close

Вход

Забыли?

вход по аккаунту

?

j.ijrefrig.2018.07.019

код для вставкиСкачать
Accepted Manuscript
Comparison of Ab- and Adsorbents for Low Temperature Heat Driven
Sorption Cooling Machines
Roland Kühn , Julia Römer , Felix Ziegler
PII:
DOI:
Reference:
S0140-7007(18)30265-2
10.1016/j.ijrefrig.2018.07.019
JIJR 4049
To appear in:
International Journal of Refrigeration
Received date:
Revised date:
Accepted date:
14 March 2018
25 May 2018
6 October 2018
Please cite this article as: Roland Kühn , Julia Römer , Felix Ziegler , Comparison of Ab- and Adsorbents for Low Temperature Heat Driven Sorption Cooling Machines, International Journal of Refrigeration (2018), doi: 10.1016/j.ijrefrig.2018.07.019
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service
to our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and
all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
Highlights:
 There are sorbents being thermodynamically more suitable for low driving temperatures
 Specialized zeotypes may have a strongly restricted operation field
 Liquid sorbents and Silica gel are quite flexible by means of operation conditions
 Ionic liquids may work at lower driving temperatures than inorganic salts
 Crystallization curves for lithium bromide are given
CR
IP
T
Comparison of Ab- and Adsorbents for Low Temperature Heat Driven
Sorption Cooling Machines
Roland Kühn1, Julia Römer2, Felix Ziegler3
1
Dipl.-Ing., Researcher at Chair of Energy Conversion Technology,
Technische Universität Berlin, Marchstr. 18, 10587 Berlin, Germany, +49 (0)30-314-22170, roland.kuehn@tu-berlin.de
2
M.Sc., Researcher at Chair of Energy Conversion Technology,
AN
US
Technische Universität Berlin, Marchstr. 18, 10587 Berlin
and CEO of Coolar UG, Wollankstr. 118, 13187 Berlin
3
Prof. Dr.-Ing., Head of Chair of Energy Conversion Technology,
Technische Universität Berlin, Marchstr. 18, 10587 Berlin
Abstract
ED
M
Thermally driven cooling machines allow for reduction of CO2 emissions when using waste heat from
industrial processes or cogeneration, or when using solar heat. The lower the necessary driving temperature
level is the more heat sources are available to be used, or the more thermal power can be drawn from a heat
source, or the smaller and cheaper the heat exchanger for a given power can be.
CE
PT
There have been debates for decades whether adsorption chillers can be driven by heat sources at lower
temperatures than absorption chillers. The paper at hand intends to continue this debate on a rather
fundamental level. The inorganic salt lithium bromide which is the most common absorbent and the common
adsorbent RD-Silica gel are compared to organic salts (ionic liquids) as designable absorbents on the one
hand, and adsorbents designed for use at low driving temperatures, on the other hand. For that purpose,
functions for the vapor pressure depression of the discussed sorbents are required. Simple correlations for the
absorbents are derived in this paper, whereas the correlations for the adsorbents are taken from literature. The
crystallization lines of lithium bromide derived from experimental data from two different sources are also
discussed shortly.
AC
From these correlations, the minimum desorption temperature at various boundary conditions is derived for
five sorbents and the influence of the desorption temperature on the change of composition in the sorbents is
shown. The latter is traced back to the shape of the sorption isotherms. At small temperature lifts the
zeotypes and the ionic liquid [mmim][DMP] can be operated with lowest driving temperatures, but at higher
temperature lifts the zeotypes are inoperable. Finally, it is concluded that the possibility to tailoring sorbents
according to the customer’s needs prevails for both liquid and solid sorption as a pre-requisite to operate with
very low driving temperatures.
Keywords: adsorption; absorption; refrigeration; low temperature heat; ionic liquids; crystallization data
1
ACCEPTED MANUSCRIPT
Nomenclature
[J kg-1]
heat of vaporization
[J kg-1]
slope
[K Pa-1]
mass of dry ab-/adsorbent
[kg]
mass of refrigerant
[kg]
ordinate intercept
[Pa]
loading
[kgrefrigerant kgsorbent-1]
maximum loading
[kgrefrigerant kgsorbent-1]
dimensionless loading q/qm
[-]
5
CR
IP
T
̃
heat of adsorption
[Pa]
evaporator pressure
[Pa]
condenser pressure
[Pa]
saturation pressure
[Pa]
saturation pressure of pure refrigerant
[Pa]
mass fraction of sorbent in mixture
[kgsorbent kgmixture-1]
̃
molar fraction of sorbent in mixture
[molsorbent molmixture-1]
A
Absorber
C
Condenser
E
Evaporator
G
Generator
̃
molar mass of ab-/adsorbent
[g mol-1]
̃
molar mass of refrigerant
[gmol-1]
̇
low temperature heat flow
[W]
̇
heat flow rejected from sorber
[W]
̇
heat flow rejected from condenser
[W]
̇
driving heat flow
[W]
absolute temperature
[K]
AC
CE
PT
ED
M
AN
US
normal pressure (10 Pa)
evaporation temperature
[K]
lowest sorption temperature
[K]
condensation temperature
[K]
highest desorption temperature
[K]
dew point temperature
[K]
boiling temperature
[K]
mass fraction of refrigerant in mixture
[kgrefrigerant kgmixture-1]
standard deviation
2
ACCEPTED MANUSCRIPT
1. Introduction
CR
IP
T
Thermally driven cooling machines or chillers allow for reduction of CO2 emissions compared to mechanical
vapor compression technology when using waste heat from industrial processes or cogeneration, or when
using solar heat. Those heat sources often feature low temperatures which are not economic for electricity
production but still have enough exergy to drive thermally driven cooling devices, which in most cases are
realized as sorption processes. Nevertheless, also to drive a sorption process a certain minimum temperature
level is needed and the lower the temperature level is the more heat sources are available to be used.
Moreover, the lower the equilibrium temperature for desorption is, the more thermal power can be drawn
from a heat source, or the smaller and cheaper the heat exchanger for a given power can be. On the other
hand, of course, at least the reversible COP will be lower, too.
AN
US
Sorption processes mostly come as absorption cycles using liquid absorbents and adsorption or chemical
reaction cycles using solid sorbents. For a long time there have been debates about which technology
provides the opportunity to be driven by heat sources at lowest temperatures (Saha et al., 2003; Henning,
2007; Lamb & Ziegler, 1996). In fact, the minimal driving temperature is often set by design constraints of
the chiller, e.g., size of the exchangers, temperature glide of the driving heat source, and circulation ratio
(Lamb & Ziegler, 1996). Thermodynamically the lowest driving temperature level is set by the sorption
material and the used refrigerant. Systems using chemical reactions theoretically have the best potential but
are less mature and suffer from other restrictions such as slow heat and mass transfer and, therefore, will not
be considered here.
