biomass and bioenergy 34 (2010) 411–416

Available at www.sciencedirect.com

http://www.elsevier.com/locate/biombioe

Modeling of the equilibrium moisture content (EMC)

of Miscanthus (Miscanthus 3 giganteus)

A. Arabhosseini a,*, W. Huisman b, J. Müller c

a
College of Abouraihan, University of Tehran, P.O.Box 11365-4117, Tehran, Iran
b
Farm Technology Group, Wageningen University, The Netherlands
c
Institute of Agricultural Engineering, University of Hohenheim, Stuttgart, Germany

article info abstract

Article history: The desorption isotherms of miscanthus, Miscanthus  giganteus (stems and leaves) were
Received 1 August 2008 determined separately by using the saturated salt solutions method at three temperatures
Received in revised form (25, 50 and 70  C) within a range from 5 to 90% relative humidity. Experimental curves of
6 December 2009 desorption isotherms were fitted to Modified Henderson, Modified Halsey, Modified Oswin
Accepted 9 December 2009 and Modified Chung-Pfost models and evaluated by Residual Sum of Squares (RSS), Stan-
Available online 28 December 2009 dard Error Estimation (SEE) and Mean Relative Deviation (MRD) methods. The Modified
Oswin model was found to be the most suitable for describing the correlation between
Keywords: equilibrium moisture content, relative humidity and temperature.
Biomass ª 2009 Elsevier Ltd. All rights reserved.
Desorption
Drying
Hygroscopic salts
Miscanthus
Miscanthus  giganteus
Postharvest

1. Introduction A lower cutting height will result in more harvested

material, higher leaf content (from fallen leaves) and therefore
Miscanthus, Miscanthus  giganteus, is a perennial rhizoma- higher moisture content. Drying in a swath might be a solu-
tous grass with the C4 photosynthetic pathway. The genus tion to this problem [2].
Miscanthus originates from the tropics and subtropics, Long-term storage is possible after drying, either in the
although different species can be found throughout a wide field or in storage facilities, provided the moisture content is
climatic range in East Asia [1]. Crops can reach a height of 3 m below 15% w.b. Since adequate drying in the spring, prior to
resembling reed and bamboo [2]. Interest in this crop as harvesting cannot always be assured, arrangements for
a resource is focused on new biomass or fiber to be used as drying in storage will be necessary [2].
a building material, for geotextiles, energy production, paper, The moisture sorption isotherm defines the relation
as a packaging material and as substrates for hydroponics [3]. between Equilibrium Relative Humidity (ERH) and Equilibrium
There is a trade-off between biomass production and Moisture Content (EMC) [5]. This information is required for
quality in terms of moisture content. The earlier the harvest the drying and storage of agricultural products as well as to be
time begins the higher the drying costs [4]. able to stop the drying process before the moisture content

* Corresponding author. Tel.: þ98 292 30 210 43; fax: þ98 292 30 254 11.
E-mail address: ahosseini@ut.ac.ir (A. Arabhosseini).
0961-9534/$ – see front matter ª 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.biombioe.2009.12.004
412 biomass and bioenergy 34 (2010) 411–416

3 3 3 3
1

Fig. 1 – The equipment for the desorption isotherm measurement: (1) cryostat; (2) water bath; (3) jars with cap including
netballs with samples, thymol and saturated salt solution; (4) water pump.

