Thin‐layer drying behavior of West Indian lemongrass (Cymbopogan citratus) leaves

Abstract The objectives of this study were to investigate the effect of temperature (40, 50, 60, and 70°C), and air velocity (0.5, 1, and 2 m/s) on the drying behavior of West Indian lemongrass (Cymbopogan citratus) leaves. Drying was carried out in a computer‐controlled tray dryer. Overall, the effect of temperature was seen to be more important than that of air velocity, but the air velocity did have an effect on drying rates at the start of the drying process at 50–70°C. Drying rate constants, diffusivity values, and activation energy were determined. Twenty‐two empirical and semiempirical thin‐layer models were tested, and although model fit varied, the Midilli model could be applied to all data with reasonable prediction of MR values.

. However, with respect to the drying method employed, there are limited studies on lemongrass which include conventional cabinet drying (Kassem et al., 2006;Lonkar et al., 2013). Previous works have instead focused on the drying of leaves in a solar dryer well as other dryers such as fluidized bed dryer, heat pump dryer, constant temperature and humidity chamber, biomass dryer, and a fixed bed dryer (Kassem et al., 2006;Waewsak et al., 2006;Kemat et al., 2008;Ibrahim et al., 2009;Fudholi et al., 2012;Sanmeema, Poomsa-ad, & Wiseet, 2012;Rahman et al., 2013;Coradi et al., 2014). The effect of drying parameters (such as temperature and velocity) on the behavior of leaves has not been reported for leaves dried in conventional ovens. Work has been reported for leaves dried in a fluidized bed dryer at a range of temperatures but at fixed or small range in air velocity (Kemat et al., 2008;Rahman et al., 2013) or at a range of relative humidity values (Ibrahim et al., 2009). With respect to drying rates, no information has been presented on the effect of drying parameters on moisture curves and drying rate curves for leaves dried in conventional ovens. Drying curves were presented by some authors (Ibrahim et al., 2009;Fudholi et al., 2012;Coradi et al., 2014). Rahman et al. (2013) calculated drying rate of leaves dried in a fluidized bed dryer in units of g/g min. Lastly, with respect to mathematical modeling of drying data, modeling has been performed for methods other than conventional drying, with limited drying variables and using selected thin-layer models, which vary in their method of model assessment (Coradi et al., 2014;Ibrahim et al., 2009;Kemat et al., 2008;Waewsak et al., 2006).
Due to the gaps in the information available, this work was therefore undertaken to systematically describe the effect of temperature and air velocity on the drying behavior of leaves dried in a conventional, drying oven. This study was conducted to collect baseline drying data for West Indian lemongrass (Cymbopogan citratus) leaves dried at three (3) drying air temperatures and velocities, and using two pretreatments. The specific aims were to develop drying curves, determine drying rate constants and diffusion coefficients, and model the drying data using twentytwo thin-layer models. Due to experimental limitations in a typical cabinet oven, the relative humidity (rh) of the drying air was not controlled, noting as well that humidity is not usually manipulated during a typical drying process.

| Rawmaterial
Mature West Indian lemongrass (Cymbopogan citratus) leaves were sanitized using a 1% bleach (sodium hypochlorite) solution, pat with paper towels to remove excess water (Kadam, Goyal, Singh, & Gupta, 2011), then allowed to air-dry for approximately 30 min. To fit on the drying trays, the leaves were then cut into pieces 20 cm in length. A total of 4.2 kg of lemongrass was used in the drying experiments.

| Dryingprocedureandtreatments
To evaluate the effect of drying air temperature and air velocity, drying was carried out in an Armfield UOP8MK-II tray dryer (Armfield, F I G U R E 1 Schematic of Armfield UOP8-KK11 tray drier  Table 1. Preliminary drying runs were conducted to investigate the effect of drying pretreatments used by Lonkar et al. (2013), namely water and chemical blanching (1% sodium carbonate solution, w/v) followed by drying at 50°C (1 m/s). It was found that these pretreatments did not significantly affect the equilibrium moisture values of dried leaves nor the time taken (min) to achieve equilibrium moisture content when compared with untreated leaves. Leaves subjected to pretreatments prior to drying experienced negative changes in color, turning yellow brown on drying.

