Flash Extraction of Pectin from Orange Albedo by Steam Injection

Biomacromolecules 2003, 4, 880-889

880

Flash Extraction of Pectin from Orange Albedo by Steam Injection† Marshall L. Fishman,*,‡ Paul N. Walker,§ Hoa K. Chau,‡ and Arland T. Hotchkiss‡ Eastern Regional Research Center, Agricultural Research Service, U.S. Department of Agriculture, 600 East Mermaid Lane, Wyndmoor, Pennsylvania 19038 and Pennsylvania State University, Department of Agricultural and Biological Engineering, 223 Agricultural Engineering Building, University Park, Pennsylvania 16802 Received November 12, 2002

Pectin was acid extracted from orange albedo by steam injection heating under pressure. Extraction times ranged from 2 to 6 min at a pressure of about 15 psi. Solubilized pectin was characterized by HPSEC with online light scattering and viscosity detection. Molar mass (M), radius of gyration (Rg), and intrinsic viscosity ([η]) all decreased with increasing extraction time when heating temperature was 120 °C. At heating times of 3 min, Mw ranged from 4.9 to 4.5 × 105, Rgz was about 44 nm, and [η]w ranged from 8.4 to 7.9 dL/g. Chromatography revealed that solubilized pectin distributions were bimodal in nature at 3 min extraction time and trimodal when the extraction time was 6 min. Scaling law exponents obtained for the highest molar mass fractions were consistent with a very compact spherical structure. For the intermediate fraction, scaling law exponents were consistent with a less compact spherical structure comparable to a random coil. In the case of the low molar mass fractions, scaling law exponents were consistent with a structure more asymmetric in shape. These results are consistent with earlier results which indicated that pectin distributions were mixtures of two or more of the following due to disaggregation during extraction: spherical aggregates, hydrogen bonded network structures, and partially or fully disaggregated components of network structures which could include branched structures, rods, segmented rods, and kinked rods. Introduction Pectin is an important food hydrocolloid in that it functions in processed and unprocessed foods as a gelling and texturizing agent.1 Much of pectin’s value as a gel arises from its spreadability and interaction with sugar. Furthermore, pectin and/or its fragments have been reported to possess nutraceutical activities that include immunostimulation, anti-metastasis, hypoglycemic, and cholesterol lowering effects.2 Pectin is distinguished chemically from other polysaccaharides in that it is mainly a copolymer of R 1-4 linked galacturonate and its methyl ester. Neutral sugars present include rhamnose, galactose, arabinose, and xylose. Biologically, pectin functions as an important cell wall component in most higher plants. Citrus fruits and apples are the main sources of food grade pectins, and these pectins are comprised of 85-95% galacturonans. Lime peels are considered the best sources of citrus pectin followed by lemon, grapefruit, and orange peels in that order.3,4 Unfortunately, pectin from orange peels, the largest potential source of citrus pectin, is considered to have the poorest structure/function properties among citrus pectins. †

Mention of brand or firm name does not constitute an endorsement by the U.S. Department of Agriculture over others of a similar nature not mentioned. * To whom correspondence should be addressed. Phone: 215-233-6450. Fax: 215-233-6406. E-mail: [email protected]. ‡ Eastern Regional Research Center. § Pennsylvania State University.

Nevertheless, we have shown that the flash extraction of orange albedo pectin using microwave heating under pressure produced pectin that was comparable or better in structural properties (i.e., molar mass, M, intrinsic viscosity, [η], or root-mean-square radius of gyration, Rg) than commercially produced pectin from lime and/or lemon peels.5,6 Optimized yields were lower than commercial citrus pectins, but heating times were shorter. Typically, yield and molecular properties were optimized after three minutes of microwave heating in aqueous HCl initially at pH 2, at a pressure of about 22 psi above ambient, and a temperature of about 120 °C. The objective of this work was to determine whether the quality and yield of pectin produced in the previous work5 was a unique result of using microwave energy to extract the pectin under optimum conditions of heating time, pressure, and temperature or whether a comparable pectin could be obtained by using another source of heat, in this case rapid steam injection under a similar regime of heating time, pressure, and temperature. Materials and Methods Pectin. A commercial citrus pectin, degree of methyl esterification (DE) of 72%, galacturonic acid content 74%, was obtained from a pectin manufacturer and used without further purification. Albedo. Fresh commercial albedo (C) was obtained from Florida early Valencia oranges (EVO). Upon arrival, a small

10.1021/bm020122e CCC: $25.00 © 2003 American Chemical Society Published on Web 05/20/2003

Flash Extraction of Pectin from Orange Albedo

Figure 1. Schematic of the extraction vessel.