has reached a specified safety value of (15%) [6] in order to save glass jars and a cryostat with an accuracy of 0.1  C. The glass
energy. jars were closed with a cap. Every 800 ml glass jar contained
EMC is defined as the moisture content of a hygroscopic 150 ml of saturated salt solution. A base plate prevented
material in equilibrium with a particular environment with contact between the samples and salt solution. Stainless steel
respect to temperature and relative humidity. EMC of the netballs were used to position the stems and leaves separately
product is the final result of moisture exchange between the in the jars (Fig. 1). In tests with RH above 60%, crystalline
product and the air surrounding the sample. In this condition, thymol was placed in the bottles to prevent microbial spoilage
the water in a product is in balance with the moisture in the [8]. Every three days, the samples were weighed to an accuracy
surrounding atmosphere [7]. The relative humidity for this of 0.1 mg. Equilibrium was acknowledged when three
condition is known as the ERH [5]. Desorption isotherms are consecutive weight measurements showed a difference of
measured during the time in which products lose moisture less than 1 mg.
starting from the wet state and adsorption isotherms show When equilibrium had been reached (after approximately 3
the amount of moisture added to products during storage weeks depending on temperature and RH), the samples were
from the dry state. dried in an oven (105  C for 24 h) in order to obtain dry matter
There is a general lack of information available concerning [9]. EMC was determined by taking measurements in triplicate
the desorption isotherms of miscanthus. Sorption isotherms on each sample.
of miscanthus were found only for ERHs above 50% and for
temperatures of 10 and 20  C [3]. As the desorption isotherm is
one of the basic parameters of the drying process, this work 2.3. Salt solution
should provide an important contribution for the optimization
of the drying of miscanthus. The objective of this study is to The saturated salt solutions method is based on the fact that
find the desorption isotherm curve of miscanthus (stem and ERHs of specific salt solutions are a well known physical
leaf) at temperatures between 25 and 70  C for ERH values in property [10]. Table 1 gives the equivalent RH of the selected
the range from 5 to 90% and to identify an appropriate model salt solutions at three temperatures. Salt solutions were
to generalize the empirical results.

2. Materials and methods Table 1 – Saturated salt solutions used in the experiments
and Equilibrium Relative Humidity at different
2.1. Plant material temperatures [10,12].
No. Salt formula Equilibrium relative humidity (ERH) (%)
Miscanthus stems and leaves have been used for this T ¼ 25  C T ¼ 50  C T ¼ 70  C
research. The crop, which was planted three years before
doing this research, was harvested at 10 cm above the roots at 1 K2SO4 97.30 95.82 94.5
2 KNO3 93.58 84.78 n.f.a
Wageningen University farm (The Netherlands) at the end of
3 KCl 84.34 81.20 79.49
the growing season (October). The immediate moisture 4 KBr 80.89 79.02 79.07
content of the crop was about 20–25% (wb). The harvested 5 NaCl 75.29 74.43 75.06
crops were kept in plastic bags in a cool room at 5  C for the 6 NaNO3 74.25 69.04 66.04
experiments. Stems and leaves were separated immediately 7 KI 68.86 64.49 61.93
after harvesting. The stems were chopped to 20–30 mm in 8 NaBr 57.57 50.93 49.70
9 K2CO3 43.16 40.91 37.37
length. Samples of 3–5 g of fresh leaves and 10–15 g of fresh
10 MgCl2 32.78 30.54 27.77
stems were used for desorption tests.
11 LiI 17.56 12.38 7.23
12 LiCl 11.30 11.10 10.7
2.2. Experimental procedure 13 ZnBr2 7.75 7.70 8.72
14 LiBr 6.37 5.53 5.23
Temperatures of 25, 50 and 70  C were maintained by utilizing
a n.f.: Not found in literature, so not applied.
an insulated water bath measuring (56  36  26 cm) with 14
 biomass and bioenergy 34 (2010) 411–416 413

0.5 0.5
25 °C Modified Henderson Modified Halsey
25 °C
50 °C
0.4 0.4 50 °C

EMC, d.b. (decimal)

EMC, d.b. (decimal)

70 °C
70 °C

0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0
0.0 0.5 1.0 0.0 0.5 1.0
ERH ERH

0.5 0.5
25 °C Modified Oswin 25 °C
0.4 0.4 Modified Chung Pfost
EMC, d.b. (decimal)

EMC, d.b. (decimal)

50 °C 50 °C
70 °C 70 °C
0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0
0.0 0.5 1.0 0.0 0.5 1.0
ERH ERH

Fig. 2 – Desorption isotherms of miscanthus leaves at 25 (,), 50 (>) and 70 (7) 8C and the fitted curves of the Modified
Henderson, Halsey, Oswin and Chung-Pfost equations.

