured. The gas holdup was calculated from the heights of the initial and expanded beds. The bubble diameter distribution was obtained by taking photographs using exposure times less than s. The distribution was measured in amplified copies by taking known internal dimensions as reference. The temperature of the polymerizing fluid was monitored with a copper-constantan thermocouple. Characterization of the Resol. The formaldehyde/ phenol molar ratio, F/P, was calculated from the amount of formaldehyde which was actually absorbed. The fraction of free formaldehyde (not l i k e d with phenolic rings) was determined using the method reported by Haslam and Soppet (1953). Total solids were obtained from the weight loss of thin samples, heated to 130-150 "C in a humidity balance (Ultra X). Infrared spectra were obtained by wetting KBr pellets with the resols and monitoring on a Perkin-Elmer 599 IR spectrophotometer. The energetics of the curing process was analyzed with a differential scanning calorimeter fitted with a DSC cell (Du Pont 990 Thermal Analyzer). Runs were carried out at a 10 OC/min scanning rate and 650 psi (nitrogen). A rapid comparison among curing rates of different resols was obtained by following the viscosity increase at constant temperature (Rotovisco RV3 Haake, T = 74 "C). The stability of different products was estimated by measuring the viscosity increase during the storage at room temperature.
Results and Discussion Bubble Column Performance. Table I shows typical operating conditions for several runs. The liquid temperature remained constant during a run, giving a solution to the problem caused by the exothermic reaction heat in the batch process. Yl and Y2are, respectively, the formaldehyde/nitrogen molar ratios at the reactor inlet and
Ind. Eng. Chem. Process Des. Dev., Vol. 22, No. 1, 1983 33
k./_,_L ,
-
1
Figure 2. Experimental holdup valuea compared with Hughmark’s correlation for a 2-in. diameter bubble column.
1
___
4
3
“ W 5
Figure 4. Volumetric mass transfer coefficient, kp,and interfacial area per unit volume, a, as a function of the superficial gas velocity, 80 O C ; ( 0 )60 O C . u,: (0)
p is the formaldehyde partial pressure at a given axial position, and p* is the pressure corresponding to the equilibrium with the liquid phase. However, as the formaldehyde is extremely soluble in the alkaline phenol solution p* -. 0;
K, -. k,
where k, is the gas-side mass transfer coefficient. Also
P = YPt/(1 + Y)
(2)
assuming a constant total pressure, pt, and neglecting the presence of water in the gaseous stream. Substituting in eq 1, we get I
e
-QN 3
4
5
dY = kga[Y/(l
+ Y)lp, dz
(3)
vs (tmjsec)
Figure 3. Sauter mean bubble diameter of the whole population as a function of the superficial gas velocity.
outlet. Y2corresponds to an average of several determinations carried out during each run. All the values were close to the average, with random deviations. The significantly low Y2 values show the high efficiency of the absorption process in exhausting almost all the formaldehyde from the inlet gases. The superficial gas velocities us, calculated as an average value between the inlet and outlet, lay in the quiescent regime range (Mashelkar, 1970). Reynolds numbers a t the sparger’s orifice lay in the intermediate regime (only run E was in the jet regime, Reo > 2100). Figure 2 shows experimental holdup values obtained by bubbling nitrogen in the reaction mixture, compared with a correlation for a 2 in. diameter bubble column (Hughmark, 1967). The Sauter mean bubble diameter, dg, is defined taking into account the surface/volume ratio of the entire bubble population. It was estimated from photographs and plotted as a function of the superficial gas velocity, as shown in Figure 3. The decrease in dB when increasing us should be mainly attributed to a reduction in the fraction of large bubbles. huming plug flow in the gas phase, the following molar balance may be written -QN dY = K@(P - p * ) dz (1) where Q N is the nitrogen molar flow rate per unit transversal area, K, is the overall mass transfer coefficient in gas phase units, a is the interfacial area per unit volume,
Integrating, we get
kp = (QN/PJ)I(YI - Y2) + In (Yi/Y2)1
(4)
Figure 4 shows (k a ) as a linear function of the superficial gas velocity. d e fact that runs at 60 and 80 “Cgave the same volumetric mass transfer coefficient, kp, supports the assumption of gas-phase control. On the other hand
