SCHOOL OF ENGINEERING AND INFORMATION TECHNOLOGY ENG470 ENGINEERING HONOURS THESIS FINAL REPORT SEPARATION OF SOLVENT FROM MICROALGAL HYDROCARBON USING NANOFILTRATION Reported by: King Zheng Lim SUPERVISORS PROFESSOR PARISA ARABZADEH BAHRI - PROFESSOR OF ENGINEERING, SCHOOL OF ENGINEERING AND INFORMATION TECHNOLOGY DR. NAVID MOHEIMANI - SENIOR LECTURER A report submitted to the School of Engineering and Energy, Murdoch University in partial fulfilment of the requirements for the unit ENG470 Engineering Honours Thesis.
96
Embed
ENG470 Engineering Honours Thesis Final Report · SCHOOL OF ENGINEERING AND INFORMATION TECHNOLOGY ENG470 ENGINEERING HONOURS THESIS FINAL REPORT SEPARATION OF SOLVENT FROM MICROALGAL
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
SCHOOL OF ENGINEERING AND INFORMATION
TECHNOLOGY
ENG470 ENGINEERING
HONOURS THESIS FINAL
REPORT SEPARATION OF SOLVENT FROM
MICROALGAL HYDROCARBON USING
NANOFILTRATION
Reported by: King Zheng Lim
SUPERVISORS
PROFESSOR PARISA ARABZADEH BAHRI - PROFESSOR OF
ENGINEERING, SCHOOL OF ENGINEERING AND INFORMATION
TECHNOLOGY
DR. NAVID MOHEIMANI - SENIOR LECTURER
A report submitted to the School of Engineering and Energy, Murdoch University in partial fulfilment of the requirements for the unit ENG470 Engineering Honours Thesis.
0 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Executive Summary
The need for searching an alternative technology to separate solvent efficiently
from the post-extraction process in the algae fuel production process has been long
researched for, and little to no convincing findings were found to rectify the current energy
crisis. This report aims to evaluate the viability of implementing nanofiltration technology
that could replace the use of a distillation column in the post-extraction process. Aspen
Plus was used to assess the thermodynamic feasibility of utilising chemical process unit
operations. This includes the following: investigation of the effect of thermodynamic
property methods to generate a more realistic separation process based on the nonideality
of the feed mixture, optimization of the simulation via sensitivity analysis, and an overall
energy balance to determine its sustainability based on the calorific value of the
hydrocarbon extracted from an algae culture. Nanofiltration experiments were carried out
to establish the applicability of the membrane purchased from Sterlitech and possibly fill a
current void in research for utilising the Duracid membrane in a heptane solution. The
experiments covered: the effect of different contact times with heptane, the effect of
pressure and feed concentration variance. A stirred cell was used to facilitate the
experiment, and several parameters were done to determine the characteristics of the
membrane, which included permeating de-ionised water, heptane, and squalene-heptane.
Results showed that prolonged contact times with heptane worsen the permeating
performance of the membrane over time, and a maximum of 6% rejection value was
attained when using Duracid membrane. Higher operating pressure and lower feed
concentration also enhanced the permeate flux. Possible explanation for such occurrence
includes the nanofiltration driving force, membrane polarity difference to the solvent, and
membrane swelling. Although GCMS showed a little rejection value for retaining squalene
in heptane solution, the finding is significant that could prove solvent separation via
nanofiltration is possible and future work is needed to improve the outcome. Alternative
separation technology and solvent resistant nanofiltration membrane had been proposed,
and that could serve as another starting point for an efficient separation process.
1 | P a g e ENG 470 ENGINEERING HONOURS THESIS
List of Nomenclature and Abbreviations
AP Aspen Plus
DC Distillation column
DI De-ionised
FS Flash separator
GCMS Gas chromatography mass spectrometry
kJ kilojoules
Lmh Litre/(m2.hour)
MRDF Molar ratio of distillate flow rate to feed flow rate
2.2 Literature review .............................................................................................................................. 13
3.2 Literature review .............................................................................................................................. 35
Figure 18: Diagram of experimental set-up .............................................................................................. 46
Figure 19: Scatter plot of cumulative permeate volume in DI water for Set 1 (Blue), Set 2
(Red) and Set 3 (Green) at 20 bar with its respective trend line and its R2 value .................... 49
Figure 20: Scatter plot of permeate flux in DI water for Set 1 (Blue), Set 2 (Red) and Set 3
(Green) at 20 bar .................................................................................................................................................. 50
5 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Figure 21: Scatter plot of cumulative permeate volume in heptane for different soaking
times at 30 bar – Set 1 (30 min), Set 2 (60 minutes) and Set 3 (90 minutes) with its
respective trend line and its R2 value .......................................................................................................... 51
Figure 22: Scatter plot of permeate flux in heptane for different soaking times at 30 bar –
Set 1 (30 min), Set 2 (60 minutes) and Set 3 (90 minutes) ................................................................ 52
Figure 23: Scatter plot of permeate flux in heptane for different soaking times at 30 bar ... 54
Figure 24: Scatter plot of cumulative permeate volume for different squalene
concentrations at 30 bar ................................................................................................................................... 56
Figure 25: Scatter plot of permeate flux for different squalene concentrations at 30 bar .... 57
Figure 26: Scatter plot of cumulative permeate volume for different squalene
concentrations at 50 bar ................................................................................................................................... 59
Figure 27: Scatter plot of permeate flux for different squalene concentrations at 50 bar .... 59
Figure 28: Scatter plot of GCMS results for different squalene concentration at 30 bar ....... 61
Figure 29: Scatter plot of GCMS results for different squalene concentration at 50 bar ....... 62
Figure 30: A spreadsheet of feed composition calculation ................................................................. 77
Figure 31: Input requirements for ‘FLASH’ column under Specification tab ............................. 78
Figure 32: Input requirement for ‘RADFRAC’ column under Configuration tab ....................... 78
Figure 33: Feed and component input requirement for ‘RADFRAC’ column under Feed
Figure 34: Input requirement for ‘RADFRAC’ column under Streams tab................................... 79
Figure 35: Setting Change on Thermodynamic Property .................................................................... 80
Figure 36: Sensitivity analysis input requirement for flash separator temperature in
‘FLASH’ model ....................................................................................................................................................... 81
Figure 37: Variable definition and input requirement for flash separator sensitivity analysis
outputs in ‘FLASH’ model ................................................................................................................................. 81
Figure 38: Sensitivity analysis input requirement for ‘RADFRAC’ reflux ratio .......................... 82
Figure 39: Variable definition and input requirement for ‘’RADFRAC’ sensitivity analysis
Figure 42: Sensitivity analysis Input requirement for 'FLASH2' reactor temperature in
‘FLASH’ model ....................................................................................................................................................... 84
Figure 43: Variable definition and input requirement for ‘FLASH2’ tank sensitivity analysis
outputs in ‘FLASH’ model ................................................................................................................................. 84
Figure 44: Sensitivity analysis Input requirement for 'FLASH2' reactor temperature in
As can be seen in Figure 3, initially, there was no heptane detected in the top
stream until there was a sharp increase in flow rate after the flash temperature reached
96 °C. Heptane flow rate reached a steady flow rate of 99.9 kg/hr at a temperature of
100 °C and squalane did not vaporise into the top stream before reaching a temperature of
100 °C (indicated by the dashed line). Therefore, no further changes had been
implemented as the process was at its optimum condition. This condition will be used for
further modelling.
2.3.4.2 ‘RADFRAC’ model
Similar to optimising ‘FLASH’ temperature, two parameters, RR and its MRDF were
used for SA to optimise heptane separation. A ‘Model Analysis Tools’ under the label of ‘RR’
was created and the input requirements can be referred as Appendix B.4. In this test, MRDF
was held constant at 0.5 and the column temperature at 1 bar. Heptane flow rate in the top
stream ‘4TOP’, labelled as ‘TOPHEPT’, was monitored in accordance with the changing
reflux ratio from ‘RADFRAC’ column, while RR varied from 0.5 to 5. Figure 4 shows the
graph of the SA.