M
The aim of this paper is to compare common absorbents like the inorganic salt lithium bromide (LiBr) (Kim
& Infante Ferreira, 2008) and common adsorbents like RD-Silica gel (Henninger et al., 2010) with organic
salts (ionic liquids: IL) as designable absorbents (Seiler et al., 2010) and adsorbents designed for use at low
driving temperatures, such as zeolites or zeotypes (Aristov, 2007). As background to compare the ad- and
absorbents with each other, functions for the vapor pressure depression of the discussed sorbents will be
presented. They describe the main differences of the boiling lines allowing particular sorption media to be
used with driving heat at low temperatures.
AC
CE
PT
ED
Using the lowest possible desorption temperature will not result in a feasible process. Only with somewhat
higher temperatures a composition change between ab- or adsorption and desorption will be achieved. This
composition change greatly influences the efficiency and investment cost of a sorption cooling machine and
will be discussed, also.
3
ACCEPTED MANUSCRIPT
2. Background
AN
US
CR
IP
T
Sorption cooling machines consist of at least two different components with the features of an evaporator E
and a condenser C on the one hand, and of an ab- or adsorber A and a desorber D on the other hand. Figure 1
shows a schematic of a sorption cycle in a logarithmic pressure over inverse temperature plot (van’t Hoff
representation). In the evaporator a low temperature heat flow ̇ enters the system and the refrigerant
evaporates. When the refrigerant content in the sorbent is small enough it will suck the refrigerant vapor.
Due to the vapor pressure depression of the refrigerant in the sorbent the refrigerant’s heat of sorption ̇ is
̇ at a higher
released at an elevated temperature
, usually to the ambient. To regenerate the sorbent, heat
temperature level has to be supplied to the desorber. The temperature has to be high enough to overcome
the vapor pressure depression, in order to increase the vapor pressure of the refrigerant in the sorbent up to a
little bit over the condenser pressure. The refrigerant flows to the condenser and is condensed there, releasing
the heat of condensation ̇ to the ambient also. The temperature in the condenser
and evaporator fix
the pressure level in the ab-/adsorber and desorber , if no flow resistance is assumed. The temperature
in the ab-/adsorber
and desorber then determines the refrigerant mass fraction in the sorbent. The
process will only work when the refrigerant content in the sorbent at the outlet of the desorber is lower than
the refrigerant content at the outlet of the ab- or adsorber. Thus it is crucial to know the vapor pressure
behavior of the sorbent/refrigerant pair, to know the temperature levels the sorption system can operate with.
E
ED
p0
C
D
T1A T1C
T2
M
log(p)
p1
T0
A
-1/T
PT
Figure 1: Schematic of a sorption cycle in a van’t Hoff diagram
2.1 Measures for composition
AC
CE
In order to describe the vapor pressure depression with respect to the refrigerant vapor pressure usually
different types of equations and even different measures for the refrigerant (R) content in the sorbent (A) are
used for adsorbents and absorbents, and the visualization is done in different types of diagrams. The
refrigerant content in solid adsorbents is normally described by the loading q, the ratio of the refrigerant mass
to the mass of the dry adsorbent:
(1)
The loading lies between zero and a maximal value, which is specific for the material used. If more
refrigerant is added the sorbent will be flooded: there will be no vapor pressure depression anymore and the
sorbent will have no influence on the phase equilibrium anymore. This maximal loading usually is, but don’t
has to be, below 1 for solids; one exception is the solid absorbent poly acrylic acid (PAA) which can adsorb
more water than its own mass (Buchholz & Graham, 1997).
The refrigerant content in liquid absorbents is commonly described by the refrigerant ( ) or dry sorbent (x)
mass fraction (eq. (2)). The mass fraction lies between 0 and 1.
(2)
4
ACCEPTED MANUSCRIPT
The loading and the mass fractions can be converted into each other by eq. (3). Moreover, the molar dry
sorbent fraction ̃ can be derived from the mass fraction x with eq. (4).
(3)
̃
((
)
̃
(4)
)
̃
2.2 Boiling lines of solutions
N
N
Li
AN
US
Cation
CR
IP
T
In Figure 3 and Figure 4 the experimental boiling data of Löwer (1960) are plotted for aqueous solutions of
LiBr and those from He et al. (2010) for aqueous solutions of 1-methyl-3-methylimidazolium
dimethylphosphate [mmim][DMP], respectively. The chemical structure of these absorbents and their molar
mass is given in Figure 2. [mmim][DMP] is proposed as good IL for water absorption (Zhao et al., 2006;
Liang et al., 2011) being stable up to high temperatures of above 160 °C (Liang et al., 2011).
O
P
O
Anion
OCH3
Br
OCH3
Molar mass
[g/mol]
222.2
(He et.al., 2010)
86.9
M
Figure 2: chemical structure and molar mass of [mmim][DMP] (left) and LiBr (right)
CE
PT
ED
In addition to the experimental data, fitting lines are shown in both vapor pressure plots. To describe the
vapor pressure lines (boiling lines) of salt solutions different methods are available, reaching from very
simple physical models (Othmer & Fröhlich, 1960) over empirical models with polynomial fitting
(Feuerecker, 1994) to complex physical models with empirical parameter fitting (He et al., 2010). In 1960
Othmer and Fröhlich proposed the description of vapor pressure lines with the Clausius-Clapeyron equation,
approximating the vapor as ideal gas, which leads to a linear dependency of the logarithmic vapor pressure
over the inverse temperature in Kelvin. Although this linear dependency is not totally true, for most
substances and solutions it is a valid approximation over a relatively wide temperature range, which can be
improved with empirical adaptations (Löwer, 1960).
2.2.1 Vapor pressure lines of aqueous lithium bromide
AC
Using this simple model of Othmer and Fröhlich (1960) for the strong electrolyte LiBr, the vapor pressure
can be described in dependence of molar salt fraction ̃ and temperature T by equation (5), leading to the
boiling lines in the vapor pressure plot (van’t Hoff diagram) in Figure 3. The resulting coefficients can be
seen in Table 3 on page 7, left column, where the coefficients and a4 in the paper at hand are derived by the
vapor pressure line of pure water.
( [
])
(
)
(5)
[ ]
Assuming Trouton’s rule with convergence of all vapor pressure lines (or boiling lines) at infinite
temperature (Othmer & Fröhlich, 1960), the slope m can be fitted with only three additional parameters, with
a relative standard deviation of the pressure of 7% at 0 °C boiling temperature and 2.3% at 80 °C. A mean
relative standard deviation over the whole temperature and concentration range of 3.7% was reached, by
using the natural logarithm of the pressure instead of the absolute pressure for the fitting procedure. Equation
(5) can be used for all salt concentrations from pure water up to the crystallization line of LiBr. The latter has
to be determined separately.