amount of residual flavedo and pulp sacs were stripped from the albedo with a paring knife. After cutting the albedo into small pieces it was stored at -20 °C in sealed polyethylene bags until extraction. Fresh lab albedo (L) was excised from the peels of Florida EVO as described previously.5 Extraction Steam Injection Heating. A 6 L pressure cooker was modified to serve as a pressure vessel as illustrated in Figure 1. Clean steam was injected at the bottom of the vessel via a 6 mm diameter line and ball valve. The clean steam was produced from deionized water and was delivered at approximately 17.2 psi pressure above ambient. A second valve was added to the pressure vessel to vent air and steam. The pressure gauge and a type T thermocouple temperature probe, the latter connected to a data logger, were used for monitoring conditions inside the vessel. Polyisocyanurate insulation 50 mm thick covering the bottom of the vessel limited heat loss to the supporting bench. Except for seals and other minor components, the pressure vessel, valves, pressure gauge, and temperature gauge were stainless steel, whereas the steam line was polypropylene. The pressure vessel was preheated to the desired treatment temperature by opening the steam valve and allowing steam to enter the pressure vessel, first with the vent valve open to expel air and, as the temperature approached the treatment temperature, with the vent valve closed to pressurize the vessel with steam. The steam valve was then closed, the steam vented, the pressure vessel lid removed, and any condensate removed. Simultaneously, about 3/4 of a 1 L solution of HCl, initially at pH 2, was poured into a 2-L beaker containing 40 g of albedo sample. The mixture was manually swirled for approximately 3 s to suspend the albedo and then placed into the preheated pressure vessel. Immediately the remainder of the 1 L acid solution was added to the beaker, swirled, and placed into the vessel to transfer remaining albedo. The vessel was then closed and the steam valve opened to inject steam into the mixture, with the vent valve remaining fully open to vent air until the temperature

Biomacromolecules, Vol. 4, No. 4, 2003 881

of the mixture reached 95-100 °C. At that point, the vent valve was closed, and subsequently, the steam valve was closed as temperature of the vessel approached the desired treatment temperature. The steam valve was repeatedly opened and closed, and occasionally, the vent valve was cracked open through the remainder of the experiment to maintain the desired temperature, while at the same time ensuring that the pressure did not exceed 15 psi above ambient which was judged to be the safe limit of the vessel. Treatment time began when the steam valve was initially opened and at the end of the prescribed treatment time, measured with a stop watch, the steam and vent valves were closed, the insulation was removed, and the vessel was plunged into an ice bath. The pressure rapidly declined due to condensation, and as the pressure approached atmospheric, the vessel was raised slightly to avoid submersion of the lid and possible water leakage into the vessel as the pressure continued to drop to partial vacuum conditions. The vent was opened to relieve the vacuum, all the while the vessel was rocked in the bath to increase the rate of cooling. When the temperature of the mixture fell to 37 °C, the vessel was removed from the ice bath and poured into a beaker. Samples were filtered with miracloth. Pectin was precipitated by adding one part of 70% isopropyl alcohol (IPA) to two parts filtrate. Then precipitated pectin was washed first with 70% IPA followed by a second washing with 100% IPA. Finally, the sample was vacuum-dried at room temperature and prepared for chromatography. Single pass extractions were run in triplicate, whereas sequential extractions were run in duplicate. Duplicate samples were combined and analyzed in triplicate. Microwave Heating. This method of extraction has been described previously.5,7 Briefly, microwave heating was performed in a CEM model MDS-2000 microwave sample preparation system. Samples were irradiated with 630 W of microwave power at a frequency of 2450 MHz. For each experiment, six equally spaced cells were placed in the sample holder, a rotating carousel. One vessel was equipped with temperature and pressure sensing devices which measured and controlled the temperature and pressure within the cell. Time of irradiation was 3 min at which time the pressure was about 22 psi above ambient and the temperature 120 °C. Cells were loaded with 1 g of albedo dispersed in 25 mL of pH 2.0 HCl solution. Solubilized pectin was isolated as described above and prepared for chromatography. Chromatography. Chromatography has been described elsewhere.6 A brief description follows. Depending on molar mass, the dried sample at a final concentration of 1 or 2 mg/mL was dissolved in 0.05 M NaNO3, stirred until dissolved, centrifuged at 50 000 g for 10 min at 35 °C, and filtered with a 0.22 or 0.45 µm millex HV filter (Millipore Corp.). The injection volume was 200 µL. Samples were run in triplicate. The flow rate was 0.70 mL/min. The solvent delivery system consisted of a model 1100 series degasser, pump, and auto sampler (Agilent Technologies). Pectins were separated by two PL-Aquagel OH-60 and one PL Aquagel OH-40 size exclusion columns in series. Column effluents were detected by a Dawn DSP multi-angle laser light scattering (LS) photometer (Wyatt Technology), model 100

882

Biomacromolecules, Vol. 4, No. 4, 2003

Figure 2. Temperature and pressure as a function of heating time during steam extraction of pectin albedo. Sample load, 40 g albedo/L of HCl initially at pH 2.

differential pressure viscometer (DPV) (Viscotek Corp.), and an Optilab DSP interferometric refractometer (DRI) (Wyatt Technology) in series. The electronic outputs from the detectors were connected to separate serial ports in the same personal computer in a manner which permitted data to be collected and processed by ASTRA (Wyatt Technology) and TriSec (Viscotek Corp.) software simultaneously.5 The change in refractive index with polymer concentration, dn/ dc, in NaNO3 for pectin from orange albedo was measured offline using the HPLC pump, the auto injector, and the Optilab refractometer. The value found for orange albedo pectin was 0.130 mL/g. Sugar Composition of Pectin. The percentage of galacturonic acid8 and degree of methyl esterification9 were determined for selected pectin samples. Results and Discussion Figure 2 contains a plot of the temperature and pressure against time, as measured in the pressure vessel during the extraction process. Time zero is just after the steam valve was opened following the addition of the albedo and HCl solution. At 119 °C maximum temperature, the pressure reading was about 15 psi. The steam saturation pressure at 119 °C is 13.1 psi which is roughly consistent with the data considering the pressure gage was read manually and considering the expected heat flux from the headspace to the liquid (temperature sensor location). The albedo/acid solution ratio (w/v) was 1:25. The temperature, pressure range, pH of the acid solution, and the albedo/acid ratio were chosen to emulate reaction conditions which were found to give optimum molecular properties for orange albedo pectin when extracted by microwave heating (MH) under pressure.5 The time course of temperature and pressure were found to be comparable for steam injection heating (SIH) under pressure as compared to those found in the MH experiments with a few minor variations. In the case of the MH, the pressure started to rise slowly almost immediately when heating commenced, whereas in the present study, the pressure did not start to rise until the reaction mixture temperature reached about 87 °C. For pectin extracted under MH, after 3 min, the pressure was 22 psi and the temperature

Fishman et al.