0.5 0.5
25 °C Modified Henderson 25 °C Modified Halsey
0.4 50 °C 0.4 50 °C
EMC, d.b. (decimal)

EMC, d.b. (decimal)

70 °C 70 °C
0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0
0.0 0.5 1.0 0.0 0.5 1.0
ERH ERH

0.5 0.5
25 °C 25 °C
0.4 Modified Oswin 0.4 Modified Chung Pfost
EMC, d.b. (decimal)

EMC, d.b. (decimal)

50 °C 50 °C
70 °C 70 °C
0.3 0.3

0.2 0.2

0.1 0.1

0.0 0.0
0.0 0.5 1.0 0.0 0.5 1.0
ERH ERH

Fig. 3 – Desorption isotherms of miscanthus stems at 25 (,), 50 (>) and 70 (7) 8C and the fitted curves of the Henderson,
Halsey, Oswin and Chung-Pfost equations.
414 biomass and bioenergy 34 (2010) 411–416

0 .5
Table 2 – comparison of the desorption data for
Leaves 50°C miscanthus stems and leaves.
Stems 50°C
0 .4 Statistical Temperature,  C
parameter
All temperatures 25 50 70
EMC d .b ., (d e ci ma l )

0 .3 P-value 0.0001 0.0001 0.3441 0.0004

Student’s t-test 4.74a 6.41a 0.99n.s 4.76a
n.s
0 .2 Not significant at 0.01 probability level.
a Significant at 0.01 probability level.

0 .1  1=C3
1
Modified Henderson : EMC ¼  lnð1  ERHÞ (1)
C1 ðT þ C2 Þ
0 .0
0 .0 0 .2 0 .4 0 .6 0 .8 1 .0  1=C3
expðC1 þ C2 TÞ
ERH, (decimal) Modified Halsely : EMC ¼ (2)
lnðERHÞ
Fig. 4 – Desorption isotherm data of miscanthus stems and
 1=C3
leaves at 50 8C. ERH
Modified Oswin : EMC ¼ ðC1 þ C2 TÞ (3)
1  ERH

 
1 ðC2  TÞ
Modified Chung-Pfost : EMC ¼ ln lnðERHÞ (4)
C1 C3
provided by mixing and dissolving salt crystals in distilled
water at a temperature 20  C higher than the previous EMC (dry basis) and ERH are given in metric units. C1, C2 and
temperature in the water bath, to maintain a saturated C3 are the model coefficients and T is temperature in  C.
solution. To avoid the formation of a condensation gradient Nonlinear regression was used to fit the four equations to
in the liquid phase, the salt solution was agitated from time the experimental data using Matlab (the Mathworks). The
to time [11]. quality of the fitted curves was evaluated by using Residual
Sum of Square (RSS), Standard Error Estimation (SEE) and
Mean Relative Deviation (MRD) [15]:
2.4. Models for desorption
X
m
 2
RSS ¼ EMCpredicted  EMC (5)
A number of models have been suggested in the literature to i¼1
describe the relationship between EMC and ERH. The Modified
Henderson, Modified Oswin, Modified Halsey and Modified ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
sP 2ffi
m 
Chung-Pfost equations [13] have been adopted by the Amer- i¼1 EMCpredicted  EMC
SEE ¼ (6)
df
ican Society of Agricultural Engineers as standard equations
for describing sorption isotherms [14]. The equations have
been transformed to obtain the equilibrium moisture content
MRD ¼
1X m
jEMCpredicted  EMCj (7)
(EMC) as a dependent parameter and the equilibrium relative m i¼1 EMCpredicted
humidity (ERH) as an independent variable.
where EMCpredicted is the predicted value of EMC, m is the number
of data points and df is the number of degrees of freedom.
0.5 The smaller the values of these statistical parameters the
T=25°C
more accurately will the data fit the model.
T=50°C The residuals of the EMC, obtained for each model, were
0.4
T=70°C plotted against the measured values and assessed visually as
EMC lea ve s (decimal)

Regression line random or patterned. If the residual plots indicated a clear

0.3 pattern, the model was not accepted [16]. Furthermore, the
experimental results of desorption isotherms of miscanthus
leaves and stems were compared by t-testing for all data as
0.2 well as the data for each separate temperature to establish the
y = 1.0272x
differences between stems and leaves.
R2 = 0.9837
0.1

3. Results and discussion

0.0
0.0 0.1 0.2 0.3 0.4 0.5
3.1. Experimental results
EMC stems (decimal)

Fig. 5 – Equilibrium moisture content EMC of miscanthus Fig. 2 shows the experimental results of desorption isotherms
leaves vs. stems for 25, 50 and 70 8C. of miscanthus leaves and the fitted curves of the selected
 biomass and bioenergy 34 (2010) 411–416 415