5
ACCEPTED MANUSCRIPT
2.2.2 Crystallization line of aqueous lithium bromide
water
x = 0.25
x = 0.4
x = 0.5
x = 0.55
x = 0.6
x = 0.65
x = 0.7
M
100
10
1
ED
vapor pressure in mbar
1000
AN
US
CR
IP
T
Various authors give data for the crystallization line of aqueous LiBr. Two well-known are Boryta (1970)
and Hüttig & Reuscher (1924). In Table 1 the crystallization line derived by linear fits to the data of Boryta
(1970) are given and in Table 2 the crystallization line derived by two parabolic fits to the data of Hüttig &
Reuscher (1924). It should be mentioned, that the temperature limits of the fitted crystallization lines are not
exactly equal to the temperatures given by Hüttig & Reuscher (1924) and Boryta (1970) for the change in
hydration stage (LiBr∙2H2O; LiBr∙H2O). This is due to the fact that the reported data for the change in
hydration do not agree to the measured crystallization data. A linear fit is used for Boryta’s data because
Boryta himself stated that the crystallization data are varying strongly between different sources (Boryta,
1970) and there is an almost linear dependency in his data within the two relevant steps of hydration above
0°C (LiBr∙2H2O; LiBr∙H2O). Comparing the data of Boryta (1970) and Hüttig & Reuscher (1924) the
difference in crystallization concentration is up to 4 mass% at 311 K. In Figure 3 it can be seen that the older
data from Hüttig & Reuscher (1924) seem to agree better to the vapor pressure data of Löwer (1960),
because the highest salt concentrations Löwer measured should crystallize at the corresponding temperatures
according to Boryta’s data. Nevertheless, the solution in the data of Löwer (1960) could have been subcooled
without crystallization, which is a common inaccuracy proposed by Boryta (1970). However, in this paper
the parabolic equation according to Table 2 (Hüttig & Reuscher, 1924) is used.
0
20
Hüttig & Reuscher (1924)
Boryta (1970)
40
60
boiling temperature in °C
80
100
120
140
PT
Figure 3: vapor pressure data of LiBr (Löwer, 1960) (van’t Hoff plot)
CE
Table 1: equations for crystallization line of LiBr fitted to data of Boryta (1970)
[ ]
AC
[ ]
Table 2: equations for crystallization line of LiBr fitted to data of Hüttig & Reuscher (1924)
( [ ])
[ ]
;
( [ ])
[ ]
;
6
ACCEPTED MANUSCRIPT
2.2.3 Vapor pressure lines of [mmim][DMP]
Meyer et al. (2015) used the same method as in 2.2.1 for fitting vapor pressure data of an ionic liquid (IL)
with the refrigerant ethanol, whereas Zegenhagen et al. (2015) adopted this method for a novel ionic liquid
with water as refrigerant. Even though in both cases the result was satisfactory, the novel IL used by
Zegenhagen et al. (2015) showed a slight deviation from the assumption of convergence of all boiling lines at
infinite temperature. Using the above method (which was derived for the strong electrolyte LiBr) for the light
electrolyte [mmim][DMP], however, leads to a remarkable and systematic deviation between fit and
measured values as can be seen in Figure 4. The resulting coefficients can be accessed in Table 3, middle
column, where the coefficients and a4 are derived by the vapor pressure line of pure water again.
1000
CR
IP
T
water
x = 0.59
x = 0.76
x = 0.85
x = 0.9
x = 0.94
100
10
1
0
20
40
AN
US
vapor pressure in mbar
10000
60
80
boiling temperature in °C
100
120
140
160
M
Figure 4: vapor pressure data of [mmim][DMP] from He et al. (2010) (van’t Hoff plot)
(fit: see Table 3, middle column)
LiBr – water
̃
;
CE
;
̃
Mean relative standard deviation
of eq. (5):
AC
[mmim][DMP] – water
Fit including pure water
̃
Fit excluding pure water
̃
PT
̃
ED
Table 3: coefficients for vapor pressure fit equation (5) page 5
crystallization; all T
Mean relative standard deviation
of eq. (5):
; 60 < T < 160°C
Mean relative standard
deviation of eq. (5):
; all T
There is a systematic disagreement of the slope of the lines with the experimental data. In this fit the slope m
is only linearly dependent on the composition, but higher polynomials improve the fit only marginally.
Obviously, the boiling lines according to the experimental data must have a much larger slope than the
boiling line of pure water, and at least the line for the largest water content will intersect with the former at
decent temperatures, using a linear extrapolation (Figure 5). If we assume that the experimental data are
correct the boiling lines cannot converge in one point with the boiling line of the refrigerant at T approaching
infinity. This behavior may indicate erroneous experimental data. Consequently, they should be used with
caution.
A much better fit is achieved when the solution data are treated separately from the pure water. The
coefficients excluding the boiling line of pure water are given in Table 3, right column and the fit is
displayed together with the experimental data in Figure 5. This fit, also linear in composition, will be used,
7
ACCEPTED MANUSCRIPT
although it should only be used in a restricted, namely the measured, range of composition. Especially an
extrapolation to very dilute solutions is surely wrong because the fit does not represent the pure water line.
1000
water
x = 0.59 (He et.al)
x = 0.76 (He et.al)
x = 0.85 (He et.al)
x = 0.9 (He et.al)
x = 0.94 (He et.al)
x = 0.6 (Zhao et.al)
100
10
1
0
20
40
60
80
boiling temperature in °C
CR
IP
T
vapor pressure in mbar
10000
100
120
140
160
AN
US
Figure 5: vapor pressure data of [mmim][DMP] from He et al. (2010) and Zhao et al. (2006) (van’t Hoff plot)
(fit: see Table 3, right column)
Data by Zhao et al. (2006) with 60 mass% of [mmim][DMP] have been plotted in Figure 5 also. They
disagree with the data from He et al. (2010), which also is an indication against the latter. The data of Zhao et
al. (2006) show the expected congruency with pure water all very dilute solutions should show, as 60 mass%
are just 10 molar%. Unfortunately this salt concentration is not in the relevant region. Consequently, in this
paper we have to rely on the data by He et al. (2010).
ED
M
The coefficients n and
are derived by the vapor pressure line of pure water for the left and middle column
of Table 3 as mentioned above. Thus, due to the linearity in composition, there is only at least one vapor
pressure point (pressure and temperature) at one known composition needed to determine
for
[mmim][DMP]. At least one vapor pressure point at three different compositions is required to determine ,
and
for LiBr.
CE
PT
Using the method excluding the vapor pressure line of the pure refrigerant at zero salt content, at least two
vapor pressure points at one composition and one vapor pressure point at a different composition are required
to determine ,
and n for [mmim][DMP]. However, of course, it is strongly recommended to measure
more vapor pressure lines to observe the behavior and to decide which method or grade of polynomial for the
concentration dependency is suitable for the sorbent/refrigerant pair, and to take into account the uncertainty
of the measurement.
AC
2.3 Vapor pressure over solid sorbents
While there is also literature about fitting the vapor pressure curves of calcium chloride impregnated Silica
gel with the Clausius-Clapeyron equation (Aristov et al., 1996), most adsorbents have no linear behavior in
the van’t Hoff plot (Gluesenkamp, 2012). Usually they are described by complex equations or higher order
polynomials which cannot be extrapolated far beyond the intervals of definition (Goldsworthy, 2014).