Figure 3. Percentage recovery of pectin from fresh albedo as a function of heating time. Sample load, 40 g albedo/L of HCl initially at pH 2. Table 1. Effect of Time on Chemical Composition of Pectins Steam Extracted at 120 °C sample time (min)

DEa

% GAb

% NSc

2(A)d

60(1)f

(B)d 3(A) (B) 6(A) (B)

77(3) 70(2) 75(3) 87(3) 65(2)

86(1) 76(2) 81(1) 77(2) 72(1) 83(2)

8.0(0.4) 10.9(0.4) 10.1(0.5) 7.6(0.6) 6.5(0.3) 8.0(0.2)

a Degree methyl esterification. b Percentage galacturonic acid in pectin (w/w). c Percentage neutral sugars in pectin (w/w). d “A” indicates the pectin fraction that floated, and “B” indicates the pectin fraction that sank to the bottom when alcohol was added. f Standard deviation of triplicate analysis.

was about 120 °C. For SIH after 3 min, the pressure was about 14.3 psi. The lower pressure in the SIH experiments was necessitated by the inability of the extraction vessel to withstand pressures greater than 15 psi. In the MH experiments, 6 vessels, 120 mL in volume, containing 25 mL of solvent were used whereas in the SIH experiments, 1 vessel, 6 L in volume containing 1 L of solvent was employed. Based on previous MH studies to extract pectin, pectin was extracted from EVO albedo using injected steam for various lengths of time, at maximum temperatures of 110 and 120 °C, respectively. At both temperatures, the pressure fluctuated between 14 and 15 psi. As shown by the data in Figure 3, pectin recovered as a percentage of dry albedo starting material increased linearly with time of extraction. Comparison of recoveries at 120 and 110 °C by the Student t test revealed that at 3 and 6 min recoveries were different with a confidence level of 92% or higher. Furthermore, the slopes of the lines differed with a confidence level greater than 95% as determined by the Student t test. Thus, we conclude that at heating times greater than 2 min a higher percentage of pectin was obtained at 120 °C than at 110 °C. Except for the data at 2 min, a higher percentage of pectin was obtained at 120 °C than at 110 °C. The effect of time on DE, % galacturonate (%GA), w/w, and % neutral sugars (%NS), w/w, of pectins which were steam extracted at 120 °C maximum is found in Table 1. It appears that samples from fractions which floated to the top when alcohol was added (“A” samples) DE increased with heating time whereas for the samples from fractions which



Table 2. Effect of Time and Temperature on Molecular Properties of Steam Extracted Pectins sample temp (°C)/ time (min)

Mw × 10-3

2/110(A)c

499(6)d

(B) 3/110(A) (B) 4/110(A) (B) 5/110(A) (B) 6/110(A) (B) 2/120(A) (B) 3/120(A) (B) 6/120(A) (B)

429(12) 434(18) 423(3) 478(7) 509(8) 415(3) 487(7) 414(20) 367(20) 403(13) 448(8) 491(10) 450(10) 355(10) 393(10)

MALLS Rgz (nm)

ba

Mw × 10-3

Rgz (nm)

LS/V [η]w (dL/g)

ab

48.7(0.2) 50.5(0.3) 47.0(0.5) 47.0(0.2) 47.6(0.4) 46.5(0.6) 50.2(0.4) 43.7(0.1) 46.2(0.3) 44.4(0.7) 47.2(0.3) 46.6(0.2) 44.3(0.4) 43.9(0.9) 36.1(0.1) 36.0(0.7)

0.23(0.01) 0.28(0.01) 0.26(0.02) 0.24(0.02) 0.21(0.02) 0.20(0.02) 0.22(0.01) 0.20(0.01) 0.22(0.03) 0.22(0.03) 0.31(0.01) 0.25(0.01) 0.16(0.01) 0.19(0.01) 0.12(0.02) 0.12(0.01)

923(55) 689(61) 664(50) 692(22) 873((3) 976(4) 627(6) 874(40) 648(60) 508(40) 585(10) 744(74) 931(20) 750(70) 571(30) 584(40)

80(2) 74(4) 67(4) 71(2) 77(1) 80(1) 68(1) 75(2) 68(3) 58(4) 65(9) 72(4) 77(2) 69(1) 56(3) 54(2)

9.9(0.2) 10.6(0.3) 8.5(0.1) 9.1(0.2) 8.8(0.1) 8.9(0.2) 9.0(0.2) 8.7(0.9) 8.4(0.3) 7.4(0.2) 8.8(0.3) 8.9(0.1) 8.4(0.4) 7.9(0.1) 5.7(0.1) 5.2(0.1)