| Analyticalmethods
To determine the moisture content (wb) of fresh and dried leaves, samples were dried at 103°C using a Mettler Toledo HB43-S (Mettler-Toledo, Columbus, Ohio, USA) moisture analyzer and moisture content expressed on a dry weight basis (g H 2 O/g DM).
Water activity of fresh and dried samples was determined using the Aqualab CX-2 water activity meter (Aqualab, Pullman, Washington, USA Based on observations during preliminary drying runs, four-point rating scales were developed to assess the color, texture, and odor of the leaves before and after drying. Color was assessed as 4) bright green 3) dull green 2) light green 1) yellow/brown. Lemongrass odor was rated as 4) strong 3) moderate 2) slight 1) none. Leaf texture after drying was assessed as 4) pliable 3) slightly pliable 2) slightly brittle 1) brittle/breaks easily.

| Dataanalysis
The sample weight data (g) at the end of the drying process and the moisture content of the final dried sample were used to backcalculate the moisture content of the respective samples at each point during the drying process (Mujaffar & Sankat, 2005, 2015. The drying rate constant (k) was determined from a plot of ln MR (1) Hue = Arc tan b * a * (degrees) TA B L E 2 Thin-layer drying models

Henderson and Pabis MR = a exp (−kt)
Modified Henderson and Pabis For this study, a total of twenty-two (22) empirical and semiempirical thin-layer models (Alibas, 2014;Kucuk, Midilli, Kilic, & Dincer, 2014;Silva, Silva, Gama, & Gomes, 2014) were applied to the MR data (Table 2). Some models are derived from the original older models. As is now a common practice in thin-layer drying studies, the performance (fit) of the models was assessed through the use of the coefficient of determination (R 2 ), root mean square error (RMSE), and the chi-squared statistic (χ 2 ). Further regression analysis and ANOVA were carried out using GenStat for Windows Discovery Edition 4 Software (VSN International Ltd., 2014).
Model fit was carried out using Curve Expert Professional software, version 2.3.0 (Hyams, 2016).

| Qualityattributes
Fresh leaves were bright green in color (rating of 4) with a strong lemongrass scent (rating of 4) and very pliable in texture (rating of 4). Figure 2 shows that with respect to the subjective ratings for dried leaves, all ratings (color, odor, and texture) were lower than those for the fresh leaves. Leaves dried at 40 and 50°C were similar and showed the least deterioration in quality attributes, while those dried at 70°C showed the greatest deterioration. At the end of drying, leaves dried at 40 and 50°C were dull green in color, while leaves dried at 60°C were lighter. Leaves dried at 70°C Subjective quality assessment of dried lemongrass leaves. Color 4) bright green 3) dull green 2) light green 1) yellow/brown. Lemongrass odor: 4) strong 3) moderate 2) slight 1) none. Leaf texture: 4) pliable 3) slightly pliable 2) slightly brittle 1) brittle/breaks easily TA B L E 3 Color attributes of fresh and dried lemongrass leaves  The color attributes of fresh and dried leaves are given in Table 3.
The loss of quality in oven-dried leaves, especially at higher temperatures, can be attributed to browning reactions at higher temperatures, a decrease in chlorophyll content and essential oils (Coradi et al., 2014;Chen & Patel, 2008;Kassem et al., 2006;Sanmeema et al., 2012). With respect to the effect of air velocity on quality of dried leaves, Coradi et al. (2014) found that increasing the air velocity from 0.8 to 1.3 m/s did not negatively affect the essential oil content of lemongrass leaves when compared with fresh leaves; however, further increasing the velocity to 1.8 m/s resulted in a statistically significant reduction in oil content. They noted the works of Buggle et al. (1999) who dried lemongrass leaves at 30-90°C and found that the highest oil content of 1.43% was found in leaves dried at 50°C. Drying at 30°C favored fungal growth and drying at 70°C resulted in significant reduction in essential oil content to 0.34%. Lightening of leaves after drying has also been reported for coriander and fenugreek leaves (Naidu et al., 2012;Shaw, Meda, Tabil, & Opoku, 2007). High temperature could lead to the replacement of magnesium ions in chlorophyll by hydrogen therefore converting chlorophyll to pheophytins (Naidu et al., 2012). Some studies showed that drying temperature had no significant effect on the color of dried mint and bay leaves (Cakmak et al., 2013;Demir, Gunhan, Yagcioglu, & Degirmencioglu, 2004;Kadam et al., 2011).