Therefore, in the paper at hand the equations (6) to (8) derived by Goldsworthy (2014), based on his own
experimental data in a pressure range from 102 to 5∙104 Pa for RD-Silica gel and the low temperature
zeotypes AQSOA-Z01 and AQSOA-Z05 are used. The measured isotherms are in a temperature range from
20°C to 160°C for RD-Silica gel and AQSOA-Z01 and between 20°C and 100°C for AQSOA-Z05. Hereby
equation (6) is an nth order spline representation for a reference isotherm at a reference temperature
,
over a certain dimensionless loading interval with lower and upper bounds ̃ and ̃ , with the dimensionless
loading ̃
, the maximum loading
and the normal pressure
. Equation (7) represents
a spline function of the ratio for isosteric heat of adsorption
to heat of vaporization , over the
8
ACCEPTED MANUSCRIPT
dimensionless loading ̃ at all temperatures. Given the first two equations equation (8) allows to calculate the
logarithm of the equilibrium vapor pressure ( ) using the saturation pressures of pure refrigerant
at
any temperature and at the reference temperature
(Goldsworthy, 2014). The required coefficients are
given in Table 4. Looking at the experimental data and fitting curves from Goldsworthy (2014), it is strictly
recommended not to extrapolate the fits of Z01 and Z05 above a pressure of 5∙10 4 Pa, especially at
temperatures above 80°C.
(̃
)
)
∑
(̃
∑ (̃
( ( ))
( (̃
̃)
̃
̃)
))
̃
̃
*
̃
̃
(
(
(6)
̃
))
CR
IP
T
(
(
( ))+
(7)
(8)
Table 4: vapor pressure equations for the adsorbents from Goldsworthy (2014)
zeotype AQSOA-Z05
̃
̃
̃
̃
̃
̃
̃
ED
PT
̃
̃
̃
̃
̃
̃
̃
̃
̃
̃
AC
CE
̃
̃
M
̃
RD-Silica Gel
AN
US
zeotype AQSOA-Z01
The following figures show the van’t Hoff plots of the solid sorbents based on the fitting curves of
Goldsworthy (2014). The minimal loading plotted is 0.01; further boiling lines are plotted in equidistant
steps of 0.02 from 0.02 on in the domain of definition given by Goldsworthy (2014).
9
ACCEPTED MANUSCRIPT
500
0.18
0.02
0.01
10
1
20
30
40
50
boiling temperature in °C
CR
IP
T
vapor pressure in mbar
water
100
60
70
80
Figure 6: vapor pressure lines of AQSOA-Z01 from fits of Goldsworthy (2014) (van’t Hoff plot)
500
AN
US
100
0.22
10
0.02
M
vapor pressure in mbar
water
1
20
30
40
ED
0.01
50
boiling temperature in °C
60
70
80
Figure 7: vapor pressure lines of AQSOA-Z05 from fits of Goldsworthy (2014) (van’t Hoff plot)
PT
0.3
CE
100
water
AC
vapor pressure in mbar
500
10
0.04
0.02
1
20
30
40
50
boiling temperature in °C
60
70
80
Figure 8: vapor pressure lines of RD Silica gel from fits of Goldsworthy (2014) (van’t Hoff plot)
The van’t Hoff plot of silica gel looks quite normal (Figure 8). The plots for the zeotypes (Figure 6 and 7)
look quite strange. The slopes of the boiling lines seem to vary arbitrarily. The boiling lines cumulate and
cross each other. Obviously, extrapolation is not possible. This behavior has important operational
consequences as will be shown later.
10
ACCEPTED MANUSCRIPT
3. Minimal desorption temperature
3.1 Boiling lines
60
AN
US
CR
IP
T
A sorption cooling machine can only work if the heat source of the desorber has a temperature level which is
high enough to allow the sorbed refrigerant to desorb from the sorbent at a given condensation temperature
level ( ). In the limiting case the total ab- or adsorption process at a given evaporation temperature level
( ) takes place at the same loading or concentration as the desorption process, resulting in no sorption
capacity. Thus, fixing the internal temperature of the evaporator , the internal temperature of the
ab/adsorber
, and the condenser
(the latter two, of course, are correlated with the ambient
temperature), the minimum temperature level of the high temperature heat source is fixed by the sorption
media only. In order to demonstrate this, in Figure 9 one limiting boiling line for each of the five presented
sorption materials for water, namely RD-Silica gel, zeotype AQSOA-Z01 and Z05, inorganic salt LiBr and
IL [mmim][DMP] is plotted. Figure 9 has the form of a Dühring-plot (dew point temperature over boiling
temperature); therefore, the boiling line of the refrigerant pure water is shown also. However, in Figure 9 the
composition or loading is not variable but it is fixed by the temperatures of the evaporator and ab-/adsorber
with
and
. Consequently, only the one resulting boiling line which satisfies these
conditions is shown for all sorbents. The ordinate of Figure 9 can be understood as variable condenser
temperature and the abscissa is the resulting minimum desorption temperature.
RD Silica gel, q=0.172
[mmim][DMP], x=0.612
50
M
40
water
30
20
10
25
30
LiBr, x = 0.50
35
40
45
50
55
60
boiling temperature in °C
CE
5
AQSOA-Z05, q=0.145
ED
25
PT
dew point temperature in °C
AQSOA-Z01, q=0.185
Figure 9: Boiling lines of the five sorbents for
and
and the refrigerant water
AC
It is obvious that, in contrast to the findings from Lamp & Ziegler (1996), there are sorbents which allow
lower desorption temperatures than a LiBr solution. While the boiling line of LiBr/water, as commonly
known, is drifting away from the water line with higher temperatures in the Dühring plot, the boiling lines of
the other media are almost parallel or even converge to the water line. Assuming, furthermore, the condenser
temperature to be equal to the ab-/adsorber temperature, the minimum desorption temperature for each
sorbent is listed in Table 5.
Table 5: Minimum desorption temperature
AQSOA-Z01
AQSOA-Z05
RD-Silica gel
LiBr
[mmim][DMP]
41.3 °C
44.8 °C
44.6 °C
47.5 °C
44.3 °C
The differences in desorption temperature are straightforward consequences from deviations from Trouton’s
rule which holds for LiBr solution only. All sorbents which follow Trouton’s rule and the ClausiusClapeyron-equation will have the same desorption temperature as LiBr solution.
11
ACCEPTED MANUSCRIPT
3.2 Change of loading
According to Table 5 above one would best choose zeotype AQSOA-Z01 to be used at low heat source
temperatures. However, the minimum desorption temperature is only one relevant parameter. For operation
the difference in loading which can be achieved between sorption and desorption is of similar importance. A
difference in loading of 0.1 to 0.2 is typical for absorption systems; in adsorption systems it may even be
larger. Moreover, bad heat and mass transfer can fully turn the tables, but this is out of the scope of this
paper.
CR
IP
T
Figure 10 (left) shows the difference in loading between sorption and desorption for the same constraints as
before, with the condensation temperature at 25°C. The desorption temperature is the abscissa again. The
difference in loading starts to rise from zero as soon as the desorption temperature is raised above its
minimum. The two zeotypes show a stepwise increase whereas the change of the Silica gel and the
absorbents is steady.