0.50(0.01) 0.54(0.02) 0.55(0.4) 0.54(0.01) 0.51(0.01) 0.48(0.01) 0.53(0.02) 0.46(0.01) 0.54(0.02) 0.56(0.04) 0.56(0.03) 0.53(0.01) 0.46(0.01) 0.50(0.01) 0.51(0.04) 0.52(0.01)

a Scaling law exponent of log R against log M plot. b Mark-Houwink exponent. c “A” indicates the pectin fraction that floated, and “B” indicates the g pectin fraction that sank to the bottom when alcohol was added. d Standard deviation of triplicate analysis.

sunk to the bottom (“B” samples) DE decreased with heating time. In the case of %GA, the value for “A” samples decreased with heating time, whereas %GA increased for “B” samples with heating time. For “A” and “B” samples, %NS did not show a discernible trend. These trends in DE and %GA were not discernible in flash extraction experiments with microwave heating. Furthermore, when heated by microwaves, the entire sample floated to the top of the beaker when alcohol was added.5 The effect of time and temperature on the molecular properties of pectin extracted from fresh commercial albedo are found in Table 2. The weight average molar mass (Mw) and the Z-average root-mean-square radius of gyration (Rgz) values were obtained simultaneously by the MALLS method and by the LS/Viscometry (LS/V) method.5 One obtains Mw and Rgz values from MALLS by measuring excess scattering intensity due to dissolved macromolecules at 15 scattering angles and plotting K*c/RΘ against sin2(Θ/2).10 As indicated by eq 1 K*c/RΘ ) 1/M + (16Π2/3λ2)(Rg2/M) sin2 (Θ/2)

(1)

M is obtained from the intercept and Rg from the slope of such a plot. K*, the optical constant, is obtained from eq 2 K* ) 2Π2no2(dn/dc)2/λ4NA

(2)

where no is the refractive index of the solvent, dn/dc is the change in refractive index with polymer concentration, c, at the wavelength, λ, of the scattered light, and NA is Avogadro’s number. The Rayleigh ratio, RΘ, is the scattering intensity at angle Θ divided by the intensity of incident light. One obtains M and Rg values from LS/V by an iteration method. Because M is estimated using the excess light scattering only at 90°, an alternate form of eq 1 is required, namely11 K*c/RΘ) 1/(MPΘ)

(3)

Initially PΘ at 90° is assumed to be 1, and the molar mass is calculated from eq 3. A new value of P90° is calculated by

assuming that the polymer under study is either a rod, random coil, or sphere and using an appropriate equation for the shape model chosen to obtain P90° from Rg.12 Rg is obtained using the Ptitsyn-Eisner13 modification of the Flory-Fox14 equation which relates [η] to Rg. Generally, three iterations are required before PΘ and M reach constant values. The approach of using molar mass, viscosity, and concentration detection in conjunction with HPSEC allows one to obtain the entire polymer distribution and various averages of the distribution properties such as the number, weight, and Z average M, Rg, and [η]. The LS/V method (aka triple detector method or Sec3 method) requires light intensity at only one scattering angle, typically 90°. The data in Table 2 reveal that LS/V values of Mw and Rgz are systematically higher than for the values from MALLS. We found a similar result for pectin that was flash extracted by microwave heating as well.5 However, the differences are greater in this study. In Figure 4A are typical superimposed chromatograms of the scattering intensity at 90° (LS), the differential pressure produced by viscosity (DPV), and refractive index (DRI) for pectin. The pectin was extracted from orange albedo after 3 min of heating by steam injection at a final temperature of 120 °C and a pressure of about 14.3 psi. The shape differences between the LS and DVP chromatograms may be interpreted as high molar mass components which contribute proportionately more area to the LS chromatogram than to the DPV chromatogram. This effect is due to molecules of equal size which exhibit different sensitivity of one technique over the other. Differences in shape between LS and DPV chromatograms could be an indication that molecules with polydispersity in shape and/or molar mass are eluting from the columns within the same size fraction. One possible result from such polydispersities would be that the molar mass obtained by the LS/V method would be too high. Figure 4B obtained using the MALLS method demonstrates a well behaved angular dependence for the excess scattering of the polymer molecules at the peak of the LS chromatogram (top). Also shown are the superimposed LS

884


Fishman et al.

Figure 4. Chromatograms of pectin which was heated to 120 °C. During extraction sample was heated for 3 min and floated on precipitation with 70% alcohol. For chromatography: sample concentration 1 mg/mL; mobile phase, 0.05 M NaNO3; flow rate 0.7 mL/min; injection volume, 200 µL; column temperature, 45 °C. (A) Generated by TriSec software. LS, scattering intensity at 90°, the DPV and the DRI. (B) Generated by ASTRA software. LS, scattering intensity at 90° and the DRI (lower). Zimm plot of peak slice marked by vertical line (upper).

and DRI chromatograms and the point (vertical line) in the distribution of pectin which is analyzed for its angular dependence (bottom). The Mw and Rgz data plotted in Figure 5 were obtained by MALLS. Figure 5A contains pectin weight average molar mass, Mw, and weight average intrinsic viscosity, [η]w, plotted against heating time of extraction for pectins extracted at a maximum of 110 °C. As mentioned previously, “A” pectin floated to the top of the beaker, whereas “B” pectin sunk to the bottom of the beaker when the precipitant, IPA, was

added. As indicated by the data in Figure 5A, Mw and [η]w values, respectively, appeared to show no trend with particle density for A and B pectins extracted for the same amount of time. On that basis and because of their relatively close values, it was concluded that we were not able to measure appreciable differences in Mw and [η]w values for these pectins. Possibly the variance in particle density among precipitated pectins is related to differences in the kinetics of precipitation. Over the course of heating times, i.e., between 2 and 6 min, Mw appeared to have decreased by



Figure 5. Effect of heating time on pectin properties. (A) Weight average molar mass and intrinsic viscosity, albedo was heated to 110 °C. (B) Weight average molar mass and intrinsic viscosity, albedo was heated to 120 °C. (C) Molar mass calibration curves superimposed upon differential molar mass distributions for pectins which floated when alcohol added, albedo was heated for 2, 3, and 6 min to 120 °C. (D) Similar to C except pectin sank to bottom when alcohol added. (E) Z average radii of gyration, albedo was heated to 110 and 120 °C.