| Dryingcurves
The initial moisture content and water activity values of lemon-  During drying, moisture content of leaves was found to decrease in the typical manner, with an initial rapid decline followed by a gradual decrease toward equilibrium.
Generally, for all temperatures and air velocity treatments, the greatest decrease in moisture content occurred during the first 200 min of drying. As temperature was increased from 40 to 70°C at 0.5 and 1.0 m/s, leaves experienced a greater decline in moisture content. This trend was also seen for leaves dried at 2.0 m/s as drying temperature increased from 40 to 60°C, but further increasing the temperature to 70°C did not cause any further change in leaf moisture content ( Figure 3c).
As shown in Figure  leaves at 50°C to a final moisture content of 11% (wb). As also given in Table 3 for the present study, the time taken for the leaves to achieve an 11% moisture value was significantly affected by temperature and air velocity (p ≤ .001) and a temperature-velocity interaction (p ≤ .05). Leaves dried at 40°C at 0.5 and at 1.0 m/s could not be dried to 11% moisture content as equilibrium moisture values averaged 20 and 18% (wb), respectively. Aside from that, the higher the temperature, the shorter the time taken to achieve a 11% mois- with this present study could possibly be due to the small sample piece size (2-cm length) used in that study. Ibrahim et al. (2009) reported that the time required for lemongrass leaves to reach a moisture content of 0.2% dry basis was 550 min at 35°C compared to 200 min at 55°C. Sanmeema et al. (2012) reported that the moisture content of lemongrass leaves (species not given) could be reduced from 180-190% (db) to 10% (db) in a heat pump dryer using hot air at 40-60°C.
Higher temperatures cause a higher reduction in moisture content as a result of increased heat and mass transfer, which favors evaporation of moisture from the leaves (Aghbashlo, Kianmehr, & Hassan-Beygi, 2010;Doymaz, 2006 Values are means ± SEM, n = 2 per treatment group. a-f Means in a column without a common superscript letter differ (p < .05) as analyzed by two-way ANOVA and the LSD test.

| Dryingratecurves
Drying rate as a function of average moisture content at the different air velocities is shown in Figure 5a- (Sablani & Rahman, 2007). This study revealed that there was a velocity effect at the start of the drying process at 50 and 60°C.
Air movement is important particularly during the early stages of drying when the external mass transfer mechanism predominates, and during this time, air velocity will impact on the rate of external mass transfer. Erbay and Icier (2010) noted that the influence of drying air temperature was higher than that of air velocity in the drying of olive leaves, adding that the highest temperature and velocity combination did not give the highest drying rate in F I G U R E 5 Drying rate as a function of average moisture content of lemongrass leaves dried at different air temperatures olive leaves dried at 50-70°C and 0.5 to 1.5 m/s. As seen in this study, higher heat transfer rates at 70°C and 2 m/s resulted in deleterious changes to the texture of the leaves which decreased the rate of moisture removal.

| Dryingrateconstant(k)andeffectivediffusion coefficient(D eff )
The drying rate constants (k) were determined from the initial straight The drying constant is said to be dependent on the material properties and the characteristics of the drying air as it represents several transport phenomena (Mujumdar, 2007). It is generally expected that as the drying rate increases, the drying rate constant will also increase, as happens with an increase in temperature. Diffusion of moisture controls the rate of drying in the falling rate period. An increase in the effective diffusivity is an indicator of lower resistance to mass transfer in the material dried. The diffusivity of water or water vapor of a material during drying is dependent on its structure or porosity and temperature (Naidu et al., 2012). Values are means ± SEM, n = 2 per treatment group. a-e Means in a column without a common superscript letter differ (p < .05).
TA B L E 5 Drying rate constants (k) and diffusion coefficients (D eff ) for dried lemongrass leaves