M
AN
US
It was stated before that AQSOA-Z01 has the lowest minimum driving temperature with 41.3 °C, but up to a
desorption temperature of 50 °C the difference in loading is just slightly increasing. The IL [mmim][DMP]
and the zeotype AQSOA-Z05 behave quite similar at their minimal driving temperature of around 45 °C. The
increase of difference in loading of the zeotype is very steep in the beginning but it flattens off soon, while
the increase in difference in loading is just slowly degrading for the IL. This leads to a higher difference in
loading of the IL from 48 °C desorption temperature on. The salt LiBr has the highest minimum desorption
temperature at the given constraints (as indicated before in Figure 9), but due to its quite continuous increase
in difference in loading, using LiBr leads to a similar behavior as AQSOA-Z01 at desorption temperatures
between 48 °C and 53 °C and a higher difference in loading than AQSOA-Z05 from 52 °C on. LiBr would
even have a higher difference in loading than [mmim][DMP] at desorption temperatures above 72°C (thin
line), but will crystallize in the absorber before (bold line). The RD-Silica gel is not competitive to the other
sorbents in the given case, i.e. at an evaporation temperature of 5 °C and a condensation and minimum
sorption temperature of 25 °C. Overall, [mmim][DMP] looks best in this case.
0.5
0.45
ED
0.45
0.5
0.4
0.4
0.3
0.15
AQSOA-Z01
AC
0.1
LiBr
AQSOA-Z05
0.25
0.2
difference in loading q
PT
0.35
CE
difference in loading q
[mmim][DMP]
45
LiBr
0.3
0.25
0.2
0.15
[mmim][DMP]
AQSOA-Z01
0.1
RD Silica gel
RD Silica gel
0.05
0.05
0
40
0.35
AQSOA-Z05
50
55
60
65
70
desorption temperature in °C
75
80
0
50
55
60
65
70
75
desorption temperature in °C
80
Figure 10: difference in loading at the temperature triple for evaporator/condenser/ad- or absorber 5/25/25 °C
(left) and 5/30/30 °C (right) for various desorption temperatures
A slight change in heat rejection temperature changes the behavior of the sorbents a lot. Figure 10 (right)
shows the difference in loading over desorption temperature for an evaporation temperature of 5 °C and a
condensation and sorption temperature of 30 °C for the different sorbents. Whereas AQSOA-Z05 has the
highest capacity increase of all sorbents till reaching nearly its maximum loading at an adsorption
temperature of 25 °C, it has nearly no capacity left at an adsorption temperature of 30 °C at any desorption
temperature. In contrast to this, zeotype AQSOA-Z01 has the highest capacity of all sorbents at 30 °C
12
ACCEPTED MANUSCRIPT
sorption temperature until almost reaching its maximum loading.
The sorbents [mmim][DMP] and LiBr reach a higher difference in loading than AQSOA-Z01 from 61 °C
and 65 °C on, respectively. This is only somewhat more than 9 K higher than the minimum desorption
temperature of the zeotype, so it is still in a reasonable range. LiBr now crystallizes after reaching a higher
difference in loading than [mmim][DMP], due to the higher absorption temperatures, but this is out of the
loading range of the diagram. Again, Silica gel has an intermediate minimum desorption temperature but is
not competitive in difference in loading to the other sorbents except to the low temperature zeotype AQSOAZ05.
CR
IP
T
Going even higher in adsorption temperature to 35 °C, also the zeotype AQSOA-Z01 achieves no reasonable
difference in loading anymore (Figure 11, left) and the RD Silica gel then is the adsorbent achieving highest
difference in loading. Increasing the evaporation temperature from 5 °C to 10 °C (10/35/35 °C) the graph
(Figure 11, right) looks like the graph for the temperature triple 5/30/30 °C (Figure 10, right), except of a
shifting of the lines to higher temperatures by 5 K and marginally smaller differences in loading.
AN
US
If all four graphs are compared, please note that the ordinate of Figure 10 and Figure 11 is the same, but in
the latter the empty space on top was cut off. The abscissae were shifted with the minimum desorption
temperature and all four were cut off at 80 °C, due to pressure restrictions of the adsorbent’s fitting curves by
Goldsworthy (2014). Comparing the graphs, an increase of the ab-/adsorption and condensation temperature
by 5 K results in a shift of minimum desorption temperature by approximately 10 K upwards, while an
increase of the evaporation temperature by 5 K decreases the minimum desorption temperature by around
5 K. This known behavior is the same for all sorbents shown here.
0.25
0.15
[mmim][DMP]
LiBr
0.1
AQSOA-Z05
0.05
0.25
0.2
M
0.2
difference in loading q
0.3
RD Silica gel
ED
difference in loading q
0.3
[mmim][DMP]
0.15
LiBr
0.1
RD Silica gel
AQSOA-Z01
0.05
AQSOA-Z01
0
60
65
70
75
desorption temperature in °C
AQSOA-Z05
0
55
80
60
65
70
desorption temperature in °C
75
80
PT
Figure 11: difference in loading at 35 °C adsorption and condensation temperature and evaporation
temperature of 5 °C (left) and 10 °C (right) for various desorption temperatures
CE
3.3 Sorption isotherms
AC
For the investigated liquid sorbents and the Silica gel the heat rejection temperature influences the possible
difference in loading qualitatively in the same way, just inversely and stronger, as the desorption
temperature. However, it has been shown that this is very different for some solid sorbents like the zeotypes
AQSOA-Z01 and Z02. In this chapter the reasons for the above seen benefits and drawbacks will be
discussed on basis of the sorption isotherms. In Figure 12 the phase equilibria are represented for several
different hypothetical materials. Obviously, it is desirable to desorb much refrigerant with a small
temperature increase. On one hand, the minimum desorption temperature
is low, when the derivative
of the boiling pressure to the boiling temperature at constant composition is large (eq. (9)).
|
⇒
(9)
This can easily be seen in a van’t Hoff representation, where low minimum desorption temperatures occur
for steep boiling lines (eq. (9) and T2minI to T2minII in Figure 12 (a)). On the other hand, a high difference in
loading
is represented by a high density of boiling lines in the van’t Hoff representation (Figure 12 (b)),
but can be better evaluated directly in the loading over dew point temperature plot. Thus, the influence of the
13
ACCEPTED MANUSCRIPT
arrangement of boiling lines and isotherms are discussed qualitatively in the following schematics in Figure
12 in van’t Hoff (a and b) as well as in loading over dew point temperature diagrams (c and d).
A steep slope of the boiling lines in the van’t Hoff plot (GI1 to GII1 in Figure 12 (a)) results in isotherms
being horizontally distant from each other (big dew point temperature difference) in the loading over dew
point temperature plot (eq. (10) and GI1 to GII2 in Figure 12 (c)).
|
⇔
|
(10)
|
|
log(p)
(
(a)
C2
material II
GII2
| ⇔
⏟
(
)
log(p)
⏟
|
AN
US
|
⇒
CR
IP
T
Additionally to a big horizontal distance (eq. (10)) a steep slope of the sorption isotherms (eq. (11)) results in
a big vertical distance (big difference in loading) of the isotherms in the loading over dew point temperature
plot (eq. (12) and qIIIG to qIIG in Figure 12 (d) and qA to qII in Figure 12 (c)). The big vertical distance of the
sorption isotherms in the loading over dew point temperature plot is transformed to a high density of boiling
lines in the van’t Hoff representation (eq. (12) and Figure 12 (b)).
