Dielectric Properties of Potato Puree in Microwave Frequency Range as Influenced by Concentration and Temperature

Abstract

Dielectric properties of potato flour-water dispersions (slurry) were measured in the frequency range of 500–2500 MHz by the open-ended coaxial probe method using a network analyzer as a function of concentration (10–25% w/w) and temperature (20–75^oC). Both commercial and laboratory prepared samples of potato flours were used. Results indicated that the dielectric constant (ε′) decreased with temperature and frequency while increased with concentration. The loss factor (ε″) increased with frequency and concentration; however, temperature showed mixed effect. Both ε′ and ε″ data in conventionally used microwave frequencies (915 and 2450 MHz) were studied as function of concentration and temperature for two sample types using a response surface methodology and found to follow 2^nd order polynomial models. Temperature and concentration contributed significantly on dielectric spectra of potato slurry and the sample source had some effect. A change in ε′ and ε″ above 70^oC could be attributed by starch gelatinization. Penetration depth (D_p) decreased with an increase in frequency and non-systematic with temperature. Addition of salt substantially reduced D_p of potato slurry.

Keywords:

• Potato
• Dielectric constant
• Loss factor
• Starch gelatinization
• Concentration
• Penetration depth

INTRODUCTION

The dielectric properties of food materials indicate how much microwave energy is absorbed, transmitted, reflected and concentrated during interaction between the food and electromagnetic energy. Dielectric properties play a significant role for effective heating and optimum process design. Understanding dielectric properties will be useful in quality sensing by radio frequency (RF) or microwave instruments.[1]

Biological materials like food have been considered as non-ideal capacitors and have the capacity store and dissipate electrical energy from an electromagnetic field, and these properties can be translated in terms of a complex notation.[2] The complex notation known as dielectric permittivity consists of two parts: a real component dielectric constant (ε′), and an imaginary component termed as dielectric loss factor (ε″).[3] The dielectric constant indicates a material's ability to store electric energy and the loss factor is a measure of its ability to dissipate the electrical energy in the form of heat.

Dielectric properties of food depend on several factors including frequency, moisture content, temperature, concentration, nature and constituents of food materials. Data on dielectric properties of foods have been abundant in the literature.[1,4–12] However, food materials are complex in their composition and in their dielectric behavior and therefore it is often necessary to measure the properties under the particular conditions of interest to obtain reliable data.[1]

Potato (Solanum tuberosum L.) production has increased significantly in many developing countries. The potato tuber contains 13–37% dry matter and starch is the major component of the dry matter. Potato starch usually consists of 20–23% amylose and 77–80% amylopectin.[13] Most of the important properties of starch arise from the fact that it is composed of two distinct polymer fractions, amylose [(1–4)-linked α-d-glucan] and amylopectin [(1–4), (1–6)-linked α-d-glucan]. Potato tuber is commonly converted to potato flours/starches to enhance shelf-life and convenience. Potato flour is one of the most viable value added product due to its versatility in function as a thickener and color or flavor enhancer. Instant potato purees are made from potatoes that have been through an industrial process of cooking, mashing, and drying to yield a packaged convenience food that can be reconstituted in the home in seconds by adding hot water or milk, producing a close approximation of mashed potatoes with very little expenditure of time and effort. Potato flour has been used to formulate soups, sauces, baking and extruded products, and also incorporated into batter dough to manufacture gluten free bakery products. Currently, food processors are looking for some innovative technologies where food products can be manufactured with minimum sensory and quality losses as well the process is faster.

Already, microwave and radio frequency heating has been used for baking and finishing of some confectionary including potato products. Various commercial potato flour/starch based soups are on the market and microwave heating can be used to reheat the rehydrated soups. To move into that direction, detailed data on dielectric properties of potato flour individually or in a composite basis are needed. Dielectric properties of mashed potato have been studied as function of temperature, moisture and salt content.[7,14] The process of manufacturing and compositions play significant role, and, therefore, it will be useful to study dielectric properties of potato puree samples from various sources. The objective of the present work was to compare dielectric properties of dehydrated commercial potato flakes intended for ready to use soup and potato flakes prepared in the laboratory as function of temperature, concentration, frequency and salt.