886


about 2-3% and [η]w decreased by about 15-30% for both samples. Between 3 and 6 min, Mw and possibly [η]w pass through a broad maximum. Figure 5B contains pectin Mw and [η]w plotted against heating time of extraction for pectins extracted at a maximum of 120 °C. In these experiments, between 2 and 6 min of heating time, Mw and [η]w decreased by about 12 and 35-42% respectively. The superimposed curves of molar mass and refractive index chromatogram against elution time for these pectins are shown in Figure 5, parts C and D. The molar mass (i.e., calibration curves) are displaced to longer elution volumes as the heating time of extraction is increased. This separation between calibration curves is larger at lower elution times. These curves indicate that, with increased heating time, the molar mass at the high end of the distribution increased but their size decreased. The changes in the course of the calibration curves are consistent with the interpretation that some of the pectin molecules fragment and re-aggregate with heating time. Because Mw decreased with heating time, a larger portion of the pectin molecules fragmented and decreased their molar mass but did not re-aggregate or re-aggregated to a molar mass lower than their original value. By way of contrast, over the range of 2.5 and 6 min of heating during extraction by microwaves, Mw and [η]w decreased by about 85 and 87%, respectively.5 The microwave experiment also had a maximum temperature of about 150 °C and a maximum pressure of about 50 psi. Figure 5E contains pectin Z-average radii of gyration plotted against heating time of extraction for pectins extracted at maximum temperatures of 110 and 120 °C. The radii of pectins extracted at 110 °C are larger than those extracted at 120 °C. At 110 °C, over a 4 min time period, the radius decreased in length by about 5-12%, whereas at 120 °C, it decreased by about 23-24%. The data in Figure 3 show that for the most part the extracted pectin yield increased with increased time of heating and temperature, whereas the data obtained concurrently in Figures 5 show that Mw, Rgz, and [η]w decrease. It appears that heating pectin for 3 min at 120 °C provides the best compromise of “good” molecular properties and yield. As indicated in Table 1, if we average values for samples A and B for the pectin extracted for 3 min at 120 °C, DE was about 73%, %GA was about 79%, and %NS was about 9%. For a commercial rapid set pectin, we found 78%, 81%, and 11% for these values, respectively. In the case of molecular parameters found in Table 2, for the pectin extracted for 3 min at 120 °C, Mw was about 470 000, Rgz was about 44 nm, and [η]w was 8.8 dL/g when A and B values were averaged. In the case of commercial citrus pectin, we found 254 000, 35 nm, and 6.1 dL/g for Mw, Rgz, and [η]w, respectively. In an effort to increase yield and without diminishing molecular properties, pectin was sequentially extracted from the same albedo sample 3 times. Sequential extraction increased the yield of microwave heated albedo pectin from 11.3% to 15.3%. In this case, extractions were for 3 min at 120 °C. After each heat treatment, as described in the section on extraction, pectin was isolated from acid solvent. In each step, the albedo or albedo residue was extracted with fresh acid. After the 3 step extraction, the

Fishman et al. Table 3. Recovery of Pectin from Sequential Extractiona number of sequence

percentage of dried albedo

1 2 3 total

6.9(0.6)b 7.5(0.3) 2.5(0.5) 16.9(1.4)

a Temperature 120 °C; time 3 min/sequence. b Standard deviation of triplicate analysis.

Figure 6. Effect of heating times on Mark-Houwink plots of extracted pectin. Albedo was heated for 2, 3, and 6 min at 120 °C.

total yield was about 16.9% (w/w) of the albedo (see Table 3). The data in Table 4 reveals that with increasing extraction step, Mw increases, but Rgz and [η]w decrease. Such behavior may result from fragmentation and reaggregation of pectin6 during heating and/or that the pectin is becoming increasingly branched as the stepwise extraction progresses. The greater discrepancy between Mw values from the MALLS method as compared to the LS/V method in steps 2 and 3 over values found in step 1 may indicate an increase in fragmentation in the latter two steps. In Figure 6, we have plotted log [η] against log M, the Mark-Houwink (M-H) plot, for pectins extracted between 2 and 6 min at a maximum of 120 °C. As the time of heating increased, the concavity of the curves increased. Concavity indicates that the monomer residue density within the polysaccharide domain (compactness) increases with an increase in molar mass. With increased heating time, the high molar mass fraction becomes even more compact. The M-H exponent “a” in Table 2 is the slope of the best linear leastsquares line drawn through the data. If the plot is linear (e.g., sample 2/120(B)), it indicates that all of the molecules in the distribution are similar in compactness and the M-H exponent is representative of the compactness of all of the molecules in the distribution. When concavity occurs in the M-H plot, then the M-H exponent represents the average compactness of the molecules in the distribution. This means


Flash Extraction of Pectin from Orange Albedo Table 4. Molecular Properties of Sequentially Extracted Pectins MALLS ext. no. 1 2 3 a

Mw ×

10-3

424(7)c 770(9) 838(10)

LS/V

Rgz (nm)

ba

53(1) 44(1) 45(1)

0.12(0.03) ≈0 ≈0

Mw ×

10-3

536(3) 1570(70) 1630(50)

Rgz (nm)

[η]w (dL/g)

ab

63(1) 80(1) 70(1)

8.3(0.1) 6.2(0.1) 4.8(0.1)

0.60(0.01) 0.30(0.01) 0.22(0.01)

Scaling law exponent of log Rg against log M plot. b Mark-Houwink exponent. c Standard deviation of triplicate analysis.