Figure 4: Heptane flow rate in top stream as a function of ‘RADFRAC’ reflux ratio (constant molar ratio of distillate to feed flow rate at 0.5 and pressure at 1 bar)
Sensitivity Results Curve
VARY 1 RADFRAC COL-SPEC MOLE-RR
TO
PH
EP
T K
G/H
R
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.048.0
48.5
49.0
49.5
50.0
50.5
51.0
51.5
52.0
TOPHEPT KG/HR
23 | P a g e ENG 470 ENGINEERING HONOURS THESIS
From the graph above, it was clear that the RR for ‘RADFRAC’ column did not affect
the phase equilibrium for the separation process. The reason for such occurrence was not
clear except for the fact that MRDF was held constant at 0.5. However, it was believed that
the amount of squalane presented in the mixed stream was too dilute to pose any effect to
the heptane separation. As a result, RR was set to 1.0 for the course of the simulation.
Next, a ‘Model Analysis Tools’ under the label of ‘MR’ was created for the MRDF
(Appendix B.4). In this test, MRDF was manipulated from 0.10 to 0.99, and the heptane
flow rate in the top stream ‘4TOP’, labelled as ‘TOPHEPT’, was recorded. Figure 5 shows
the graph of the SA.
Figure 5: Heptane flow rate in top stream as a function of ‘RADFRAC’ MRDF (constant reflux ratio of 1 and pressure at 1 bar)
SA shows that as MRDF increased, the amount of heptane flow in the top stream
increased accordingly. This was expected because the molar ratio determines the desired
amount of light key component to being recovered. Since heptane was chosen to be the
light key component, the higher the MRDF, the greater the amount will be recovered.
Hence, the value of 0.99 was chosen for the MRDF in this simulation.
‘RADFRAC’ model, nearly 100% of squalane from ‘5BOTTOM’ stream had been successfully
separated to the bottom stream ‘7SQUAL’, which is highlighted in the blue box.
2.3.6 SA on heptane recovery for the final model
Having some trace of heptane still left in ‘6TOP’ and ‘7SQUAL’ stream, another SA
was carried out for both models. Similar SA procedure was taken, shown in Appendix B.5.
2.3.6.1 ‘FLASH’ model
The condition for performing the SA around the second FS was as follows: tank
pressure at constant and the tank temperature varied from 90 to 200 °C. The responding
variables, heptane flow rate in the top stream ‘6TOP’, labelled as ‘TOPHEPT’. Squalane flow
rate in the bottom stream ‘7BOTTOM’, labelled as ‘BOTSQUAL' was monitored to determine
the optimum reactor temperature (Figure 8). Note that the feed basis for both heptane and
squalane in the mixed stream was 0.468 kg/hr and 0.1 kg/hr, respectively.
Figure 8: Heptane flow rate in top stream and squalane flow rate in bottom stream as a function of ‘FLASH2’ temperature (‘FLASH’ model, constant pressure 1 bar)
Similar to ‘FLASH’ model optimisation, the condition for SA was as follows: tank
pressure held at constant 1 bar and the tank temperature varied from 90 to 200 °C. The
responding variables, heptane flow rate in the top stream ‘6TOP’, labelled as ‘TOPHEPT’.
Squalane flow rate in the bottom stream ‘7BOTTOM’, labelled as ‘BOTSQUAL' was
monitored to determine the optimum reactor temperature (Figure 9). Note that the feed
basis for both heptane and squalane in the mixed stream was 0.698 kg/hr and 0.1 kg/hr,
respectively.
Figure 9: Heptane flow rate in top stream and squalane flow rate in bottom stream as a function of ‘FLASH’ temperature (‘RADFRAC’ model, constant pressure 1 bar)
A similar trend can be seen when compared to Figure 8. Heptane formed a plateau
at 0.670 kg/hr mark after reaching 145 °C, and squalene in the bottom stream, ‘7SQUAL’,
stayed constant at 0.1 kg/hr until it reached 145 °C. Hence, the FS, ‘FLASH’, used in
‘RADFRAC’ model should be at 145 °C. Table 8 below shows the stream table from using
the new operational parameters. (‘FLASH1’ temperature = 100 °C, ‘FLASH2’ temperature =
1. Solvent absorption – The polymer surface absorbs the solvent in.
2. Solvent penetration – The polymer surface being penetrated by the solution, which
the solvent molecules first occupy the free volume and then diffuse into the
polymer.
43 | P a g e ENG 470 ENGINEERING HONOURS THESIS
3. Polymer expansion – The polymer structure expands as a result from the trapped
solution in the pores, swelling the network of the polymer chains.
As mentioned prior, membrane swelling can influence the transport mechanism of
different types of membrane, namely dense and porous membrane. Firstly, the expansion
of free volume in a dense membrane could cause the membrane pores to increase, allowing
larger molecules to pass through, thus, increasing membrane permeability, decreasing
selectivity and lowering rejection (Ebert 2005; Farid 2010). Secondly, the compaction of
membrane pores in a porous membrane could lead to an increase in selectivity and
decrease in permeability, results in higher rejections (Ebert 2005; Farid 2010).
Study has found that membrane swelling is a good indication for permeation as the
so-called channels from the membrane are formed, thus increasing solvent flux in organic
solvents such as n-alkanes, i-alkanes and cyclic compounds, through a dense PDMS
composite NF membrane (Robinson et al. 2004). However, it was argued that swelling for
porous membrane could cause the membrane to be ‘less open’, resulted in higher rejection
(Robinson et al. 2005).
3.2.4.5 HP4750 Stirred Cell
The HP4750 Stirred Cell used in this experiment has a high durability against
pressure due to its stainless steel (316L) cell body construction that can withstand a
maximum rating of 1000 psig (6900 kPa). With this feature, this stirred cell is capable of
performing a wide selection of membrane separation and stimulating the flow dynamics of
microfiltration, ultrafiltration, reverse osmosis and nanofiltration. Moreover, this stirred
cell is also chemically resistant to a wide range of liquid and gas chemicals, making it an
ideal choice to filter both aqueous and non-aqueous solutions.
This stirred cell is considered suitable for simulating the flow dynamics of NF
systems, in particular, a dead-end filtration, where the feed flow is perpendicular to the
membrane, resulting in the retained particles accumulating on the surface of the
membrane (Nobel and Terry 2004).
44 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Its features and technical specifications are provided in Appendix C.
Figure 17: Image of HP4750 Stirred Cell (Sterlitech 2015)
3.3 Materials and methods
3.3.1 Chemicals
The solvent used for dissolving squalene was n-heptane, which was purchased
from Rowe Scientific Pty Ltd, Perth. The n-heptane purchased was a technical grade.
Squalene was used to facilitate the concentration variation for the nanofiltration
performance, and it was purchased from VWR International Pty Ltd. Table 11 shows the
essential physical properties of the chemicals used in this experiment.
Table 11: Physico-chemical properties of n-heptane and squalene
Chemicals n-heptane Squalene
Chemical structure
Molecular weight 100.21 g/mol 410.72 g/mol
Density 679.5 kg/m3 854.0 – 856.0 kg/m3
Flash point -4.0 °C 200.0 °C
Boiling point 98.4 °C 470.3 °C
3.3.2 Nanofiltration membrane
The NF membrane used for the filtration experiment was GE Osmonics KH Duracid
Series TFC NF Membrane, and it was purchased from Sterlitech Corp. Table 12 shows the
technical specification of the studied membrane:
45 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Table 12: Technical specifications for GE Osmonics Duracid NF membrane (Sterlitech 2015)
Series Feed Type pH
range Flux Pressure
MgSO4 Rejection
MWCO
Duracid Industrial/ Commercial
Acid Purification,
Mineral Concentration
0 – 9
10 – 19 Gfda/
17 – 32.3 Lmhb
225 psi / 1551 kPa / 15.51 bar
98.0 % ~150 –
200 Daltons
a Gfd – gallons/ft2/day; b Lmh – litre/m2/hour;
The purchased membrane was a flat sheet measuring 30.5 cm by 30.5 cm, and it
was cut into a circular disc of 46 cm diameter using a print-out that had a circle of diameter
46 cm as a guide. For each experiment trial, a new membrane was used, and the membrane
was immersed in deionized (DI) water for at least 24 hours before any experimental work.