MATERIALS AND METHODS

Materials

Potato puree, a commercial potato product in dry form was procured from Knorr, Turkey (commercial sample, CL) and the second set of potato flour sample was prepared in the laboratory (McGill University) (Laboratory sample, LS). For the laboratory sample, fresh potatoes (Blanches, Quebec, Canada) were procured from a local Montreal market. Potatoes were boiled, peeled, and then mashed manually. Mashed potatoes were dried in a dryer until the moisture content reached about 8%, (wet basis) and finally powdered by a grinder. Dried potato flour was packed in airtight container and stored at room temperature (21^oC) until further use. The proximate composition of both potato samples were reported in Table 1

Table 1 Composition of potato flakes

	Commercial puree	Laboratory prepared puree
Moisture (%)	6	7.8
Ash (%)	3.5	2.7
Protein (%)	8	10.5
Crude fat (%)	1	1.1
Carbohydrate (%)	72	73.1
Others ingredients	Mono-and diglyceride, disodium diphosphate, sodium metabisulphite, spice extract and ascorbil palmitate	–

Experimental Design

A response surface methodology was used to investigate the main effect of the process variables on the dielectric properties of potato slurry. The experimental design adopted was a modified configuration of Box's central composite design (CCD) for three variables. Independent variables were type of sample type (two discrete levels), concentration and temperature. These variables were selected to optimize responses dielectric constant and loss factor in the microwave range (Table 2). The complete design included 40 experiments. Effect of salt addition on dielectric properties of potato slurry was studied at constant concentration (15% w/w) between commercial and laboratory prepared samples at selected temperature (Table 2) and total 12 experiments were carried out for the experiments.

Table 2 Variables and levels for a rotatable central composite design used in dielectric properties of potato slurry experiments

Variables	Symbol	Coded variable levels
		−1.00	−0.556	−0.333	−0.111	0.333	0.778	+1.000
Potato slurry type	X1	CS∗						LS∗
Concentration (%)	X2	10		15		20		25
Temperature (^oC)	X3	30	40		50	60	70	75
∗CS-commercial sample; LS-laboratory sample.

Sample Preparation

As per design described above, selected concentrations of commercial (CS) and laboratory (LS) made potato slurry were made by using required volume of distilled water. A constant salt concentration (0.5% w/w) was directly added to 15% slurry (both) and stirred well before dielectric measurement. The salt was added to compare commercial sample with similar salt concentration and secondly to enhance dielectric properties of potato slurry.

Dielectric Properties Measurement

A vector network analyzer (Model: Agilent 8722ES, Agilent Technology, Palo Alto, CA) with an open-ended coaxial cable (#8120-6192, Hewlett Packard) connected to a probe (85070C, Agilent Technology, Palo Alto, CA) in electronic calibration module (E-cal, 4693A) was used to measure dielectric properties of potato slurry. The slurry sample was placed in a wide glass tube (50 mL) and the open co-axial probe (probe diameter ≈2.4 mm) was set into the tube. The sample was heated from 30 to 75^oC and dielectric properties were measured in the frequency range of 500–3000 MHz. The detailed procedure of the dielectric measurement was reported elsewhere.[10] However, for modeling purpose only mostly used frequency for microwave heating 915 and 2450 MHz were considered. All the measurements were carried out in triplicate and were reproducible to ±5%. Mean values and standard errors were calculated from 3 replicates.

The penetration depth is an important parameter in characterizing temperature distribution in microwave heated foods. Penetration depth (D_p) is generally described as the distance from the surface of a dielectric material where the incident power decreased to 1/e of the incident power. The magnitude of D_p was calculated using the dielectric constant and loss factor at frequencies of 915 and 2450 MHz for potato slurries[15] as:

(1)

where λ_o = 0.328 m at 915 MHz and 0.122 m at 2450 MHz.