Table 3 – Estimated coefficients and quality of fitting of four desorption isotherm models for stems and leaves of
miscanthus at 25, 50 and 70 8C.
Parameters Estimated values and the variance of the models and statistical parameters

Sorption models

Henderson Halsey Oswin Chung-Pfost

Stems
C1 0.222  0.038 4.438  0.109 0.132  0.005 0.156  0.011
C2 1.123  0.049 (2.84  0.30)  102 (1.10  0.07)  102 0.026  0.073
C3 17.011  5.685 2.104  0.070 2.406  0.066 1.147  0.234
RSS 0.010 0.009 0.005 0.039
SEE 0.017 0.016 0.012 0.034
MRD 0.244 0.229 0.119 0.487
Residual Systematic Random Random Systematic

Leaves
C1 0.435  0.089 4.491  0.139 0.155  0.007 (16.54  1.19)  102
C2 1.349  0.067 0.036  0.004 (1.37  1.13)  103 9.008  4.504
C3 2.166  4.002 2.361  0.093 2.729  0.102 117.24  21.72
RSS 0.012 0.012 0.009 0.037
SEE 0.018 0.018 0.016 0.034
MRD 0.218 0.175 0.119 0.401
Residual Systematic Random Random Systematic

equations at 25, 50 and 70  C. The typical S-shaped (Sigmoid) The Modified Oswin equation is the most suitable as it had the
curves were found for all three temperatures. When lowest values for RSS, SEE and MRD and a random pattern of
comparing the three temperatures at constant RH, it can be residuals.
seen that the EMC increases with decreasing temperature.
The desorption data for miscanthus stems, at the three
3.3. Discussion
temperatures, is shown in Fig. 3. The same typical S-shaped
curves were also found for stems, but the levels are slightly
Sorption isotherms of miscanthus were available in the liter-
lower compared to those of leaves.
ature but only for ERHs above 50% and for temperatures of 10
The results of miscanthus leaves and stems were
and 20  C [3]. Those results were close to our results at similar
compared at the same condition to check for differences
conditions. As the availability of sorption isotherms of mis-
between both fractions. Fig. 4 presents the experimental data
canthus was somewhat limited in the literature, the results
of desorption of miscanthus stems and leaves at a tempera-
were also compared with data of other crops. The comparison
ture of 50  C. The level of the leaves was only a little higher
showed that the EMC of miscanthus is almost similar to maize
than that of the stems. Fig. 5 shows data of leaves compared to
straw; rice straw and wheat straw, especially at relative
stems for 25, 50 and 70  C. The inclination of the regression
humidities lower than 60% [17].
line exhibits a factor of 1.03, again indicating a slightly higher
Miscanthus is often harvested in late winter–early spring,
EMC of the leaves compared to the stems. The results of the t-
when the biomass has reached a low moisture content
test showed a significant difference between the stems and
leaves at 25 and 70  C and for the entire data (Table 2).
6
T=2 5
3.2. Fitting of sorption models to experimental data T=5 0
4
T=7 0
The Modified Henderson, Modified Halsey, Modified Oswin
Res idual, %

and Modified Chung-Pfost models (Eqs. (1)–(4)) were fitted to 2

the experimental sorption data of the miscanthus (leaf and
stem). The resulting coefficients of the equations for leaves 0
and stems are shown in Table 3 along with RSS, SEE, MRD and
the observations of the residual plots.
-2
The residuals of EMC of the selected models at the three
temperatures were plotted against the experimental data. The
-4
pattern was observed to be, either randomly or systematically 0.0 0.5 1.0
distributed. In most cases the Modified Oswin equation had
ERH
a random distribution in the residual plots (Fig. 6) and the
distribution showed a systematic pattern for the other equa- Fig. 6 – Residual plots for the Modified Oswin equation for
tions. The residual plots for the Modified Henderson equation desorption data of miscanthus leaves at temperatures with
with a systematic distribution is shown as an example (Fig. 7). random distribution.
416 biomass and bioenergy 34 (2010) 411–416

3 references

2
[1] Lewandowski I, Clifton-Brown JC, Scurlock JMO, Huisman W.
1 Miscanthus: European experience with a novel energy crop.
Res idual, %

Biomass Bioenergy 2000;19(4):209–27.