3.3.3 Filtration experiment set-up and procedure
Dead-end filtration experiments were performed with a stirred cell; model
StelitechTM HP4750 Stirred Cell. The stirred cell was pressurised by industrial grade
nitrogen gas and the maximum operating pressure for this cell was Pa
(1000 psi or 69 bar). The effective membrane area of the stirred cell was m2,
allowing an active membrane diameter of 4.31 cm. This stirred cell had a processing
volume of 300 mL, while its liquid hold-up volume was 1 mL.
For the filtering process, the applied pressure in this experiment was 20 bar, 30 bar
and 50 bar, respectively. Meanwhile, the pressure on the permeate side was approximated
to be at atmospheric pressure under all conditions as the permeate tube was being held in
the atmosphere.
Prior to filtration, the compaction process mentioned in the previous section was
carried out. All experiments were performed in batch mode, whereby the feed solution was
charged through the membrane in the cell, leaving the larger product on the membrane
surface. The permeate samples flowed out from the bottom of the cell and were collected at
a 1-minute interval for every trial over 1 to 1.5 hours, and the cumulative volume was
measured using a measuring cylinder.
46 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Filtration experiment was done using DI water and a mixture of organic solution as
the feed media. The organic solution comprised of squalene and heptane was used as the
solvent. For the purpose of evaluating the performance of the nanofiltration in organic
solution, different concentration of squalene was tested at 1.0 %, 3.0 % and 5.0 % v/v. The
concentration of the squalene samples was based on volume percentage of the squalene
content and the volume percentage is defined as Equation 3.
Equation 3: Volume percentage determination
Note: The solute throughout this experiment was squalene and the solvent used to make
up the solution was heptane.
The experiments were conducted in pairs to check the replication of the membrane
performance. All experiments were conducted at ambient temperature of . A
schematic view of the experimental set-up is shown in Figure 18.
Figure 18: Diagram of experimental set-up
47 | P a g e ENG 470 ENGINEERING HONOURS THESIS
3.3.4 HP4750 Stirred Cell assembly, maintenance and operation
Before assembling the stirred cell, all of the necessary components should be
verified and present, and the complete set of the stirred cell is shown (Appendix D). All the
assembling procedures and precautions are described in Appendix E.
3.3.5 Chemical analysis
To investigate the solute rejection in the nanofiltration process, the feed and
separation permeate samples were analysed using Shimadzu Gas Chromatography system
that comprised of GCMS-QP2010S gas chromatograph-mass spectrometer (GC-MS), GC-
2010 gas chromatograph and AOC-20i+S auto-injector and an auto-sampler. For each
chemical analysis, an approximate 1 mL of sample was collected in a 4 mL screw top glass
vial. The equipment and method conditions used for running the GCMS analysis are shown
in Table 13. Appendix F shows the detail method parameters for running the GCMS
analysis.
Table 13: List of equipment and method condition used for GCMS
Parameters Description
Column BP-5, 30 m long, 0.25 µm thickness and 0.25 mm
internal diameter
Carrier gas Ultra-high purity helium gas
Injector port temperature 300 °C
Column oven temperature Kept at 220 °C for 1 minute, ramped to 260 °C at
the rate 2.0 °C min-1
Total run time 35 minutes
Mass spectra range 45.0 to 1000 (mass to charge) m/z
3.3.6 Membrane permeance analysis
The performance of the nanofiltration membrane was examined after the filtration
experiment. The cumulative weight of permeate recorded from each experiment was used
48 | P a g e ENG 470 ENGINEERING HONOURS THESIS
to determine the filtration system efficiency such as the permeate flux and the rejection
value. The permeate flux, (L/m2.hr or Lmh) was obtained using Equation 3.
Equation 4: Expression of permeate flux
where is the cumulative volume difference (L), is the time difference (min), and is
the active membrane area (m2).
While the rejection value, could be obtained with the following equation:
Equation 5: Expression of rejection value
where and is the permeate concentration and feed concentration (% vol),
respectively.
49 | P a g e ENG 470 ENGINEERING HONOURS THESIS
3.4 Nanofiltration performance – Results and Discussion
Prior to solvent separation using nanofiltration, two tests were carried out to verify
its capability to permeate different types of solvents, permeating polar and non-polar
solvent. It was necessary to determine whether the membrane purchased from
Sterlitech Co. was applicable for the purpose of this experiment and to investigate the
factors that could affect the membrane performance using different solvents.
3.4.1 Permeating DI water
The filter membrane was tested using de-ionised water as feed media. This was
done to verify if filtration was possible from using the pre-soaking method. After pre-
soaking the filter overnight, the stirred cell was fitted with the filter and was charged with
100mL of deionised (DI) water at a constant pressure of 20 bar by N2 gas. This set-up was
repeated three times to evaluate the reliability of the results. The following Figure 19 and
Figure 20 show the cumulative permeate volume for three replicates and its flux over a
period of 10 minutes, respectively.
Figure 19: Scatter plot of cumulative permeate volume in DI water for Set 1 (Blue), Set 2 (Red) and Set 3 (Green) at 20 bar with its respective trend line and its R2 value
y = 0.4109x + 0.0909 R² = 0.9982
y = 0.4127x + 0.0455 R² = 0.9994
y = 0.4209x + 0.0318 R² = 0.9991
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
4.5
0 2 4 6 8 10 12
Cu
mu
lati
ve V
olu
me
(m
L)
TIme (min)
Permeate Volume Set 1, 2 & 3 - DI Water
Set 1
Set 2
Set 3
Linear (Set 1)
Linear (Set 2)
Linear (Set 3)
50 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Figure 20: Scatter plot of permeate flux in DI water for Set 1 (Blue), Set 2 (Red) and Set 3 (Green) at 20 bar
According to Figure 19, it can be seen that the cumulative permeate volume was
relatively consistent for all of the filtration performed. An average R-square value of 0.998
was achieved when permeating DI water. Likewise, from Figure 20, the permeate flux
stabilised after filtering for 4 minutes and yielded an average permeate flux of 17.33 Lmh.
At the beginning of the experiment the permeate flux in Set 1 was slightly higher than the
other sets. However, it was clear that the filter could retain a stable and consistent
permeate flux over time. This was to be expected, as the molecular weight of water
(18.01 Dalton) is smaller than the MWCO of the membrane (200 Daltons), which suggested
that water could permeate through the nanofilter with consistent results. Thus, this
resulted in a relatively higher permeate flux (Kim, Jegal and Lee 2002). Such linear
relationship between the permeate volume and the time taken suggested that the pre-
soaking method was necessary as it removed the conditioning agent on the surface of the
filter upon manufacturing the filter (Vandezande, Gevers and Vankelecom 2008).
It is worth mentioning that there was little to no literature review to compare the
permeance results obtained for the Duracid membrane. However, the results shown by
permeating DI water through the filter suggested that there was no significant impact from
0.00
5.00
10.00
15.00
20.00
25.00
30.00
0 2 4 6 8 10 12
Pe
rme
ate
Flo
w R
ate
(Lm
h)
TIme (min)
Permeate Flux Set 1, 2 & 3 - DI Water
Set 1
Set 2
Set 3
51 | P a g e ENG 470 ENGINEERING HONOURS THESIS
membrane-solvent interaction, as the filter did not change in physical appearance upon
contacting with DI water. According to the technical specification for the membrane
purchased from Sterlitech, refer to Table 12, the permeate flux fell within the expected
range, which implied that the filter was functioning properly.