Polynomial Equations and Statistical Analysis

A quadratic polynomial regression model was considered for predicting individual Y variables. The proposed model for each response is:

(2)

where b₀ is a constant and b_i, b_ii, and b_j are regression coefficients of the model and X_i and X_j are the independent variables in coded values. The regression analysis was performed using Matlab™ custom made routines. The significance of the coefficient of the model was calculated by means of RMSEC (root mean square error in calibration) and RMSCEV (root mean square error in cross validation) was used for a first testing of the predictive model. The Leave-One-Out Cross validation is performed by computing as many models as samples, each time removing one sample from the data set and predicting its response using the coefficients so computed. It was also compared to the pooled standard deviation calculated from the experimental variance of the sets of replicate experiments. The contour plots for the model were plotted as a function of two variables keeping the third variable at a constant value.

RESULTS AND DISCUSSION

Effect of Frequency on Dielectric Properties of Potato Purees

Typical dielectric spectra for samples of commercial (CS) and laboratory (LS) formulated potato slurries (10% w/w) at 30ºC are shown in Fig. 1. Both ε′ and ε″ values for LS and CS samples decreased similarly from 500 to 1650 MHz and, thereafter increased at higher frequency range. The LS sample consistently showed higher values of both ε′ and ε″ as compared to the CS sample, especially at the lower frequency. The relatively higher dielectric values for LS was attributed by higher moisture content. Change in loss factor is attributed by ionic conductivity at lower frequencies, to bound water relaxation, and to free water relaxation near the top of the frequency range.[16] Dielectric behavior noticed in this study was very similar to our previous study on Basmati rice slurry.[10] Decrease in dielectric parameters as function of frequency is common for various foods.[9,17–18]

Figure 1 Typical dielectric spectra of commercial (CS) and laboratory prepared (10%) potato slurry at 30^oC.

Effects of Concentration, Temperature, Sample Type, on Dielectric Constant

Statistical analysis revealed that the model fits were significant, and there was satisfactory correlation between actual and fitted values (R² ≥ 0.86 and adjusted R² ≥ 0.82). The 2^nd order polynomial equations for dielectric constant at 915 [Eq. (3) and 2450 [Eq. (4)] (computed on the coded values) take the following form (only significant coefficients are reported):

(3)

(4)

Concentration (X₂) and temperature (X₃) had a negative influence on the dielectric constant of the potato slurry and interaction terms (between sample type and concentration) enhanced the magnitude of ε′. Since the sample type is a discrete variable, the equation needs to be separated for the two sample types in order to characterize the influence of temperature and concentration for the two sample types. This can be done by substituting X₁ = −1 for the commercial sample and X₁ = +1 for the lab formulated sample. So for the commercial sample (CS) at 915 MHz, Eq. (3) becomes:

(3a)

and for the lab sample (LS) the equation will be

At 2450 MHz, the equation 4 for the commercial sample (CS) will be

(4a)

and for the lab sample, it is

(4b)

The dielectric constant as a function of temperature vs. concentration for each potato source is plotted individually at 915 and 2450 MHz respectively (Figs. 2 and 3). They demonstrate that an increase in temperature and concentration markedly reduced ε′ value for both sources and both frequency levels. A decrease in dielectric constant as function of temperature and concentration for various foods are reported in the literature.[7,10,19] The small influence of potato source on ε′ becomes apparent when the subplots a and b are compared in Figs. 2 and 3 respectively. Only the concentration coefficients for the two samples showed a large difference, with the commercial sample showing a greater effect.

Figure 2 Surface plots for dielectric constant of potato slurry (unsalted) as a function of temperature and concentration at 915 MHz: a) sample 1; and b) sample 2.

Figure 3 Surface plots for dielectric constant of potato slurry (unsalted) as a function of temperature and concentration at 2450 MHz: a) sample 1; and b) sample 2.