Figure 7. Molar masses determined by LS/V method plotted against molar masses determined by the MALLS method for pectins in Table 2.

the distribution could have a mixture of shapes. Another possibility is to have a mixture of shapes within the same fraction as mentioned above. In the case of pectin, we have shown evidence that during heating there is a transition from branched to more linear extended molecules due to heat induced disaggregation of side chains.5,15 Thus, it is quite possible to have molecules with different shapes or compactness but similar radii of gyration eluting within the same fraction. In that case, there could be large disparities in [η] as well as M within the same size fraction. Because the MALLS method depends on the angular dependence of the light scattering intensity whereas the LS/V method depends on the intrinsic viscosity and the light scattering at 90°, this will lead to disparities in values for M and Rg measured by the two methods on the same sample. In Figure 7, we have plotted Mw from LS/V against Mw values from MALLS for data shown in Table 2. The r2 value for the linear least squares line through the data was 0.94, and the slope of the line was 3.06. This indicates that LS/V values of molar mass are greater than MALLS values by a constant factor of about 3. In that case, the lines in the M-H plots in Figure 6 would be shifted to lower molar masses by about log 0.5, but the slopes would be unchanged; that is, the M-H exponents would be unchanged. To more precisely specify the mixture of molecular shapes present in the samples heated for 3 and 6 min at 120 °C, we have integrated the chromatograms by parts. We chose integration limits at points in the distribution where changes in slopes of the molar mass calibration curves occur. These changes in the slope of the calibration curve are identified initially by changes in a slope or an inflection point on either the 90° light scattering chromatogram or the differential refractive index chromatogram (see Figure 8). Using the elution volumes of these integration limits, the ASTRA

Figure 8. Chromatograms for pectins extracted after 6 min of steam heating. Generated by ASTRA software. Light scattering intensity at 90°, LS, refractive index, DRI. Sample concentration 1 mg/mL; mobile phase, 0.05 M NaNO3; flow rate 0.7 mL/min; injection volume, 200 µL; column temperature, 45 °C. Vertical lines indicate boundaries of fractions obtained by integration by parts.

software was used to divide the calibration curve into linear segments. The same integration limits were used with TriSec software to obtain M-H exponents of fractions given in Table 5. Based on integration by parts, we found that the M-H plots could be approximated by two straight lines (i.e., two M-H exponents) for samples heated for 3 min. On the other hand, samples heated for 6 min required that M-H plots be approximated by three straight lines (i.e., three M-H exponents). Table 5 contains Mw, Rgz, [η]w, and “a” values for these fractions as determined by LS/V and, Mw and Rgz values determined by MALLS. To determine which model, sphere, random coil, or rod, best described the molecules, we calculated Mw and Rgz in each fraction by the LS/V method using all three shape models. We chose the model for each fraction which gave the closest agreement in Mw values between the two methods of analysis. Also, we recalculated new global LS/V values of Mw and Rgz using eqs 4 and 5 based on the Mw and Rgz values found for the fractions16 Mw ) (Rgz 2)1/2 ) {

∑wi Mwi

∑wi Mwi (Rgz2)i / ∑wi Mwi}1/2

(4) (5)

888


Fishman et al.

Table 5. Molecular Properties of Pectin Fractions sample temp (°C)/ time (min) 3/120(A)d

3/120(B)

6/120(A)

6/120(B)

Fa 1 2 Gf Cg 1 2 G C 1 2 3 G C 1 2 3 G C

wi b 0.16 0.84 1 1 0.18 0.82 1 1 0.27 0.49 0.24 1 1 0.28 0.5 0.22 1 1

MALLS Mw × 10-3 Rgz (nm) 1280(20)e 362(2) 491(10) 586 1180(60) 258(70) 450(10) 427 1120(30) 290(5) 128(4) 355(10) 364 913(20) 245(6) 120(3) 393(10) 406

48.1(0.4) 40.5(0.3) 44.3(0.4) 53.5 48.0(0.3) 39.2(0.4) 43.9(0.9) 49.5 36.9(0.5) 32.0(0.7) 28.0(0.6) 36.1(0.1) 34.9 36.9(0.5) 32.5(0.6) 28.8(0.7) 36.0(0.7) 35.1

Mw × 10-3 1680(60) 440(3) 931(20) 640 1340(20) 417(30) 750(70) 586 1620(10) 390(9) 135(3) 571(30) 416 1160(8) 294(4) 126(2) 584(40) 501

LS/V Rgz (nm) 65.9(1.0) 51.8(0.3) 77(2) 58.2 58.6(0.3) 49.5(0.3) 69(1) 53.5 54.7(0.6) 44.2(0.8) 20.0(0.3) 56(3) 44 47.2(2) 38.8(0.2) 19.2(0.2) 54(2) 43.7

[η]w (dL/g)

ac

14.7(0.3) 7.9(0.1) 8.4(0.4)