| Moistureratioandthin-layermodels
Moisture ratio (MR) calculated based on Equation 4 plots are given in Figure 6a-c. Moisture ratio values were significantly affected by drying time, temperature, and a time-temperature interaction (p ≤ .001). As seen previously for the drying curves, increase in temperature resulted in increased decline in MR values, with the exception of leaves dried 2 m/s at 70°C.
With respect to lemongrass, Ibrahim et al. (2009) presented MR curves for leaves dried at 35-55°C and 30-50% rh, reporting that the main factor influencing drying to be temperature. Coradi et al. (2014) presented MR plots for lemongrass leaves dried at 40-70°C showing a temperature effect beyond 60 min of drying. Moisture ratio curves given by Kadam et al. (2011) for mint leaves revealed a trend of increasing decline in MR as temperature increased.
Of the twenty-two thin-layer models applied to the MR data, the coefficients for the five models which best fit the drying data obtained at 40-70°C and 0.5 to 2.0 m/s are given in Table 6 a to c.
Model fit couple be expected to differ because the shapes of the MR curves for leaves dried under different conditions of temperature and velocity also differ. Beyond the models that best fit the data, the other models showed regression coefficients of <0.900 and some models failed.
Although model fit differed with the drying conditions of temperature and air velocity, the Midilli model could be applied to all data with reasonable prediction of MR values as shown in Figure 7 for the drying data at 50°C at 0.5 to 1.5 m/s. Other models which best fit the data in this study included the Other studies on lemongrass leaves have reported model fit to differ with drying method. Waewsak et al. (2006) found the Wang and Singh model to best describe the drying data for lemongrass leaves dried in a biomass dryer at 60°C, of the thirteen models tested. Ibrahim et al. (2009) found the Newton model to best fit the data for lemongrass leaves dried in a constant temperature and humidity chamber at three drying temperatures (35, 45, 55°C) and three relative humidity conditions (30, 40, and 50%) at a fixed air velocity of 1 m/s. Coradi et al. (2014) reported that the two-term model best fits the drying data (40-70°C) for cut leaves dried in a fixed bed dryer.
In general, thin-layer model fit varies widely with respect to leafy materials and is found to depend on drying method and temperature. The Midilli model has been reported to best describe the drying data for bay leaves, verbena leaves microwave-dried spinach leaves (Barbosa et al., 2007;Cakmak et al., 2013;Doymaz, 2014;Simha & Gugalia, 2013). The Page model has been reported to best describe the MR data for hot air-dried bay leaves, coriander, fever leaves, and spinach (Gunhan, Demir, Hancioglu, & Hepbasli, 2005;Shaw et al., 2007;Simha & Gugalia, 2013;Sobukola & Dairo, 2007) while the Verma, logarithmic, and two-term models were used in other studies (Doymaz, 2006;Kadam et al., 2011;Premi et al., 2010). Erbay and Icier (2010) reported that the modified Henderson and Pabis model TA B L E 6 (Continued) (Continues) best fits the MR data for olive leaves dried at 50 to 70°C at 0.5 to 1.5 m/s.

| CON CLUS IONS
Increasing temperature from 40 to 60°C resulted in a dramatic decrease in total drying time, increase in drying rate, and decrease in equilibrium moisture content of leaves. Increasing temperature to from 40 to 60°C resulted in an average decrease in total drying time of 71%. The effect of air velocity was more important during the initial stages of drying, and insignificant at lower moisture values.
Additionally, drying of leaves at higher velocities of 2.0 m/s is not recommended as the leaves blow about as they dry and become lighter. Drying at all temperatures and velocity combinations took place in the falling rate period. Overall, the average equilibrium moisture values of leaves dried at 40°C were approximately 35% higher than the average value for leaves dried at 50°C and 80% higher than the average value for leaves dried at 60°C. Increasing temperature from 40 to 60°C resulted in an average decrease in total drying time of 71%, while further increasing the temperature to 70°C resulted in an average decrease of 78%. Drying rate constants and diffusivity values were successfully determined.
Model fit varied according to specific temperature-velocity com- F I G U R E 7 Comparison of predicted versus experimental moisture ratio values for lemongrass leaves dried at 50°C using the Midilli model in terms of subjective ratings was adversely affected at temperatures above 50°C. As the best possible combination of drying temperature and air velocity based on drying time and overall quality attributes, drying of lemongrass leaves at 50°C is recommended at a velocity of 1 m/s, as further increasing the air velocity to 2 m/s will not improve drying time.

CO N FLI C TO FI NTE R E S T
None declared.