Dielectric loss factor of potato slurry fitted well 2^nd order polynomial equations as function of sample, type, concentration and temperature. The equations for loss factor at 915 [Eq. (5)] and 2450 [Eq (6)] MHz take the following form:

(5)

(6)

Again, these were simplified for CS (5a) and LS (5b) at 915 MHz as

(5a)

(5b)

and CS (6a) and LS (6b) at 2450 MHz

(6a)

(6b)

At 915 MHz, all both variables have positively influenced the loss factor, however, the negative square term for temperature effect somewhat shadowed the positive linear effect. However, since the coefficients (2.39 and 1.83) were much higher than the variable values (-1 to +1, coded), the positive linear term dominated the temperature effect, and hence the overall effect was still positive (loss factor decreased as temperature and concentration increased). At 2450 MHz, however, only the linear terms were significant, and the negative effect of temperature and the positive effect of concentration on the loss factor were distinct. The coefficients for the temperature and concentration for the two samples were very different indicating differences in the magnitude of loss factors between the two samples.

The effects of concentration and temperature on loss factor at 915 and 2450 MHz are illustrated in Figs. 4 and 5, respectively. Higher ε″ values were observed for LS slurry at similar concentration and temperature studied over commercial sample and this difference became more pronounced at higher concentration. At 915 MHz, ε″ increased as temperature was increased from 30 to 75^oC for both slurries. The trend was similar at higher frequency (2450 MHz); however, the magnitudes were relatively lower than at 915 MHz. For CS, ε″decreased as temperature was increased and a non-symmetrical increase of ε″ was observed for LS sample. Regier et al.[14] observed similar decreasing value of ε″ for mashed potatoes with temperature at 2450 MHz.

Figure 4 Surface plots for loss factor of potato slurry (unsalted) as a function of temperature and concentration at 915 MHz: a) sample 1; and b) sample 2.

Figure 5 Surface plots for loss factor of potato slurry (unsalted) as a function of temperature and concentration at 2450 MHz: a) sample 1; and b) sample 2.

The loss factor increased as concentration of slurry was increased from 10 to 25% and found to be more significant at higher temperature. The increased value of the loss factor at higher concentration (low water content) could be due to the increased viscosity of the slurry. An increase in loss factor as function of concentration has been reported recently by Motwani et al.[19] for corn starches. This trend was found to be opposite to our earlier study on rice flour slurry[10] where ε″ decreased as concentration increased from 30 to 50%. The variation could be affected by type of starch and ratio of amylose to amylopectin in the studied starch.

A combined effect of temperature and concentration on ε″ is presented by contour plot (Figs. 4b, 5b). An increase in ε″ value as function of temperature and concentration was observed at 915 MHz and ε″ values at 2450 MHz were insignificantly increase or almost constant. Guan et al.[7] observed similar increase in loss factor while temperature was increased from 20 to 120^oC. A relatively greater ε″ value of LS sample which influenced by temperature (>70^oC) as compared to CS could suggest higher degree of gelatinization of potato starch that took place in the studied temperature range where as dielectric values of CS indicated samples either pre-gelatinized or completely gelatinized (no significant change in ε″) before manufacturing.

Combined Effect of Temperature and Type Sample on Dielectric Constant of Salt Added Slurry

Addition of salt (0.5% w/w) to potato slurry did not change dielectric constant significantly. Combined effect of temperature and type of sample at constant concentration (15%) and salt level (0.5% w/w) on dielectric constant were evaluated by 2^nd order polynomial model [Eq. (7 at 915 and Eq (8) at 2450 MHz]; however, only the temperature term showed significance:

(7)

(8)

Hence, only temperature influenced the dielectric constant of salted potato slurry. The type of sample and their concentration had no effect on ε′ value at either of the two frequencies studied. In our earlier study on Basmati rice flour slurry,[10], similar trend was observed for dielectric constant while 1% salt was added. The results indicated that the dielectric constant of salted slurry decreased as temperature increased from 30 to 75^oC for both 915 and 2450 MHz. Guan et al.[7] observed similar trend for dielectric constant of salted mashed potato. Addition of salt to water influences the dielectric properties in two ways.[2] The dissolved ions obtained from salt bind the water molecules and the degree of binding is a function of nuclear charge which in turn depends on the size and the charge of dissolved ions. This results in water polarization and consequently decreases in ε′.