0.28 0.6 0.46

13.1(0.2) 7.2(0.1) 7.9(0.1)

0.36 0.58 0.5

8.7(0.3) 5.9(0.1) 2.9(0.1) 5.7(0.1)

0.17 0.46 1.46 0.51

7.9(0.2) 5.3(0.1) 2.7(0.1) 5.2(0.1)

0.18 0.54 1.57 0.52

a Fractions from integration by parts. b Weight fraction of fractions from integration by parts. c Mark-Houwink exponent. d “A” indicates the pectin fraction that floated, and “B” indicates the pectin fraction that sank to the bottom when precipitated by 70% isopropyl alcohol. e Standard deviation of triplicate analysis. f Global samples calculated from whole chromatogram. LS/V method assumed random coil. g Corrected global sample calculated from integration by parts. LS/V method assumes Fraction 1 is comprised of hard spheres; Fraction 2, random coils; Fraction 3, rods.

Here wi is the weight fraction of the ith fraction, Mw is the weight average molar mass, (Rgz2)1/2 is the same as Rgz or z average root-mean-square radius of gyration. As a control, also using eqs 4 and 5, we recalculated new global MALLS values of Mw and Rgz based on the Mw and Rgz values found for the fractions. As indicated in Table 5, global corrected (C) values of Mw and Rgz for the two methods are in better agreement than uncorrected global (G) values of Mw and Rgz. The uncorrected MALLS global values (G) were obtained by integrating the entire chromatogram whereas the corrected MALLS values (C) were obtained by integrating fractions and summing the Mw and Rgz values obtained for each fraction in the chromatogram according to eqs 4 and 5, respectively. For both G and C MALLS values, no assumption as to the shape of the molecules was required. In the case of Mw and Rgz values determined by the LS/V method, uncorrected global values (G) were obtained by assuming the random coil model and integrating the entire chromatogram. Corrected global LS/V values were obtained by integrating fractions and summing the Mw and Rgz values obtained for each fraction in the chromatogram according to eqs 4 and 5, respectively. For all samples, models assumed were a hard sphere model for fraction 1, a random coil model for fraction 2, and a rodlike model for fraction 3. We only obtained a fraction 3 for the sample heated for 6 min. As shown by our previous work on pectin molecules imaged from dilute solution by transmission electron microscopy (TEM), a host of shapes and sizes may coexist in a pectin preparation.17,18 Here we have shown that, with decreasing molar mass in a distribution, scaling laws using models predict a transition in shape from spheres to random coils to rods. These models as predictors of shape should not be taken literally. More accurately, these models predict that the organization of monosaccharide residues within the pectic polysaccharide domain becomes less compact with decreasing molar mass. Such would be the case for monomer

Figure 9. Schematic models of possible pectin structures based on electron micrographs.

residues if a polymer molecule transitioned from sphere to random coil to rod. Our earlier TEM work suggests that upon appropriate perturbation infinite pectin networks disaggregate into spherical microgel networks. Upon further perturbation, these disaggregate into their component parts which initially are branched polysaccharides. Moreover, upon further perturbation branched pectins may disaggregate into rods, segmented rods, and kinked rods. A schematic diagram of these various pectin forms based on images obtained by electron microscopy is shown in Figure 9. The M-H exponents which we have observed with decreasing molar mass are consistent with the progressive dissociation observed by microscopy. Previously, we have observed the tendency for pectin to disaggregate by diluting its concentration,19,20 the addition of hydrogen bond disrupting materials,17,18,21 by following time dependent microwave heating during extraction,5 and by measuring its molecular properties after dissolving in various solvents.6 The data in Table 6 allows a comparison of pectin extracted by steam injected heating, “A” or “B”, with pectin extracted by microwave heating (MH), both from orange albedo and with commercial citrus pectin (CCP). Also found in Table 6 is a comparison of results of pectin from albedo


Flash Extraction of Pectin from Orange Albedo Table 6. Comparison of Molecular Properties of Various Pectins MALLS sample temp (°C)/ time (min) Mw × 10-3 Rgz (nm) 3/110(A)b Lc

516(15)d

(A)C 3/120(A)L (B)L 3/120(A)C (B)C 3/120MLe MC CCP

434(18) 513(2) 450(10) 491(10) 450(10) 373(3) 437(10) 254(4)

51.6(0.2) 47.0(0.5) 45.7(0.1) 43.9(0.9) 44.3(0.4) 43.9(0.9) 50.5(0.4) 42.8(0.1) 35(2)

LS/V [η]w (dL/g)

aa

10.5(0.2) 8.4(0.1) 8.8(0.2) 7.9(0.1) 8.4(0.4) 7.9(0.1) 9.4(0.1) 7.3(0.1) 6.1(0.1)

0.57(0.01) 0.55(0.04) 0.52(0.01) 0.50(0.01) 0.46(0.01) 0.50(0.01) 0.64(0.01) 0.54(0.01) 0.64(0.01)

a Mark-Houwink exponent. b “A” indicates the pectin fraction that floated, and “B” indicates the pectin fraction that sank to the bottom when alcohol was added. Samples with “A” or “B” were extracted with steam heat. c “L” lab prepared albedo, “C” commercially prepared albedo. d Standard deviation of triplicate analysis. e “ML” microwaved from lab albedo, “MC” microwaved from commercial albedo, “CCP” commercial citrus pectin.