GII1
(b)
GA
T0
(c)
material I/II
T1A
A
T0
q
material I material II
T2minI
T2minI
GI1
GII2
qA
Δq
xIIIG < xIIG
xII = xIIIG > xA
T1A/C
T2minII
T2
-1/Tboil
(d)
material II/III
T1A
A
material II material III
T2 T2
GA
GIII
qIIIG
qIIG
qII
GIII
A
xA
T1C2 T2minII T2minI -1/Tboil
AC
qA
T1A/C1
CE
q
E
xII
PT
xA
material III
C
M
A
material II
material I
GI1
ED
E
(12)
)
material II/III
C1
(11)
GII
GI1
T0
T1C1
T1C2
Tdew
T0
T1C
Tdew
Figure 12: schematic of sorption equilibrium; (a) and (b): boiling lines; (c) and (d): isotherms; (a) and (d)
change in slope; (b) and (c) change in density
14
ACCEPTED MANUSCRIPT
Of course the shape of the isotherms of real media is differing from the simple schematic displayed in Figure
12. In Figure 13 the isotherms of the liquid absorbents are shown. The isotherms are plotted in equal
temperature steps and for better understanding the loadings for the temperature quadruples for
evaporator/condenser/ad- or absorber/desorber 5/20/20/40 °C and 5/30/30/70 °C are marked with horizontal
lines. Additionally there are two-sided arrows between the loadings at the specific isotherms at the end of
sorption and at the end of desorption. The LiBr isotherms are ending at low loads at their crystallization
point.
CR
IP
T
For both liquid absorbents it can be seen that the dew point temperature steps (horizontal distance) between
the isotherms do not vary a lot neither with loading nor with dew point temperature. Thus, the minimum
desorption temperature will just be shifted when shifting the overall working temperature level. The
horizontal distance between the sorption isotherms is larger for [mmim][DMP] making it more suitable to
reach low driving temperatures than LiBr (compare Figure 12 (c) and Table 5).
10
1.5
60°C
20°C
20
30
40
dew point temperature in °C
30°C
40°C
50°C
70°C
60°C
1
80°C
90°C
70°C
0.5
CE
AC
0
0
50°C
PT
loading [g/g]
1
0.5
40°C
loading [g/g]
30°C
20°C
ED
1.5
M
AN
US
The slope of the sorption isotherms increases with increasing loading, while obviously the loading at fixed
dew point temperature decreases with increasing desorption temperature. Consequently, the vertical distance
between the sorption isotherms decreases with increasing desorption temperature, resulting in the decrease in
composition change with increasing desorption temperature in Figure 10 and Figure 11. The isotherms´
slopes are larger for [mmim][DMP] in comparison to LiBr at high loadings and the other way round at small
loadings. Thus, the increase in difference in loading with increasing desorption temperature is higher for
[mmim][DMP] than for LiBr starting at the minimum desorption temperature. Due to the fact, that the
change of slope of the LiBr isotherms within the examined region is much less than for [mmim][DMP], the
increase in loading is less affected by the increasing desorption temperature for LiBr; therefore the difference
in loading of LiBr surpasses that of [mmim][DMP] at higher desorption temperatures (Figure 10 and 11).
The fact, that with increasing absorption temperature level the loading at the end of the absorption process
decreases, is more unfavorable for [mmim][DMP] than for LiBr. Consequently, LiBr outperforms
[mmim][DMP] closer to the minimum desorption temperature with increasing absorption and condensation
temperature level (Figure 10, left to right).
80°C
90°C
50
0
0
10
20
30
40
dew point temperature in °C
50
Figure 13: Sorption isotherms of inorganic salt LiBr (left) and IL [mmim][DMP] (right)
The sorption isotherms of RD-Silica gel in Figure 14 behave quite similar to [mmim][DMP] at low dew
point temperatures, but they do not continue to infinite loading (infinite dissolution) but end at a maximum
loading of approximately 0.32. When the dew point is increased further there will be normal condensation
and the loading will not change. The sigmoidal shape is not uncommon for solid sorbents (Brunauer et al.,
1940). It should be stated here, that hysteresis of the isotherms between adsorption and desorption is
neglected for all solid sorbents discussed in the paper at hand.
The isotherms of Silica gel distant to the maximum loading look quite similar to those of the absorbents at
small loadings. The horizontal distance between the isotherms is nearly independent from loading and dew
15
ACCEPTED MANUSCRIPT
point temperature as for the absorbents and their distance is between the values of LiBr and [mmim][DMP],
what is reflected by the minimum desorption temperature, lying in between as well (Figure 11 and Table 5).
The slope of the isotherms of Silica gel is small in comparison to that of the absorbents, what cannot be seen
directly in Figure 14, due to scaling effects. However, there is no significant difference in slope within and
among the sorption isotherms in the examined temperature range for Silica gel in contrast to the absorbents.
The first fact leads to a smaller increase in difference in loading compared to the absorbents, but the second
fact in return results in a nearly constant increase in difference in loading (Figure 11).
10°C
30°C
20°C
loading [g/g]
0.25
40°C
50°C
CR
IP
T
0.3
60°C
0.2
70°C
0.15
80°C
0.1
90°C
0
0
10
AN
US
0.05
20
30
40
dew point temperature in °C
50
Figure 14: Adsorption isotherms of RD-Silica gel
ED
M
As can be seen in Figure 15 the isotherms of the zeotypes AQSOA-Z01 and AQSOA-Z05 basically, of
course, behave like those of the Silica Gel. The horizontal distance of the sorption isotherms is larger for the
zeotype Z01 than for Z05 and for both larger than for the absorbents and the RD-Silica gel. However, there is
an increase in horizontal distance with increasing loading, just slightly for Z05 but noticeable for Z01. The
big horizontal distance explains the lower minimum desorption temperature of Z01 in comparison to Z05 and
of both in comparison to the other sorbents discussed (Figure 10 and Table 5). The increasing horizontal
distance with loading of Z01 explains why the difference in minimum desorption temperature between
AQSOA-Z01 and Z05 becomes larger with decreasing heat rejection temperature, and higher loadings
respectively (Figure 10, right to left).