Added salt significantly influenced loss factor of potato slurry. The increase in ε″ with salt addition is due to electrophoretic migration of dissolved salts that also relate to size and charge of the dissolved ions. Similar observation has been made by Bircan and [19] during investigation of salt–starch interaction on dielectric properties. Changes of loss factor as function of temperature and potato type were well represented by 2^nd order polynomial equations at 915 and 2450 MHz [Eqs. (9 and 10)]. A high coefficient of determination (R² = 0.97) indicated that the developed model for the loss factor fitted adequately.

(9)

(10)

Equation (9)(loss factor at 915 MHz) was significant only with the linear term of temperature. For the two samples, the relationship will only be different with respect to the intercept (CS = 35.8, LS = 55.8) and hence the linear curve for the two samples will be parallel to each other with the loss factor directly increasing with temperature.

However, with the loss factor at 2450 MHz, the temperature relationship was a bit more complex. Concentration term was not significant. The equation represents loss factor as a quadratic function of temperature for both samples (10a for CS and 10b for LS):

(10a)

(10b)

Again, higher A higher loss factor value was observed for LS than CS. The temperature effect was rather weak in view of the relatively large size of the intercept. For the commercial sample, there was a small decrease in loss factor as the temperature increased while there was a small increase in loss factor with temperature with LS.

Penetration Depth

Penetration depth (Dp) of potato slurry provides information related to depth to which heat is dissipated uniformly. An increase in penetration depth will enhance the temperature uniformity in microwave heating. Temperature dependence of penetration depth is complex, because ionic and dielectric losses interact and, therefore, no generalized trend was observed between D_p and temperature of 15% potato slurry (Table 3). A change of D_p above 70^oC could be attributed by gelatinization of potato starch. However, penetration depth of slurry at 915 MHz was higher compared to sample measured at 2450 MHz. This is in close agreement with earlier observation of Wang, Wig, Tang, and Hallberg.[17] Addition of salt significantly reduced penetration depth and similar observation was observed in our earlier study on rice slurry. The potato starch gelatinization process appeared to play a significant role on dielectric properties of rice flour slurry.

Table 3 Effect of salt addition on penetration depth of 15% potato slurry as function of temperature

	CS		CSS		LS		LSS
Temp. (^oC)	915 MHz	2450 MHz	915 MHz	2450 MHz	915 MHz	2450 MHz	915 MHz	2450 MHz
30	3.36	1.18	1.61	0.82	2.19	1.02	0.99	0.63
40	3.51	1.28	1.51	0.85	2.07	1.04	0.92	0.62
50	3.07	1.31	1.38	0.84	1.81	1.02	0.83	0.60
60	2.91	1.38	1.22	0.80	1.50	0.94	0.76	0.53
70	2.38	1.33	1.11	0.76	1.34	0.89	0.70	0.53
75	2.28	1.30	1.51	0.85	1.34	0.89	0.79	0.55

CONCLUSIONS

Dielectric properties of potato slurries were studied as function of concentration, temperature and variation of in the frequency range, 500–2500 MHz with or without added salt. Results indicated that dielectric constants were increased as function of concentration. Temperature was found to contribute significantly on dielectric parameters; type of sample had the least effect on dielectric behaviour. Addition of 0.5% salt to slurry significantly increased loss factor and reduced penetration depth. Dependence of dielectric properties on concentration, sample type, and temperature were well represented by 2nd order polynomial equations under selected conditions. The potato starch gelatinization process appeared to play a significant role on dielectric properties of potato flour slurry.