3.4.2 Permeating heptane
The filter was prepared using the same method as described in Section 3.3.3. Since
three filters were pre-soaked in heptane, only one filter can be used at a time. Filters were
tested by its particular total soaking time in heptane, which was 30, 60 and 90 minutes. It
was noted that when the stirred cell was charged at 20 bar, there was no solution
permeating from the stirred cell outlet. Hence, the stirred cell was fitted with the filter and
was charged with 100mL of heptane solution at a constant pressure of 30 bar by N2 gas.
The permeate performance of the filters was recorded accordingly, and is shown in Figure
21 and Figure 22.
Figure 21: Scatter plot of cumulative permeate volume in heptane for different soaking times at 30 bar – Set 1 (30 min), Set 2 (60 minutes) and Set 3 (90 minutes) with its respective trend line and its R2 value
y = 0.0864x + 0.1109 R² = 0.9979
y = 0.0737x + 0.1615 R² = 0.9921
y = 0.059x + 0.1904 R² = 0.9874
0.0
0.5
1.0
1.5
2.0
2.5
3.0
0 5 10 15 20 25 30
Cu
mu
lati
ve V
olu
me
(m
L)
Time (min)
Permeate Volume Set 1 & 2 - Heptane
Set 1 (30 min)
Set 2 (60 min)
Set 3 (90 min)
52 | P a g e ENG 470 ENGINEERING HONOURS THESIS
As shown in Figure 21, the total cumulative permeate volume decreased as the
soaking time in heptane solution increased. It was noted that the longer the soaking time in
heptane, the harder it is for the heptane to permeate through. Observed from Set 3, the line
of best line yielded a relatively weak measure of correlation compared to other sets. Its
final cumulative permeate volume was 1.9 mL over the period of 20 minutes, which was
about 70.3 % of the final cumulative permeate volume for Set 1 (2.7 mL). Noted from the
R-square values, it can be seen that the linear correlation deviated over prolonged contact
time with heptane. Such trend was not expected, as preliminary experimentation did not
take into account of the different effect after soaking in heptane solution over a different
period of time.
Figure 22: Scatter plot of permeate flux in heptane for different soaking times at 30 bar – Set 1 (30 min), Set 2 (60 minutes) and Set 3 (90 minutes)
As a result from a decreased in cumulative permeate volume as seen from Figure
21, the final permeate flux had decreased over the increasing soaking time from 3.7 Lmh
(Set 1) to 2.6 Lmh (Set 3). It can be seen that due to the different soaking time in heptane
solution, the permeate flux for using the filter was not only consistent, but also displaying a
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
0 5 10 15 20 25 30
Pe
rme
ate
Flo
w R
ate
(Lm
h)
TIme (min)
Permeate Flux Set 1 & 2 - Heptane
Set 1 (30 min)
Set 2 (60 min)
Set 3 (90 min)
53 | P a g e ENG 470 ENGINEERING HONOURS THESIS
decreasing trend. Moreover, the permeate flux for filtering heptane solution was noticeably
lower (3.17 ± 0.55 Lmh) when compared to filtering DI water (17.33 ± 0.11 Lmh).
Table 14: Physical properties of solvents used (Engineeringtoolbox 2016)
Some reasons could explain such phenomena in the filtration trials. Upon
investigation, it was found that the differences in the physical properties of solvents, as
shown in Table 14, could affect the results of the permeate flux (Kim, Jegal and Lee 2002).
The physical properties included molecular weight, dielectric constant, and viscosity.
According to their studies, despite viscosity of water being higher than heptane, it was
observed that the polar solvent (water) had resulted in a higher flux using polyamide TFC
membrane, while the nonpolar solvent (heptane) resulted in lower flux
(Kim, Jegal and Lee 2002). This was mainly due to heptane having significantly lower
dielectric constant, thus its high hydrophobicity. Dielectric constant is a measure of a
substance’s ability to insulate electric charges from each other. It measures the polarity of
the material, and that is, the higher the dielectric constant, the higher polarity the solvent,
the greater the ability to stabilise charges (Hardinger 2016). Hence, due to such factor, the
permeate flux from permeating heptane resulted in lower flux compared to permeating
water.
According to Van der Bruggen et al. (2002), non-polar solvent such as n-heptane
affected negatively the solvent flux through hydrophilic membranes. Since the membrane
manufacturer did not disclose any information in regards to the materials used for
fabricating Duracid membrane, it could be postulated that Duracid membrane is a
hydrophilic membrane due to such low permeate flux in heptane. A similar study
suggested that the polarity and the hydrophobicity of membrane surface play a crucial role
in solvent permeation (Jimenez-Solomon et al. 2013). According to these studies, the
Solvents Molecular Weight (g mol-
1)
Dielectric Constant (at
20°C)
Viscosity (cP)
(at 27°C)
Water 18.02 80.1 0.890
Heptane 100.21 1.9 0.376
54 | P a g e ENG 470 ENGINEERING HONOURS THESIS
support membrane of TFC membrane had a significant impact on the physicochemical
properties of different solvents, particularly with n-hexane, a non-polar solvent.
The filter manufacturer had prescribed the method of soaking the filter in DI water
only. However, upon consulting with the Process Development Product Manager from
Sterlitech, it was recommended that the Duracid membrane was the best membrane they
could offer for this experiment. It is also worth mentioning that due to insufficient data for
the membrane’s compatibility with heptane from Sterlitech, the Product Manager
suggested that Duracid membrane could be tested for its applicability to the nonpolar
organic solution and investigate its filtration outcome, thus, the usage of heptane as the
solvent.
A further investigation on the effect of a longer soaking time in heptane was carried
out. The filter membrane was pre-soaked in DI water for overnight, followed by soaking
the membrane in heptane for 100 minutes and 120 minutes. A chart of its permeate flux
was generated along with the previous finding to illustrate the effect of soaking time
(Figure 23).
Figure 23: Scatter plot of permeate flux in heptane for different soaking times at 30 bar
0.00
2.00
4.00
6.00
8.00
10.00
12.00
0 5 10 15 20 25 30 35
Pe
rme
ate
flo
wra
te (
Lmh
)
Time (min)
Permeate Flux - Heptane
30 min
60 min
90 min
100 min
120 min
55 | P a g e ENG 470 ENGINEERING HONOURS THESIS
It was found that there was a decreasing trend of the permeate flux as a function of
soaking time in heptane. Based on the observation, it can be seen that the membrane that
had been immersed in heptane for 120 minutes resulted in the highest permeate flux. Such
occurrence was not expected, as the previous trial experiments did not match with the
findings that suggested the longer the soaking time, the more resistive the membrane
could be, that is lower permeate flux.
It was hypothesised that the prolonged contact time with heptane after per-soaking
in DI water had negatively affected the polymeric arrangement of the membrane surface.
However, since there was limited information in regards to the characteristics of the
membrane purchased such as the surface-coated material and the monomers used for
fabricating the membrane, it was hard to correlate the solvent performance with the
membrane characteristics (Kim, Jegal and Lee 2002). It was unclear as to why the filter
membrane behaved in such a way that the permeate flux varied so much under the slight
variation of soaking time. Nonetheless, it could be postulated that the membrane was
‘damaged’ when it was immersed for too long.
Literature review found that the external surface characteristics of PA TFC
membranes are commonly known for its hydrophilicity, meaning PA TFC membranes are
commonly used in aqueous applications. In order to increase its permeability of nonpolar
solvents, its surface properties needed to be modified via surface chemistry to increase its
hydrophobicity (Jimenez Solomon, Bhole and Livingston 2013).