u
6t/dB
(5)
where t may be calculated from Hughmarks correlation (Figure 2). The functionality a vs. us, plotted in Figure 4, is approximately linear; thus, the gas-side mass transfer coefficient, k,, seems not to be affected by the superficial gas velocity. This has also been quoted by Mashelkar (1970). It is interesting to compare the experimental values with theoretical predictions for mass transfer inside a completely stagnant bubble. Solution of the equation for transient diffusion in spherical geometry gives (Sherwood et al., 1975)
where kg is the average gas-side mass transfer coefficient corresponding to a logarithmic mean of inlet and outlet values of p , D is the molecular diffusion coefficient of formaldehyde in nitrogen, and t is the contact time, given by t = €Z/U, (7)
34
Ind. Eng. Chem. Process Des. Dev., Vol. 22, No. 1, 1983
Table 11. Experimental and Theoretical Gas-Side Mass Transfer Coefficients (mol/cm3 s atm) run A B
c
D E F G
kg ( a p t l ) 2.56 x 10-5 2.37 x 10-5 2.66 x 10-5
2.60 x 10-5 3.85 x 10-5 1.07 x 2.56 x 10-5
kpa (exptl) 3.14 2.90 3.17 2.81 4.36 1.42 3.05
x x x x x x x
10-5 10-5 10-5 10-5 10-5
10-5
I 2
kpa (theor) 2.42 2.42 2.35 2.12 7.19 0.93 2.00
x x x x x x x
3
10-5 10-5 10-5 10-5 10-5 10-~ 10-5
Taking into account that terms beyond the first one are negligible, eq 6 leads to kp(theor) = (1/RT)((u,/Z)(ln ( r 2 / 6 ) )+ 4r2Dt/dB2) (8) arises from the On the other hand, the experimental balance
6
-
Q N ( Y-~Yz) = k,aZ
(PI
- p2)/ln
Wp2)
(9)
50
i i
0
I
I
200
80
t (min)
L
Figure 5. Increase in the liquid viscosity, measured at 30 " C , as a function of reaction time at 80 "C and different formaldehyde flow rates (g/h): (1) 60.9; (2) 44.0; (3) 21.8.
G G
(exptl) may be easily calculated From eq 2,4, and 9, as a function of kp (exptl), Yl, and Y2. In order to estimate (theor) from eq 8, diffusion coefficient were calculated using the Chapman-Enskog equation (Bird et al., 1960). The Lennard-Jones constants were obtained by using critical parameters of the formaldehyde molecule (Brandani et al., 1980). This gave D(80 "C) = 0.184 cm2/s D(60 "C)= 0.165 cm2/s
I
100
I
I
3
(10)
The Sauter mean bubble diameter was calculated from Figure 3, and the holdup from Hughmark's correlation (Figure 2). Table I1 shows the comparison between experimental and theoretical mass transfer coefficients. There is satisfactory agreement in the order of magnitude, giving indirect evidence of the gas phase control in the absorption process. It is worthwhile to point out that the gas-phase control is not the usual situation in deep-bed contactors. Sharma and Mashelkar, as quoted by Mashelkar (1970), are the only workers who had given some results on the determination of gas-side mass transfer coefficients in bubble columns. They had also found good agreement between experimental results and the stagnant sphere model. The Chemical Reaction. The reactions taking place between phenol and formaldehyde under alkaline conditions are well known (Knop and Scheib, 1979). The first step is the formation of methylol compounds by addition of formaldehyde to 0,o'andlor p positions of the phenolic ring. The second step is condensation through methylene or ether linkages, leading to a mixture of mono- and polynuclear methylolated molecules, called resol. The properties of commercial resols are strongly dependent on the particular application. Usual properties are: formaldehyde to phenol molar ratio, F I P = 1to 1.5, % solids content = 40 to 80, viscosity (cP)at room temperature = 10 to 3500, storage stability = 3 weeks to 12 months. When the reaction is carried out in a bubble column, the required resol type arises from the following operating conditions. (1)FIP is set from the formaldehyde flow rate (&/ Y J ,the initial phenol amount, and assuming complete absorption. This, in turn, determines the reaction time. (2) The final viscosity and the % solids content are fixed by varying both the reactor temperature and the amount of water deleted with a partial condenser. For example, Figure 5 shows the increase in viscosity, measured at 30 "C, as a function of time for three different
Figure 6. Rate of Viecosity increase at 74 "C for different resols; (1) = 1100 cP, % sol = 72; (2) F/P = 1.23, h , l m a c = 160 cP, % sol = 64; (3) F/P= 1.43, /.tolmoc = 50 cP, % sol = 54.