AC
CE
PT
A more obvious difference to the absorbents and the RD-Silica gel is that at small loading and low dew point
temperature the isotherms almost overlap, which is more distinctive for Z05 (Figure 15, right) than for Z01
(Figure 15, left). With increasing dew point temperature one isotherm after the other separates. Then, the
initially very small slope of the isotherms becomes very steep. In this region a huge difference in loading can
be achieved, larger than for the discussed absorbents and the RD-Silica gel. The achieved change in
difference in loading in this region is bigger for AQSOA-Z05 (Figure 10, left), due to the fact, that at these
conditions the influence of the isotherms’ steeper slope of Z05 is bigger than the influence of the slightly
bigger horizontal distance of the isotherms of Z01. While the slope of the isotherms of Z05 at intermediate
loadings becomes even slightly steeper with increasing dew point temperature, the slope of the isotherms of
Z01 at intermediate loadings decreases with dew point temperature. Thus the achievable difference in
loading between two sorption isotherms in this region is decreasing with desorption temperature, leading to
the sigmoidal shape of the difference in loading of the plot of Z01 in Figure 10.
The slope of the isotherms of Z01 and Z05 is smaller close to the maximum loading and smallest close to the
minimum loading in comparison to the other discussed sorbents.
16
ACCEPTED MANUSCRIPT
0.25
0.25
20°C 30°C
50°C
40°C
40°C
50°C
60°C
0.2
70°C
0.1
0.15
0.1
0.05
0.05
90°C 70°C
80°C
90°C
10
20
30
40
dew point temperature in °C
50
80°C
CR
IP
T
0.15
0
0
30°C
60°C
loading [g/g]
loading [g/g]
0.2
20°C
10°C
10°C
0
0
10
20
30
40
dew point temperature in °C
50
Figure 15: Adsorption isotherms of AQSOA-Z01 (left) and AQSOA-Z05 (right)
AN
US
Therefore, the positive effect of the big change in loading with increasing dew point temperature can be used
only in the very limited working temperature range, where the slope is steep. This is exemplified in Figure 15
with the temperature quadruples 5/20/20/40 and 5/30/30/50. As stated above, the state points of the end of
adsorption and desorption respectively are linked by a two-sided arrow. For the first quadruple with the very
low reject heat temperature of 20°C the difference in loading is comparable for both zeotypes, namely 0.04
for Z01 and a little bit larger, 0.05 for Z05. For the second quadruple the difference in loading is almost 0.14
for Z01 and it is almost naught for Z05. This means that Z01 fits extremely well to these temperatures and
Z05 not at all. Both zeotypes will work below optimal performance for the first quadruple, but they will
work.
M
A high difference in loading in the cooling process is reached for sorbents with S-shaped isotherms when the
following rule-of-thumb is obeyed:
ED
- the evaporator temperature should be above the dew point temperature at which the adsorption isotherm
corresponding to the reject heat temperature exhibits its steepest ascent.
PT
- the condenser temperature should be below the dew point temperature at which the desorption isotherm
exhibits its steepest ascent.
CE
With an evaporator temperature of 5 °C and condensation temperatures between 20 °C and 30 °C this is true
for adsorption isotherms below 30 °C and desorption isotherms above 60 °C for AQSOA-Z01. It is true for
AQSOA-Z05 with adsorption isotherms below 20 °C and desorption isotherms clearly above 50 °C. This
shows that AQSOA-Z05 should be used at temperature levels of around 10 K lower than Z01.
AC
4. Conclusions
In the paper at hand vapor pressure data have been used to proof that the minimum driving temperature of
sorption cooling machines depends on the sorption material used. To do so a physical vapor pressure fit
known from inorganic salt solutions was successfully adapted to organic salt solutions and crystallization
data of lithium bromide were discussed. It has been shown that each absorbent or adsorbent features
individual benefits and drawbacks with regard to achieving lowest desorption temperatures, depending on the
temperature level of heat rejection and evaporation. The respective operational field can be most easily seen
in the usual logarithmic pressure over inverse temperature plot (van’t Hoff representation), in which the
vapor pressure curves should be steep for achieving a low desorption temperature, and dense for achieving a
large difference in loading, at the pressures and temperatures of operation. In the representation of loading
over dew point temperature, the less dense the sorption isotherms in the operational field are the lower the
minimum driving temperature will be. A low density of sorption isotherms also indicates high possible
differences in loading. Additionally, the steeper the sorption isotherms are, the higher the possible difference
17
ACCEPTED MANUSCRIPT
in loading at a given driving temperature level will be. Due to the physical minimum loading of zero for all
sorbents and a maximum loading for all adsorbents, steep sorption isotherms make adsorbents much more
sensitive to ambient temperatures and change in the choice of evaporation temperatures.
When using zeotypes it is possible to achieve a very large difference in loading with small driving
temperature levels if the zeotype is adapted to the working temperatures. However, the examined zeotypes
cannot be used effectively when a maximum heat rejection temperature is surpassed, whatever desorption
temperature is used. The absorbents LiBr and [mmim][DMP] as well as RD-Silica gel are more flexible in
their working conditions due to a more constant slope of the sorption isotherms with dew point temperature
of RD-Silica gel and no maximum loading for the two absorbents, but they do not offer said design option.
AN
US
CR
IP
T
It has been shown that the absorbent [mmim][DMP] can outperform all other examined sorbents at low
evaporation temperature and at heat rejection temperature below 25 °C as far as minimum desorption
temperature and loading capacity is concerned, just surpassed in loading capacity by LiBr at high desorption
temperatures. However, usually the heat rejection temperature is in the range between 20 °C and 40 °C. Then
the zeotype AQSOA-Z05 is the best choice of the analyzed sorbents for heat rejection temperatures of 25 °C
and slightly above, and the zeotype AQSOA-Z01 is preferable at heat rejection temperatures around 30 °C.
Both zeotypes Z05 and Z01 cannot be used at heat rejection temperatures of 30 °C and 40 °C and higher,
respectively. From the five discussed sorbents only the RD-Silica gel and the IL [mmim][DMP] remain for
operation with low evaporation temperatures and heat rejection temperatures higher than 50°C or so, but, of
course, there are more sorbents available.
M
At this point, it should be stated once more that the specific findings of this paper rely on specific
experimental data which, unfortunately, contain unknown uncertainties. Especially the experimental data by
He et al. (2010) can be questioned. This matter deserves further attention. Still, the fundamental relationships
which have been discussed in this paper will prevail, although the minimum desorption temperature of
[mmim][DMP] will rise if the parameter n in equation (5) would be smaller than supported by the data of He
et al. (2010).
ED
5. Outlook
AC
CE
PT
In the paper at hand only the vapor pressure data are used to discuss if the chosen sorbents are appropriate for
operation at low driving temperatures. The hysteresis between ad- and desorption of the zeotypes was
neglected so far. However, especially in low temperature applications the temperature differences which are
necessary to drive the heat and mass transfer are almost in the same order of magnitude as temperature lift
and thrust. Therefore, to select a sorbent to achieve lowest driving temperatures also properties which
influence the heat and mass transfer, such as viscosity (Meyer et al., 2015) of the liquids and thermal
conductivity (Ziegler, 1997) of the solids as well as the coefficient of diffusion for both (Meyer et al., 2015;
Ziegler, 1997) have to be considered. For instance, an IL might have a very high viscosity even at
intermediate salt mass fractions, usually going along with a low diffusion coefficient, making it challenging
to gain a good heat and mass transfer (Meyer et al., 2015). This may result in a remarkable deviation from
the equilibrium state, assumed in the discussions above. This imbalance might even become bigger with
increasing desorption temperature for IL’s (Meyer et al., 2016). Thus, the RD-Silica gel might be the best
choice of the observed sorbents for high heat rejection temperatures, assuming that it is easier to find a
competitive design for good heat and mass transfer for the Silica gel than for the IL.