3.4.3 Permeating a binary solution of squalene and heptane
Having studied the effect of permeating heptane using the Duracid membrane, the
experiment further investigated the effect of permeating squalene in heptane solution to
examine the potential for solvent separation. Similar to permeating heptane, preliminary
tests showed that no permeate was observed when 1%, 3% and 5%v/v squalene solution
(heptane as solvent) was charged in the stirred cell at a pressure of 20 bar. Therefore, the
56 | P a g e ENG 470 ENGINEERING HONOURS THESIS
applied pressure was increased from 20 bar to 30 and 50 bar. A set with no squalene in the
feed solution, labelled as ‘Control’, was included to serve as a control variable.
3.4.3.1 Permeating squalene-heptane solution at 30 bar
Figure 24 and Figure 25 show the filtration performance of the filter membranes
operated at 30 bar. Its respective soaking time in heptane had been labelled accordingly.
Figure 24: Scatter plot of cumulative permeate volume for different squalene concentrations at 30 bar
It can be seen that 3% v/v squalene yielded the highest cumulated permeate
volume among all the experiment sets, as shown in Figure 24. The final permeate volume
from permeating 3% v/v squalene was almost three times as much as that of the control
(8.8 mL from 3% v/v compared with 3.4 mL from control). Such significant difference
could be noted from the pre-soaking time in heptane solution. As indicated, the membrane
used for permeating 3% v/v had soaked for the longest time among them all, almost 6
hours of immersion in heptane solution. It was clear that the filter membrane was soaked
in heptane for too long that the membrane ‘damaged’ and allowed easier permeation.
y = -1E-04x2 + 0.0337x + 0.2323 R² = 0.9894
y = -0.0004x2 + 0.1267x + 0.1793 R² = 0.9995
y = -6E-05x2 + 0.0215x + 0.1196 R² = 0.9909
y = -0.0002x2 + 0.052x + 0.2206 R² = 0.9952
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
0 20 40 60 80
Cu
mu
lati
ve v
olu
me
(m
L)
Time (min)
Permeate Volume
1% v/v (3 hr 45 min)
3% v/v (5 hr 50 min)
5% v/v (3 hr 30 min)
Control (heptane only, 1 hr40 min)
57 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Figure 25: Scatter plot of permeate flux for different squalene concentrations at 30 bar
From Figure 25, it can be seen that the permeate flux for permeating 3% v/v was
significantly higher than the other experimental sets and twice as much as the control set.
Several experiments were carried out to test the performance of the filter
membrane when soaked in heptane for a longer time. In general, the membrane cannot be
used after being immersed in heptane for too long. It was found that after an immersion of
over 5 hours, the membrane was damaged from permeating heptane solution. From the
observation, once pressure started to charge into the stirred cell, the membrane was
unable to retain any heptane solution. Similar outcome was observed when the membrane
was reused after its first filtration process.
Regarding the effect of prolonged soaking time in heptane, it can be seen that the
higher the feed concentration, the lower the final permeate volume was (Figure 24). 1%
v/v squalene solution had a lower final permeate volume than that of control, and 5% v/v
squalene had the lowest of them all. Such occurrence was expected mainly due to an
increase in the osmotic pressure in the feed solution, which consequently reduced the
driving force of the nanofiltration. The following equation shows the Van’t Hoff’s equation
that relates osmotic pressure and solute concentration:
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
0 20 40 60 80
Pe
rme
ate
flo
w r
ate
(Lm
h)
Time (min)
Permeate flux 1% v/v (3 hr 45 min)
3% v/v (5 hr 50 min)
5% v/v (3 hr 30 min)
Control (heptaneonly, 1 hr 40 min)
3.9 Lmh
1.55 Lmh
1.14 Lmh
0.75 Lmh
58 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Equation 6: Van’t Hoff’s first equation of osmotic pressure relating to solute concentration
Where, Π is the osmotic pressure, is the molar concentration of solute, is the ideal
gas constant, and is the temperature.
From Equation 6, the osmotic pressure increases proportionally as the molar
concentration of squalene increases, provided that pressure remains constant. As such, the
driving force, will be reduced. Hence, the observed permeate
volume and permeate flux for 1% v/v and 5% v/v squalene were lowered comparatively to
the control set.
Equation 7: Expression of driving force with effect to concentration
3.4.3.2 Permeating squalene-heptane solution at 50 bar
A similar procedure from the previous section was carried out for operating the
stirred cell at 50 bar, however, with an exception of the soaking time of the membrane in
heptane for not longer than 4 hours. Figure 26 and Figure 27 show the filtration
performance of the filter membranes operated at 50 bar, labelled with its respective
soaking time in heptane.
In general, higher operating pressure resulted in an increase in final permeate
volume for all of the experimental sets when compared to previous findings. Table 15
illustrates the difference in total permeate volume for each set operated at 30 bar and 50
bar.
59 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Figure 26: Scatter plot of cumulative permeate volume for different squalene concentrations at 50 bar
Figure 27: Scatter plot of permeate flux for different squalene concentrations at 50 bar
Table 15: Comparison of total permeate volume with respect to operating pressure
y = -0.0016x2 + 0.2672x + 0.4165 R² = 0.9982
y = -0.0015x2 + 0.1816x + 0.2333 R² = 0.9976
y = -0.0006x2 + 0.0809x + 0.2624 R² = 0.9926
y = -0.0013x2 + 0.2753x + 0.4089 R² = 0.999
0.0
2.0
4.0
6.0
8.0
10.0
12.0
14.0
0 20 40 60
Cu
mu
lati
ve v
olu
me
(m
L)
Time (min)
Permeate Volume
1% v/v (1 hr 10 min)
3% v/v (3 hr)
5% v/v (2 hr)
Control (heptane only, 30min)
0.00
5.00
10.00
15.00
20.00
25.00
0 10 20 30 40 50 60
Pe
rme
ate
flo
w r
ate
(Lm
h)
Time (min)
Permeate Flux
1% v/v (1 hr 10min)
3% v/v (3 hr)
5% v/v (2 hr)
Control (heptaneonly, 30 min)
30 bar 50 bar Differences by factor
Control (heptane only) 3.4 mL 12.3 mL 3.62
1% v/v 2.5 mL 10.9 mL 4.36
3% v/v 9.5 mL 6.1 mL 0.64
5% v/v 1.6 mL 3.2 mL 2
60 | P a g e ENG 470 ENGINEERING HONOURS THESIS
According to the table above, all the experimental sets (except 3% v/v) showed a
significant increase in total permeate volume when operated at higher pressure. This can
be explained by the fact that increasing applied pressure to the system will increase the
driving force for the filtration process, provided that the osmotic pressure of the solution
stays constant for each solute concentration. For 3% v/v squalene-heptane solution, the
comparison was unable to establish due to the fact that the membrane was damaged for a
longer soaking time, as mentioned previously (Figure 24 and 25). The following equation
illustrates the idea of the aforementioned explanation:
Equation 8: Expression of driving force with effect to applied pressure
Notice that from Figure 26, the ‘Control’ yielded the highest final permeate volume,
which suggested that the decrease in driving force due to an increase in osmotic pressure
in the feed solution was consistent. Furthermore, since the membranes did not soak in
heptane for more than 4 hours, the permeance results showed a consistent performance in
accordance with the effect of pressure and squalene concentration increase.
3.5 GCMS results
Upon review on the permeance of squalene-heptane solution, it is crucial to
perform a rejection test using the results obtained from GCMS analysis to determine
whether the membrane can separate squalene from the heptane solution.
3.5.1 Permeating squalene-heptane solution at 30 bar
A graphical representation of the GCMS results from the nanofiltration performed
at 30 bar is shown in Figure 28.