prepared in the laboratory (L) with that of albedo prepared commercially (C). Interestingly, larger values of Mw, Rgz, and [η]w were obtained from both heat injected and microwaved pectin heated for 3 min than from CCP. Furthermore, the polydispersity values, Mw/Mn and Mz/Mn (data not shown), were lower for the steam or microwave extracted pectin than for CCP indicating a more narrow molar mass distribution. Even for samples 6(A) and 6(B) located in Table 6, which have higher Mw values, comparable Rgz values and slightly lower [η]w values have comparable or slightly poorer molecular properties than CCP. Moreover, as indicated in Table 1, values of DE and %GA are comparable to values of DE and %GA found for commercial high methoxyl pectin.19 Comparison of data in Table 6 revealed that pectin from lab prepared albedo had slightly better properties than commercially prepared albedo. Conclusion As judged by their properties, it appears that pectin rapidly extracted from orange albedo by steam injection under pressure was comparable to the same pectin produced by microwave heating under pressure. Furthermore, both of these were comparable or better than commercial citrus pectin prepared either from lime and/or lemon peels. Yield of sequentially extracted orange pectin was about 16.9% of dry matter which was lower than the 25-30% of dry matter found for commercial citrus pectin.4 Nevertheless factors which tend to mitigate the lower recovery of pectin from the heating process described here compared to commercial heating processes are savings in energy due to shorter heating times, premium in selling price due higher quality and the lower cost and greater availability of orange peels over lemon or lime peels. Disparities between molar masses obtained

from the MALLS method as compared to the LS/V method could be reduced significantly by correcting for curvature in the M-H plot. Plots were corrected by modeling linear segments of the lines as spheres, random coils, or rods in that order as the molar mass decreased. Our data here is consistent with earlier results which indicated that pectin distributions were mixtures of two or more of the following: spherical aggregates, hydrogen bonded network structures, and partially or fully disaggregated components of network structures which could include, branched structures, rods, segmented rods, and kinked rods. Acknowledgment. We thank Andre White and Tung Nguyen for their technical assistance in obtaining monosaccharide composition data and Kevin Hicks for helpful comments. We also thank John Phillips for help with statistical analysis of the data. References and Notes (1) Fishman, M. L. In Wiley Encyclopedia of Food Science and Technology, 2nd ed.; Francis, F. J., Ed.; John Wiley & Sons: New York, 2000; Vol. 3, pp 1858-1862. (2) Yamada, Y. In Pectins and Pectinases; Visser, J., Voragen, A. G. J., Eds.; Elsevier: Amsterdam, 1996; pp 173-190. (3) Kertez, Z. I. The Pectic Substances; Interscience Publishers: New York, 1951; p 457. (4) Braddock, R. J. Handbook of Citrus By-Products and Processing Technology; John Wiley & Sons: New York, 1999; pp 191-197. (5) Fishman, M. L.; Chau, H. K.; Hoagland, P.; Ayyad, K. Carbohydr. Res. 2000, 323, 126-138. (6) Fishman, M. L.; Chau, H. K.; Kolpak, F.; Brady, J. J. Agric. Food Chem. 2001, 49, 4494-4501. (7) Fishman, M. L.; Chau, H. K. United States Patent 6,143,337, 10/7/ 2000. (8) Filisetti-Cozzi, T. M. C. C.; Carpita, N. C. Anal. Biochem. 1991, 197, 157-162. (9) Voragen, A. G. J.; Schols, H. A.; Pilnik, W. Food Hydrocoll. 1986, 1, 65-70. (10) Wyatt, P. J. Anal. Chim. Acta 1993, 272, 1-40. (11) Viscotek, TriSec GPC Software Reference Manual, V3, pp A-18. (12) Kratochvil, P. In Light Scattering of Polymer Solutions; Huglin, M. B., Ed.; Academic Press: New York, 1967; pp 335-384. (13) Ptitsyn, O. B.; Eizner, Y. E. SoV Phys. Technol. Phys. 1960, 4, 10201036. (14) Fox, T. G.; Flory, P. J. J. Am. Chem. Soc. 1951, 73, 1904-1908. (15) Fishman, M. L.; Chau, H. K.; Coffin, D. R.; Hotchkiss, A. T., Jr. In AdVances in Pectin and Pectinase Research; Voragen, F., Schols, H., Visser, R., Eds.; Kluwer Academic Publishers: Boston, MA, 2003; pp 107-122. (16) Fishman, M. L.; Doner, L. W.; Chau, H. K.; Hoagland, P. D. Int. J. Polym. Anal. Charact. 2000, 5, 359-379. (17) Fishman, M. L.; Cooke, P.; Levaj, B.; Gillespie, D. T.; Sondey, S. M.; Scorza, R. Arch. Biochem. Biophys. 1992, 294, 253-260. (18) Fishman, M. L.; Cooke, P.; Hotchkiss, A.; Damert, W. Carbohydr. Res. 1993, 248, 303-316. (19) Fishman, M. L.; Pfeffer, P. E.; Barford, R. A.; Doner, L. W. J. Agric. Food Chem. 1984, 32, 372-378. (20) Fishman, M. L.; Gillespie, D. T.; Sondey, S. M.; Barford, R. A. J. Agric. Food Chem. 1989, 37, 584-591. (21) Fishman, M. L.; Gross, K. C.; Gillespie, D. T.; Sondey, S. M. Arch. Biochem. Biophys. 1989, 179-191.

BM020122E

Flash Extraction of Pectin from Orange Albedo by Steam Injection

Recommend Documents