In the light of the statement above, the smaller the temperature lift is, the more deviation from the above
derived minimum desorption temperatures is to be expected, and even more, the more ineffective the heat
and mass transfer, due to bad design or fluid properties, is. The design of a sorption cooling machine has to
consider all these properties to decide which sorbent is most suitable to achieve lowest driving temperatures.
And, obviously, there are a lot more absorbents and adsorbents to be examined than those five in the paper at
hand.
18
ACCEPTED MANUSCRIPT
Acknowledgements
AC
CE
PT
ED
M
AN
US
CR
IP
T
The knowledge about the ionic liquids was obtained in the research project E-Norm, project number
03ET1129, financed by the German Federal Ministry for Economic Affairs and Energy. Many thanks to the
research group of the E-Norm project. The adsorption data were analyzed in the process of application for
the research project AHA, project number 03ET1412, financed by the German Federal Ministry for
Economic Affairs and Energy.
19
ACCEPTED MANUSCRIPT
AC
CE
PT
ED
M
AN
US
CR
IP
T
References
Aristov, Y. (2007). Novel Materials for Adsorptive Heat Pumping and Storage: Screening and Nano
Tailoring of Sorption Properties. Journal of Chemical Engineering of Japan 40 (13), pp. 1242-1251.
Aristov, Y., Tokarev, M. M., Cacciola, G. & Restuccia, G. (1996). Selective water sorbents for multiple
applications, 1. CaCl2 confined in mesopores of silica gel: Sorption properties. Reaction Kinetics
and Catalysis Letters 59 (2), pp. 325–333.
Boryta, D. A. (1970). Solubilty of Lithium Bromide in Water between -50°C and +100°C. Journal of
Chemical and Engineering Data, 15(1), 142-144.
Brunauer, S., Deming, L. S., Deming, W. E. & Teller, E. (1940). On a Theory of the van der Waals
Adsorption of Gases. Journal of the American Chemical Society, 62(7), 1723-1732.
Buchholz, F. L. & Graham, A. T. (1997). Modern Superabsorbent Polymer Technology. Wiley-VCH.
Feuerecker, G. (1994). Entropieanalyse für Wärmepumpensysteme: Methoden und Stoffdaten. München: TU
München.
Gluesenkamp, K. (2012). Development and Analysis of Micro-Polygeneration Systems and Adsorption
Chillers. Dissertation. University of Maryland, College Park.
Goldsworthy, M. (2014). Measurements of water vapour sorption isotherms for RD silica gel, AQSOA-Z01,
AQSOA-Z02, AQSOA-Z05 and CECA zeolite 3A. Microporous and Mesoporous Materials 196,
pp. 59-67.
He, Z., Zhoa, Z., Zhanga, X. & Feng, H. (2010). Thermodynamic properties of new heat pump working
pairs: 1,3-Dimethylimidazolium dimethylphosphate and water, ethanol and methanol. Fluid Phase
Equilibria 298, pp. 83-91.
Henning, H.-M. (2007). Solar assisted air conditioning of buildings - an overview. Applied Thermal
Engineering 27, pp. 1734-1749.
Henninger, S., Schmidt, F. & Henning, H.-M. (2010). Water adsorption characteristics of novel materials for
heat transformation applications. Applied Thermal Engineering 30, pp. 1692-1702.
Hüttig, G. F. & Reuscher, F. (1924, July 7). Studien zur Chemie des Lithiums - I. Über die Hydrate des
Lithiumschlorids und Lithiumbromids. Journal of Inorganic and General Chemistry, 137(1), 155180.
Kim, D. & Infante Ferreira, C. (2008). Solar refrigeration options - a state-of-the-art review. International
Journal of Refrigeration 31, pp. 3-15.
Lamb, P. & Ziegler, F. (1996). Comparison of different liquid and solid sorption systems with respect to low
temperature driving heat. Proceedings of International Absorption Heat Pump Conference, (pp.
269-276). Montreal.
Liang, S., Chen, W., Cheng, K., Guo, Y. & Gui, X. (2011). The Latent Application of Ionic Liquids in
Absorption Refrigeration. In S. Handy, Applications of Ionic Liquids in Science and Technology
(pp. 467 - 494). Croatia: InTech.
Löwer, H. (1960). Thermodynamische und physikalische Eigenschaften der wässrigen LithiumbromidLösung. Dissertation. Technische Hochschule Karlsruhe, Karlsruhe.
Meyer, T., Kühn, R., Ricart, C., Zegenhagen, T. & Ziegler, F. (2015). Simulation of an absorption
refrigerator working with ionic liquids and natural refrigerants. Proceedings of the 24th IIR
International Congress of Refrigeration. Yokohama, Japan.
Meyer, T., Winker, M., Bergemann, S., Kühn, R., Ricart, C. & Ziegler, F. (2016). Ergebnisse von
Experimenten mit einer Absorptionskälteanlage mit ionischer Flüssigkeit und Ethanol. Kassel:
DKV.
Othmer, D. & Fröhlich, G. (1960). Correlating vapor pressures and heats of solution for the ammonium
nitrate-water system: An enthalpy-concentration diagram. AIChE Journal 6 (2), pp. 210-214.
Saha, B., Koyama, S., Kashiwagi, T., Akisawa, A., Ng, K. & Chua, H. (2003). Waste heat driven dual-mode,
multi-stage, multi-bed regenerative adsorption system. International Journal of Refrigeration 26,
pp. 749-757.
Seiler, M., Schneider, M.-C., Kühn, A. & Ziegler, F. (2010). New high-performance working pairs for
absorption chillers and heat pumps. Proceedings of the International Conference: Innovative
Materials for Processes in Energy Systems, (pp. 350-360). Singapore.
Zegenhagen, T., Kühn, R., Meyer, T., Ricart, C. & Ziegler, F. (2015). Investigation of a Liquid Disiccant
System for Air Dehumidification working with an Ionic Liquid in a Two-Stage Refrigeration
System for Cold Stores. Proceedings of the 24th IIR International Congress of Refrigeration.
Yokohama, Japan.
Zhao, J., Jiang, X.-C., Li, C.-X. & Wang, Z.-H. (2006). Vapor pressure measurement for binary and ternary
systems containing a phosphoric ionic liquid. Fluid Phase Equilibria 247, pp. 190-198.
Ziegler, F. (1997). DKV Forschungsberichte: Sorptionswärmepumpen (Vol. 57). Stuttgart: DKV.
20
Документ
Категория
Без категории
Просмотров
0
Размер файла
1 889 Кб
Теги
ijrefrig, 019, 2018
1/--страниц
Пожаловаться на содержимое документа