From the chemical analysis, it showed that the membrane did not separate
squalene from the feed solution successfully. The peak area of squalene detected from
GCMS analysis for the feed (Before NF) and the permeate (After NF) did not produce a
61 | P a g e ENG 470 ENGINEERING HONOURS THESIS
significant solute rejection due to its similarities. Its respective rejection value was
calculated using Equation 5 and Table 16 summarises the outcome of the nanofiltration
process.
Figure 28: Scatter plot of GCMS results for different squalene concentration at 30 bar
Table 16: Summary of the nanofiltration process outcome operating at 30 bar
From the table above, it was noted that the squalene-heptane concentration of
3% v/v yielded the highest rejection value among them all. Referring to previous findings
(Figure 24 and Figure 25), when permeating 3% v/v squalene-heptane solution with the
filter membrane pre-soaked for 5 hours 50 minutes, the permeate flux was the highest.
Here, the GCMS results showed its rejection value was the highest as well. Thus, this
suggested that higher permeate flux resulted in higher rejection value. However, according
1.0
3.0
5.0
0.0
50,000,000.0
100,000,000.0
150,000,000.0
200,000,000.0
250,000,000.0
300,000,000.0
350,000,000.0
400,000,000.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0
Pe
ak a
rea
of
squ
ale
ne
Concentration of squalene-heptane solution (% v/v)
Conc. Before & After NF
Before NF
After NF
Concentration
Peak area of
squalene in Feed
(Before NF)
Peak area of
squalene in
Permeate (After NF)
Rejection
value
Permeate
flux
1.0% v/v 163,417,513 161,493,759 1.2% 1.14 Lmh
3.0% v/v 286,497,258 264,489,450 7.7% 3.90 Lmh
5.0% v/v 333,965,981 315,256,960 5.6% 0.73 Lmh
62 | P a g e ENG 470 ENGINEERING HONOURS THESIS
to Darvishmanesh et al. (2011), they found that using a hydrophobic membrane
manufactured from SOLSEP, higher permeate flux of oil-hexane solution did not result in
higher oil rejection, but instead a relatively constant rejection value was shown for 10%,
20% and 30% w/w oil concentration. It was unclear as to why high permeate flux could
result in higher rejection value.
Nonetheless, from the GCMS results, it can be concluded that the filter membrane
used did not perform well in separating squalene from heptane solution due to such poor
rejection value (less than 10%).
3.5.2 Permeating squalene-heptane solution at 50 bar
Similarly, a scatter plot of GCMS results for the nanofiltration performed at 50 bar
had been generated and it can be referred to Figure 29.
Figure 29: Scatter plot of GCMS results for different squalene concentration at 50 bar
A similar trend can be observed from Figure 29 when compared to Figure 28. As
observed, the membrane did not appear to be separating squalene from the feed solution
as the peak area of squalene from the feed (Before NF) yielded similar results to the
1.0
3.0
5.0
-
10,000,000.00
20,000,000.00
30,000,000.00
40,000,000.00
50,000,000.00
60,000,000.00
0.0 1.0 2.0 3.0 4.0 5.0 6.0
Pe
ak a
rea
of
squ
ale
ne
Concentration of squalene-heptane solution (% v/v)
Peak Area vs Conc. Before & After NF
Before NF
After NF
63 | P a g e ENG 470 ENGINEERING HONOURS THESIS
permeate (After NF). Its corresponding rejection value had been computed, Table 17,
showing a similar trend from Table 16, with an exception of the minor difference in
rejection value.
Table 17: Summary of the nanofiltration process outcome operating at 50 bar
Likewise, permeating 3% v/v squalene-heptane solution yielded the highest
rejection value of 6.1%. According to the permeate flux profile on rejection value, both
parameters did not show any correlation to describe the effect of solute concentration to
the rejection value. According to Darvishmanesh et al. (2011), it was demonstrated that
increasing oil concentration decreased the permeability of TFC membrane, while the
retention of the membrane stayed constant over 10%, 20% and 30% w/w oil
concentration. It was not clear as to why the results of the rejection value to squalene-
heptane concentration were inconsistent with Darvishmanesh.
However, similar to the outcome from previous section (Section 3.5.1 and 3.5.2),
the separation performances for permeating 1.0 %, 3.0 % and 5 % v/v squalene-heptane
solution were not successful due to the low rejection value (less than 10%). One of the
possible explanations for the low rejection value could be due to membrane swelling.
Heptane, being a non-polar solvent, could have been repelled due to the hydrophobic
interaction with the Duracid membrane, causing surface repulsion, presuming the
membrane surface has a hydrophilic characteristic (Farid 2010). It was postulated that
when the membrane swelled, the polymer chains that make up the membrane structure
were stretched due to surface repulsion from the hydrophobicity from heptane
(Tarleton, Robinson and Salman 2006). As a result of that, free volume in the space
Concentration Peak area of squalene
in Feed (Before NF)
Peak area of
squalene in
Permeate (After NF)
Rejection
value
Permeate
flux
1.0% v/v 7,318,756 7,239,073 1.1% 7.74 Lmh
3.0% v/v 35,944,452 33,744,414 6.1% 4.18 Lmh
5.0% v/v 56,160,475 53,896,920 4.0% 2.19 Lmh
64 | P a g e ENG 470 ENGINEERING HONOURS THESIS
between the polymers increased, therefore, increased solvent permeability and lowered
rejection.
65 | P a g e ENG 470 ENGINEERING HONOURS THESIS
3.6 Summary of the findings
To sum up, based on the observations from permeating DI water, heptane and
squalene-heptane, it can be concluded that Duracid membrane was, in fact, a hydrophilic
membrane due to its high permeability in DI water. Because of its membrane nature,
Duracid membrane did not show consistent permeance for both heptane and squalene-
heptane solution. One of the factors that could cause a variance in results was due to
heptane being a non-polar solvent and hydrophobic to the Duracid membrane.
Consequently, this could lead to surface repulsion on the Duracid membrane. Another
factor was due to membrane swelling; an occurrence where the free volume in between the
polymer that made up the membrane structure was occupied by heptane molecules, hence
restricting permeability and selectivity. From the GCMS result, a maximum rejection value
of 6 % was achieved, which suggested that the Duracid membrane had retained only 6 % of
squalene in the heptane solution. With all the evidence putting together, it was concluded
that Duracid membrane was not suitable for separating this non-polar organic-
hydrocarbon system and the membrane will be ‘damaged’ when come into contact with
heptane for too long.
66 | P a g e ENG 470 ENGINEERING HONOURS THESIS
4 Conclusion
With vast amounts of research and development being conducted around the
necessity to provide improvements in the current commercial scale algae fuel production
process, the limited research on the SRNF application have restricted the growth of more
eco-friendly and efficient algae fuel production. It was the aim of this thesis to investigate
an alternative technique of separating microalgal hydrocarbon from a biocompatible
solvent, in particular, heptane. Its secondary objective also included an investigation of the
efficiency for operating a chemical process separating the mixture of hydrocarbon and
organic solvent using a computer simulation. Through experimentation, model simulation
and simultaneous literature review, this report has revealed results that could be useful for
future work as a reference and serves as a starting point for further development.
To evaluate the thermodynamic feasibility of the commercial algae fuel production,
Aspen Plus was employed to model a separation process using two types of unit
operations, namely flash separator and distillation column. After several modifications to
the original proposal, the effect of thermodynamic property methods and optimisation
using sensitivity analysis, a heat duty required to carry out the process had been computed.
Upon conducting preliminary energy balance around the simulation, it was found that the
conventional method was not feasible as the process was not able to sustain itself.
However, the simulation was able to achieve at least 99.98% heptane recovery and attain
100% hydrocarbon recovery. Further modelling will be required to implement effluent
recycle to achieve an outcome that could provide positive energy output, as well as to seek
for an alternative way to extract algae fuel and concentrate the extraction more efficiently.