F/P = 1.04, j&ec
formaldehyde flow rates. Reaction temperature and carrier flow rate were kept constant. As expected, the greater the formaldehyde flow rate, the steeper the viscosity increase. Properties of the Resols. Products obtained in the bubble column were characterized by measuring several properties. For molar ratios F I P in the range 1-1.5, the free formaldehyde amount was negligible. Only when FIP was set equal to 3.4 was a significant amount of free formaldehyde present (7.39%). Figure 6 showa the viscosity increase at 74 "C in a plate and cone viscometer, as a measure of curing rates for different resols. Curing rate increased with FIP, po, and % solids. Product 3, obtained at 60 "C, had a high molar ratio FIP but a relatively low viscosity. Both products 1 and 2 were obtained at 80 "C. The viscosity and % solids of product 1were increased by deleting water from the reaction mixture, using a partial condenser. Note the process versatility with regard to the type of resol which is desired. Figure 7 shows the IR spectrum of a typical resol arising from the bubble column. Bands were assigned according to Secrest (1965) and Colthup et al. (1975). Important bands are indicated in Table 111. Methylolated phenolic compounds were present, together with a small amount of free phenol. Ether linkages could not be detected (band
Ind. Eng. Chem. Process Des. Dev., Vol. 22, No. 1, 1983 35 moo0
t
I
2
I lb00
1400
1200
1000
8M
600
3 (cm-1 Figure 7. IR spectrum for a resol characterized by F/P = 1.43, hlmec = 160 cP, % sol = 56. 01
25
Figure 9. Increase in resol viscosity, measured at 30 O C , a8 a function of the storage period at 20 OC: (1) F/P = 1.04,~101s-c= 1100 cP, % 501 = 72; (2) F/P= 1.33,p&,mc 1380 cP, % 801 68; (3)F/P= 1.43,h l ~ e =c 160 cP, % 801 = 56; (4)F/P= 1.52,h l a o c = 110 cP. % sol = 64.