[2] Venturi P, Huisman W, Molenaar J. Mechanization and costs
0
of primary production chains for Miscanthus  giganteus in
The Netherlands. J Agric Eng Res 1998;69:209–15.
-1 [3] Huisman W, Kortleve WJ. Mechanization of crop
T=2 5 establishment, harvest, and post-harvest conservation of
-2 T=5 0 Miscanthus sinensis giganteus. Ind Crops Prod 1994;2(4):289–97.
T=7 0
[4] Danalatos NG, Archonotoulis SV, Mitsios I. Potential growth
and biomass productivity of Miscanthus giganteus as affected
-3
0.0 0.5 1.0 by plant density and N-fertilization in central Greece.
ERH Biomass Bioenergy 2006;31(2(3)):145–52.
[5] Soysal Y, Oztekin S. Equilibrium moisture content equations
Fig. 7 – Residual plots for the Modified Henderson equation for some medicinal and aromatic plants. J Agric Eng Res 1999;
for desorption data of miscanthus stems at three 74(3):317–24.
[6] Hamer PJC, Knight AC, McGechan MB, Cooper G. Model for
temperatures with systematic distribution.
predicting the field drying characteristics of grass
conditioned by maceration (severe treatment). J Agric Eng
Res 2000;75(3):275–89.
[7] Silakul T, Jindal VK. Equilibrium moisture content isotherms
avoiding added costs for biomass drying. Nevertheless during
of mungbean. Int J Food Prop 2002;5(1):25–35.
winter, the crop loses a considerable part of dry matter, which [8] Menkov ND. Moisture sorption isotherms of chickpea seeds
could be retrieved with an early harvest before winter sets in at several temperatures. J Food Eng 2000;45(4):189–94.
[18]. Late harvesting causes low quality and less drying costs [9] Park KJ, Vohnikova Z, Brod FPR. Evaluation of drying
and vise versa. Drying costs are also related to the final parameters and desorption isotherms of garden mint leaves
moisture content in the drying process. Based on our results (Mentha crispa L.). J Food Eng 2002;51(3):193–9.
a final moisture content of 10% w.b. can be recommended for [10] Greenspan L. Humidity fixed points of binary saturated
aqueous solutions. J Res Nat Bur Stand A Phys Chem 1976;
a safe storage at 25  C. In further research on drying, advice
81A(1):89–96.
can be given on the optimum harvesting time by considering [11] Kouhila M, Belghit A, Daguenet M, Boutaleb BC. Experimental
the quality and cost factors involved. determination of the sorption isotherms of mint (Mentha
viridis), sage (Salvia officinalis) and verbena (Lippia citriodora). J
Food Eng 2001;47(4):281–7.
[12] Labuza TP, Kaanane A, Chen JY. Effect of temperature on the
4. Conclusion moisture sorption isotherms and water activity shift of 2
dehydrated foods. J Food Sci 1985;50(2):385–91.
Applying different models to fit experimental data of [13] Chen C. Sorption isotherms of sweet potato slices. Biosyst
desorption isotherms of miscanthus has revealed the Modi- Eng 2002;83(1):85–95.
[14] ASAE. Moisture relations of plant-based agricultural
fied Oswin Equation as being the most accurate one. Coeffi-
products. In: Standards engineering practice data. 50th ed.
cients for the various models are now available for ASAE Standards; 2003. p. 538. D245.5,.
a temperature range from 25 to 70  C, i.e. EMC can now be [15] Sun DW. Comparison and selection of EMC ERH isotherm
estimated for all temperature within that range. The EMC equations for rice. J Stored Prod Res 1999;35(3):249–64.
level of the leaves of miscanthus has shown to be slightly [16] Chen CC, Morey RV. Comparison of 4 EMC/ERH equations.
higher than that of the stems. That means if dried together, Trans ASAE 1989;32(3):983–90.
[17] Singh RN. Equilibrium moisture content of biomass
the EMC of the leaves should be used as a critical value. As
briquettes. Biomass Bioenergy 2004;26(3):251–3.
mould growth occurs at RH above 70%, then ERH of 60% is
[18] Ercoli L, Mariotti M, Masoni A, Bonari E. Effect of irrigation
recommended for safe storage. That means a moisture and nitrogen fertilization on biomass yield and efficiency of
content of 0.11 d.b. or 10% w.b. has to be reached for the energy use in crop production of Miscanthus. Field Crops Res
drying of miscanthus. 1999;63(1):3–11.