Having proved that separation via phase change is not energetically efficient, an
alternative technology of implementing nanofiltration was carried out. Since no
information in regards to the materials made for the membrane purchased was disclosed,
it was hard to correlate its nanofiltration performance with the effect of solvent contact
time. However, it was found that the membrane was not suitable for permeating non-polar
solvent, and the permeance was greatly correlated to the solvent soaking time. Some
67 | P a g e ENG 470 ENGINEERING HONOURS THESIS
plausible reasons had been proposed including the effect of polarity and the occurrence of
membrane swelling. The chemical analysis also did not show any positive results, which
the membrane only accounted for a maximum solute rejection value of 6 %. Even though
such rejection value was low compared to literature value for implementing different types
of nanofiltration membrane, it is still a relatively useful piece of finding that shows
nanofiltration in the organic phase is possible. In spite of the limited convincing finding
that could show membrane filtration is better than other concentration methods,
undoubtedly, more work and in-depth research are needed to develop further in favour to
the algae fuel production.
4.1 Recommendations for future work
1. Given the findings obtained from the NF experiments were not positive, it is necessary
to seek out for some other membrane manufacturers to provide the appropriate
membrane for the investigation. Some membrane manufacturers had been found
during the literature review, and these are summarised in Table 18.
Table 18: Different Solvent Resistant Nanofiltration and Their Properties Provided by its Respective Manufacturers (Othman et al 2009; Sterlitech 2015; Evonik 2015).
Membrane
type Manufacturer
Membrane
class Polymer Type
Pore Size,
MWCO
Tmax,
°C
pH
Tolerance
Desal-DL GE Osmonics - Polyamide 150-300 90 2-11
Desal-DK GE Osmonics - Polyamide 150-300 90 2-11
MPF-34 Koch Dense PDMSb 200 40 0-14
MPF-44 Koch Dense PDMSb 250 40 3-10
STARMEMTM
120 METa
Dense Polyimide 200 50 -
NF30 Nadir Dense Polyethersulfone 400 - -
STARMEMTM
122 METa
Semi-
porous Polyimide 220 50 -
DURAMEM® Evonik - P84® polyimide 150-900 50 -
a Membrane Extraction Technology, London, UK.; b Polydimethylsiloxane.
68 | P a g e ENG 470 ENGINEERING HONOURS THESIS
2. Given the modelling aspect did not yield a positive energy output that could allow the
chemical process to sustain itself, some techniques had been proposed to improve the
separation process. This includes
i. Due to a dilute mixture of squalene in the feed stream, it is necessary to
implement a recycle stream by recirculating bottom stream from either flash
separator or distillation column back to the feed stream to increase the squalene
concentration. This could reduce the amount of heating energy to vaporise
heptane from the feed stream and yield a better squalene recovery.
ii. A separation unit operator ‘Extract’ could be used for the solvent extraction
process to further development the solvent extraction process prior to the
solvent separation process. The mass and energy balance obtained for the overall
extraction and separation process could be used for performing a techno-
economic assessment to determine its feasibility.
3. Future research on ways to synthesise solvent resistant membrane could be another
recommendation for future work if no appropriate membrane could be found
(Darvishmanesh, Degrève and Van der Bruggen 2009; Jimenez Solomon, Bhole and
Livingston 2013; Lau et al. 2012; Tarleton, Robinson and Low 2009). This provides the
user to selectively choose the right materials to fabricate the membrane via a technique
called interfacial polymerization (IP). Extensive literature review, consultation with the
vendor for the appropriate materials must be carried out. Equipment used for the
process must also be considered, and help must be sought.
4. Two membrane technologies can also be researched on to determine its viability for
the purpose of solvent separation in a biological system. Technologies which includes:
i. Membrane distillation – a thermally driven separation that is enabled due to
phase change. A hydrophobic membrane is used as a barrier for the liquid phase,
allowing the vapour phase (e.g. water vapour) pass through the membrane's
The following spreadsheet was developed to facilitate the feed composition calculations
DW 0.89 g/L *From Navid et al 2013b total oil % 30% of DW *From Navid et al 2013b
HC % 49% of total oil % *From Navid et al 2013b
HC % 14.7% of DW
HC 0.13083 g/L HC = hydrocarbon, assuming all of them are bot-oil
ρ of bot-oil 835 g/L
if culture volume 1000 mL Algae 0.89 g DW
HC 0.13083 g DW HC 0.156683 mL
liquid content 99.94%
*From Schnurr et al 2013 Solid content 0.06%
*From Schnurr et al 2013
ρ of fresh water 1000 g/L liquid content 999.4 g Based on the liquid content and the cell density (DW)
For every 1L culture
heptane 200 mL Based on 1:0.2 ratio (Culture:heptane) ρ of heptane 684 g/L
heptane 136.8 g
HC 0.10 % heptane 99.90 % 100 %
Figure 30: A spreadsheet of feed composition calculation
78 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Appendix B: Aspen Plus Program Input Setups
Appendix B.1: Input Entry for flash separator and distillation column in the
Initial Stage of Separation Process
The input entry for FS and DC can be made by going under the ‘Blocks’ tab. As
mentioned previously, the reflux ratio for this simulation was set at 1 initially and the
column pressure profile was set to 1 bar. For the purpose of extracting heptane from the
mixture stream, the main component to be separated should be heptane under ‘Feed basis’.
The following figure shows the overall input specification that needed to be entered in the
RADFRAC column.
Figure 31: Input requirements for ‘FLASH’ column under Specification tab
Figure 32: Input requirement for ‘RADFRAC’ column under Configuration tab
79 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Figure 33: Feed and component input requirement for ‘RADFRAC’ column under Feed Basis
Figure 34: Input requirement for ‘RADFRAC’ column under Streams tab
80 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Appendix B.2: Setting Change on Thermodynamic Property Method
Setting change on thermodynamic property method can be done by selecting the
method that is desired under ‘Global’ tab in ‘Properties’ section in Data Browser. This is can
be seen by referring it to Figure 43. Notice that property methods such as NRTL and
UNIFAC can be changed just by clicking the drop box on ‘base method’.
Figure 35: Setting Change on Thermodynamic Property
81 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Appendix B.3: Input Entry for Sensitivity Analysis on ‘FLASH’ Temperature
A sensitivity test was created under ‘Model Analysis Tools’, with the sets of inputs
that needed to be filled in, such as the manipulated variable, its type and the manipulated
variable limits. Figure 44 and Figure 45 show the input requirement for the sensitivity
analysis.
Figure 36: Sensitivity analysis input requirement for flash separator temperature in ‘FLASH’ model
Figure 37: Variable definition and input requirement for flash separator sensitivity analysis outputs in ‘FLASH’ model
82 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Appendix B.4: Input Entry for Sensitivity Analysis on ‘RADFRAC’ Reflux Ratio
and Molar Ratio of Distillate to Feed Flow Rate
Firstly, to optimize the reflux ratio, a sensitivity test was created under ‘Model
Analysis Tools’, with the sets of inputs that needed to be filled in. Figure 46 and Figure 47
shows the input requirement for the sensitivity analysis.
Figure 38: Sensitivity analysis input requirement for ‘RADFRAC’ reflux ratio
Figure 39: Variable definition and input requirement for ‘’RADFRAC’ sensitivity analysis outputs
83 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Next, the molar ratio of distillate to feed flow rate (MRDF) was optimized using a
‘Model Analysis Tools’. The following figures show the input requirement for the sensitivity
analysis.
Figure 40: Sensitivity analysis input requirement for ‘RADFRAC’ MRDF
Figure 41: Variable definition and input requirement for ‘’RADFRAC’ sensitivity analysis outputs
84 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Appendix B.5: Input Entry for Sensitivity Analysis on ‘FLASH’ Temperature
The following figures show the sensitivity analysis input requirement parameters
for the second flash tank in ‘FLASH’ model.