50
IO0
150
~ ( y ) 200
Figure 8. DSC thermogram for a resol characterizedby F/P= 1.43, hlS0.c 160 cP, % 501 56. Table 111. Description of Some of the IR Bands Shown in Figure 7
band, cm-' 1510 1470 1463 1040 1020 1000 875-885 825 770-785 765 690
origin
PhOH substituted in 0,p, and o , p positions PhOH substituted in o,o', and o,o',ppositions methylene linkage -CH,PhOH substituted in an o position PhOH substituted in a p position methylol -CH,OH PhOH substituted in o,o',p positions PhOH substituted in p and o , p positions methylene linkage -CH,PhOH substituted in o and o,o' positions free PhOH
at 1065 cm-l). After curing at 150 "C a significant intensity decrease of bands at 690,755,825,1000, and 1510 cm-' was observed, indicating that further condensation took place. A DSC thermogram of the same resol, after a 3 months storage, is plotted in Figure 8. The Curing is characterized by the exothermic band at 150 "C. The reaction heat was estimated as (-AH) = 261 J / g of dry resin. The storage stability of several products at room temperature is shown in Figure 9. Stability decreases with an increase in FIP,b,and % solids. This is in agreement with previous reported results (Adabbo et al., 1979). Conclusions The technical feasibility of producing resols in gas-liquid contaeting devices has been shown. The main advantages are: (a) control of the reactor temperature resulting from the limitation of the reaction rate by the formaldehyde supply; (b) versatility for obtaining products of different F/Pmolar ratios, initial viscosities and 70solids; (c) high
efficiency of the absorption process for exhausting almost all the formaldehyde from the inlet gases; (d) decrease in operation cost with respect to the batch process (absence of an intermediate absorption step and increase in the reactor output). The formaldehyde absorption rate in a bubble column was controlled by mass transfer in the gas phase. The volumetric mass transfer coefficient could be estimated with equations describing transient diffusion in spherical geometry. Characteristics of resols obtained in a bubble column were the usual ones of commercial resols. Acknowledgment The authors wish to thank the Comisidn de Investigaciones Cientificas de la Provincia de Buenos Aires (Argentina) for financial support. Nomenclature a = interfacial area per unit volume D = molecular diffusion coefficient of formaldehyde in nitrogen dB = Sauter mean bubble diameter F = moles of formaldehyde H = enthalpy 4, = local gas-side mass transfer coefficient k, = average gas-side mass transfer coefficient corresponding to a logarithmic mean of inlet and outlet values of p K, = overall mass transfer coefficient in gas phase units P = moles of phenol p = formaldehyde partial pressure p* = formaldehyde partial pressure corresponding to the equilibrium with the liquid phase pt = total pressure QN = molar flow rate of nitrogen per unit transversal area of column R = gas constant Reo = Reynolds number at the single-orificesparger t = contact time T = temperature u, = superficial gas velocity Y = molar ratio, mol of formaldehyde/mol of nitrogen t = axial coordinate Z = height of the expanded liquid bed Greek Symbols t = holdup p = viscosity
Xnd. Eng. Chem. Process Des. Dev. 1985, 22, 36-38
38 p u
= density = surface tension
Subscripts 0 = initial value w = water Registry No. Formaldehyde-phenol copolymer, 9003-35-4.
Literature Cited Adabbo. H. E.; Wit, E. M.; Robs, A. J.; Williams, R. J. J. Rev. flast. Mod. (Spain) 1979. 277,70. Bkd. R. B.; Stewart, W. E.; Lightfoot, E. N. “Transport Phenomena”; Wlley: New York. 1960 pp 16-19. Brandani, V.; DI Giacomo, G.; Foscolo, P. U. Ind. €ng. Chem. frocess D e s . Dev. 1960, 79, 179. BrunnmClller, F. La BASF 1975, 25, 29. Colthup, N. B.; Dab, L. H.; Wlberley, S. E. “Introduction to Infrared and Raman Spectroscopy”, 2nd ed.; Academic Press: London, 1975; p 395.