Figure 42: Sensitivity analysis Input requirement for 'FLASH2' reactor temperature in ‘FLASH’ model
Figure 43: Variable definition and input requirement for ‘FLASH2’ tank sensitivity analysis outputs in ‘FLASH’ model
85 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Next, the reactor temperature for ‘RADFRAC’ model was optimized using a ‘Model
Analysis Tools’. The following figures show the input requirement for the sensitivity
analysis.
Figure 44: Sensitivity analysis Input requirement for 'FLASH2' reactor temperature in ‘’RADFRAC’ model
Figure 45: : Variable definition and input requirement for ‘FLASH2’ tank sensitivity analysis outputs in ‘’RADFRAC’ model
86 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Appendix B.6: Input Entry for Sensitivity Analysis on ‘FLASH’ Temperature
The following figures show the input parameters for ‘1SQUAL’ stream in both
‘FLASH’ and ‘RADFRAC’ model.
Figure 46: Input requirements for stream ‘1SQUAL’ for water content variable
87 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Appendix C: HP4750 Stirred Cell Features and Specifications
The following summarises the essential features and technical specification of the HP4750
Stirred Cell that is made by Sterlitech Corporation.
Table 19: HP4750 Features and Technical Specifications (Sterlitech 2015)
Parameter Description
Membrane size 47 to 49 mm diameter
Active membrane area m2 (43.12 mm diameter)
Processing Volume 300 mL
Hold-up volume 1 mL
Maximum Pressure 69 bar (69000 kPa or 1000 psig)
Maximum Temperature 121 °C (250 °F) @ 55 bar (55000 kPa or 800 psig)
pH range Membrane dependent
Connections:
Permeate Outlet
Pressure Inlet
1/8-inch diameter 316L SS tubing
¼ inch FNPT
Wetted materials of construction:
Cell Body
O-rings
Gaskets
Stir Bar
316L stainless steel
Buna-N
Buna-N
PTFE-coated magnet
Dimensions:
Cell Body diameter
Cell Top width*
Cell Bottom width*
Cell height
Assembled weight
5.1 cm
10.2
13.3
22.1
2.72 kg
Autoclavable Yes
* Measurement included assembling with clamp/coupling
88 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Appendix D: HP4750 Stirred Cell Components
The HP4750 Stirred Cells was shipped with the following components and the complete set
can be referred as Figure 30.
Figure 47: HP4750 parts and components (Sterlitech Corp 2015)
1. Stainless steel cell body 2. Cell top 3. Cell bottom 4. Cell top coupling 5. Cell bottom coupling 6. Porous stainless steel membrane support disk 7. Two O-rings 8. Top gasket 9. Permeate tube 10. Stir bar assembly 11. Stir bar retriever
89 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Appendix E: HP4750 Stirred Cell Assembly
The following figures illustrate the procedure on how to assemble the stirred cell
appropriately:
1. O-ring insertion:
Ensure that the O-rings were wetted with the fluid to be processed and the
insertion was properly fitted in the grooves
Figure 48: Outer O-ring (left) and inner O-ring (right) insertion
2. Membrane and porous membrane support disk insertion:
Ensure that the active side of the membrane, which usually have a shiny, coated
surface, facing toward the cell reservoir, while a dull side facing the other way
Followed by the stainless steel porous membrane support disk being placed on
top of the membrane to hold it in place.
Figure 49: Membrane (left) and porous membrane support disk (right) insertion
3. Cell Bottom and bottom clamp assembly:
Ensure that the alignment between the Cell Bottom and Cell Body is properly
done by fitting the circular ridge on the Cell Body onto the circular groove on the
Cell Bottom
Ensure that the high pressure coupling is properly tighten using the appropriate
wrench
Figure 50: Cell Bottom fitting (left) and high pressure coupling assembly (right)
90 | P a g e ENG 470 ENGINEERING HONOURS THESIS
4. Permeate Tube assembly and Stir Bar insertion:
Ensure that the Permeate Tube is tighten using the appropriate wrench
Ensure that the Stir Bar is lowered into the cell body using the Stir Bar Retriever,
preferably not dropping it. After the Stir Bar is in place, the feed solution can be
poured in and filtered out upon assembly completion.
Figure 51: Permeate Tube assembly (left) and Stir Bar insertion (right)
5. Cell top insertion and top clamp assembly:
Ensure that the gasket and the Cell Top are fitted accordingly
Ensure that the high pressure coupling is properly tighten using the appropriate
wrench
Figure 52: Gasket assembly (left), Cell Top installation (middle) and high pressure assembly (right)
6. High pressure hose and pressure regulator connection:
A thermoplastic, non-conductive 7N Series 6.4 mm Swagelok hose was attached
to the fitting on the Cell Top, connecting the other end of the hose to Victor
CutSkill® TPR250 ¼’’ flare fitting pressure regulator using the appropriate
wrench
Figure 53: High pressure hose attachment (left) and pressure regulator connection (right)
91 | P a g e ENG 470 ENGINEERING HONOURS THESIS
The pressure regulator was then assembled on an industrial grade nitrogen gas
cylinder, which was purchased from BOC Gases. A Permeate Collection Vessel was placed
under the Permeate Tube and the stirred cell was placed on a magnetic stirrer.
Before commencing any filtering process, the membrane was pre-conditioned by
gradually pressurizing the stirred cell to check for leaks and to ensure consistent
performance. Once the filtration process had been completed, the pressure source was
turned off and the stirred cell was depressurized by opening the pressure discharge port
slowly via the relief valve. It was highly recommended not to depressurize the stirred cell
by loosening the coupling, as it would cause sudden burst upon opening the stirred cell.
Once the stirred cell was depressurized to ambient pressure, the cell was emptied,
cleaned with heptane and dried with paper towel. Upon reviewing the choice of the
appropriate cleaning regime for the stirred cell from the HP4750 Operation Manual
(Sterlitech Corp 2015), it was found that n-heptane was chemically compatible to the
material used for gasket and O-rings.
As a safety precaution, during the assembly, operation and cleaning processes, the
stirred cell was kept in a fume hood to ensure no heptane vapour escaped into the
environment.
Figure below illustrates the schematic view of the HP4750 Stirred Cell system:
92 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Figure 54: HP4750 System Configuration (Sterlitech 2015)
93 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Appendix F: GC-MS Method Parameters
Model used in the GC-MS chemical analysis:
GCMS-QP2010S Gas Chromatograph-Mass Spectrometer
GC-2010 Gas Chromatograph
AOC-20i+S Auto Injector and Auto Sampler
The following tables summarise the method parameter used for the chemical analysis of
the feed and permeate solution:
Table 20: GC-2010 Gas Chromatograph parameters
Parameter Description
Column Oven Temperature 220.0 °C
Injection Temperature 300.0 °C
Injection Mode Split
Flow Control Mode Linear Velocity
Pressure 118.9 kPa
Total Flow 54.0 mL/min
Column Flow 1.0 mL/min
Linear Velocity 38.9 cm/sec
Purge Flow 3.0 mL/min
Split Ratio 50.0
High Pressure Injection Off
Carrier Gas Saver Off
Splitter Hold Off
Oven Temperature Program
Rate Temperature (°C) Hold Time (min)
- 220.0 1.00
2.00 260.0 1.00
External Wait No
Equilibrium Time 3.0 min
Table 21: MS Mass Spectrometer parameters
Parameter Description
Start Time 1.50 min
End Time 11.00 min
ACQ Mode Scan
Event Time 0.50 sec
Scan Speed 2000
Start m/z 45.00
End m/z 1000.00
94 | P a g e ENG 470 ENGINEERING HONOURS THESIS
Table 22: GCMS-QP2010 Gas Chromatograph-Mass Spectrometer parameters
Parameter Description
Ion Source Temperature 200.0 °C
Interface Temperature 200.0 °C
Solvent Cut Time 1.00 min
Detector Gain Mode Relative
Detector Gain 0.00 kV
Threshold 1000
Table 23: AOC-20i/S Auto Injector and Auto Sampler parameters