Crusos. A. Meter. flast. (Bucherest) 1976, 75, 160. Haslam. J.; Soppet, W. W. J . Appl. Chem. 1953, 3,328. Hughmark. G. A. Ind. Eng. Chem. RocessDes. Dev. 1967, 6 , 218. Ingraham, T. R.; Fraser, D. Proceedings 3rd Toronto Symposium Thermal Analysls, Toronto, 1969 Chemlcal Institute of Canada; X , 101. Knop, A.; Schelb. W. ”Chemistry and AppllcauOn of Phendlc Resins”, Polym. Proper. Appl. 3; Springer-Verlag: Beriln-Heidalberg, 1979; p 35. Mashelkar, R. A. Br. Chem. Eng. 1970, 15, 1297. Secrest, P. J. J . Paint Tech&. Eng. 1965, 37, 187. Sherwood, T. K.; Pigford, R. L.; Wllke, Ch. R. “Mass Transfer”; McGraw-Hill: New York, 1975; p 227. Strelcow. W. I.; Zukowskala, N. S.; Mujina. L. A. Zavcd. Lab. 1972, 7 , 78. Struplnskaya, 0. P.; Ivanov, P. S.; Anlslmova, A. V.; Demkln, V. M.; Struplnskll, V. A. flast. Massy 1966, 72, 18 (Eng. Transi. RAPRA 44 C11-723). Urbanski, J.; Czerwlnskl, W.; Janlcka. K.; Majewska, F.; Zowall, H. Handbook of Analysis of Synthetlc Polymers and Plastics”; Ellis Horwood: England and Wydawnlctwa Naukowo-Technlczne: Poland, 1977; p 28.
Received for review June 22, 1981 Accepted June 14,1982
Produckrg Fermented Starches by Accelerated Processes and Their Use as Yarn Sizing Agents Galowrt M. Elgal’ and Rkardo H. Wade
‘ U.S. Department of Agricutture, Agricuttural Research Service,
Southern Regional Research Center, New Orleans, Louisiana 70179
When ordinary starch is boiled and applied to yarn as sizing agent, it forms a water-insoluble film that requires an energy-consuming process to desire. However, degraded starch can be used, which is soluble in cold water and requires less energy to wash it off. One of the processes that produces degaded starch with suitable properties is fermentation;however, normally it reqolres approximately two to three weeks for process completion. This project investigated methods that reduced the fermentation time to 2 to 3 days. Experimentally it was determined that degraded starch from either the slow or accelerated process produces a product with equivalent sizing properties.
Introduction In a previous publication (Elgal et al., 1978) the application of fermented starches as sizing agents was reported as a potential commercial process. The main purpose of applying sizing agents to textile yarns is improved tensile strength and abrasion resistance. These two yarn qualities are essentiaI for preventing yarn deterioration and breakage during weaving. Ordinarily, after being woven, the fabric is desized because the sizing agent impedes further textile finishing. The use of fermented starch sizing agents allows their removal without use of expensive steam and hot water. Efforts in the current study have been to shorten the period of starch fermentation from 2 to 3 weeks to a few days. Two basic processes were investigated, natural aerobic fermentation with bacteria originating from the air with baker’s yeast. The degraded starches produced by both processes were found to be interchangeable in application. The principal use of these degraded starches in this effort was the application on yarn as sizing agent. Because this sizing agent is soluble in ambient temperature water, desizing can be accomplished with reduced energy consumption. The quality of products from these starch degradation processes was evaluated by utilizing yarn ‘One of the facilities of the Southern Region, Agricultural Research Service, U S . Department of Agriculture.
testing instruments and viscosity measurements.
Starch Fermentation The concentration of starch solution utilized throughout this study was 12% by weight in water. The starch used was Crown thin boiling starch (manufactured by Penick and Ford). Lower grade starches may be used to achieve the fermentation more economically. The powdered starch was thoroughly mixed in ambient temperature water. This mixture was then heated to near its boiling point of 90-100 “C and held at that temperature for 5 min with stirring. The resulting solution was used to prepare degraded starches. (The term “solution” is used qualitatively in this writing and may be subject to argument whether the mixture is a “true” solution.) To prepare fermented starch, the above conventional starch solution was allowed to cool, loosely covered to permit venting of gases, and then maintained at ambient temperature during the fermentation process. Daily stirring and aeration with a spatula enabled bacteria from the air to mix into the solution. The solution pH was recorded as the fermentation progressed and produced acid. Butyric acid was identified by odor and by gas chromatographic methods. The pH stabilized at approximately 3.5 (Figure 1, upper curve) with the effect of controlling any further fermentation. The pH was used as a measure of completion.
This article not subject to U.S. Copyright. Published 1982 by the American Chemical Society
