WET OXIDATION AND
NANOFILTRATION OF HTC
PROCESS WATER
Separation of organic matter
Nathalie SöderlindSpring 2020
Degree Project in Bioresource Technique Engineering, 30 credits
Examiner: Tomas Hedlund
I
Abstract
Wastewater treatment in forest industries and municipalities produce large amount of sludge, that without treatment may cause environmental problems, such as eutrophication, oxygen depletion and toxic aquatic environment. There is an acute need for a sustainable application of the sewage sludge from the wastewater treatment, as bans and regulations prevent previously used applications. Hydrothermal carbonisation is a method used to upgrade sewage sludge to hydrochar but causes formation of a process water (HTC-water) that needs treatment before it can be transferred to a wastewater treatment plant. In this study wet oxidation (WO) will be compared with nanofiltration (NF), and with a combination of WO and NF, to evaluate the reduction of organic, nitrogen-, phosphorous- and sulphur-containing compounds in the HTC-water. WO was performed in 230°C, 2 h and with a pressure of O2 equivalent to 150% of the COD of the HTC-water. Two thin filmed composite
polyamide flat sheet membranes, with a molecular weight cut-off (MWCO) of 150 – 300 Da (denoted ‘DL’) respectively 300 – 500 Da (denoted ‘NFW’), and recommended pressure of 220 and 110 psi, respectively, were used in the NF. The DL membrane had a higher flux despite its lower MWCO but because of the higher pressure. It also had a better separation of analytes than the NFW membrane. NF provided a more efficient reduction than WO, by reducing the analytes TOC (55 %), COD (15 %), BOD7 (100 %), P-containing compounds (275 %) and N-containing
III
List of abbreviations
BOD Biochemical oxygen demand
BOD7 Biochemical oxygen demand on a 7 day basis
COD Chemical oxygen demand
DM Dry matter
DOC Dissolved organic compounds HTC Hydrothermal carbonisation LC Liquid chromatography
LC-MS Liquid chromatography – mass spectrometry
MF Microfiltration
MS Mass spectrometry
MW Molecular weight
MWCO Molecular weight cut off m/z Mass to charge ratio in MS
NF Nanofiltration
RO Reverse osmosis
SDG Sustainable development goals TOC Total organic carbon
UF Ultrafiltration
UV-vis Ultraviolet visible spectroscopy
WO Wet oxidation
Author contribution
WO on HTC-water, nanofiltrations of HTC-water and WO HTC-water, DM measurements and LC-MS analysis have been executed by the author.
The HTC-water used in this study was provided by C-Green Technology AB.
The analysis of TOC, COD, BOD7, P, S, N, NO2--N, NO3--N and NH4+-N was
V Table of contents
Abstract ... I Author contribution ... III
1 Introduction ... 1
1.1 HTC, a solution to sludge problem ... 1
1.2 Wet oxidation ... 1
1.3 Membrane filtration... 2
1.4 Liquid chromatography-mass spectrometry ... 3
1.5 Content of organic, nitrogen-, phosphorus- and sulphur-containing compounds ... 4
1.6 Aim of the diploma work ... 4
2 Popular scientific summary including social and ethical aspects ... 5
2.1 Popular scientific summary ... 5
2.2 Social and ethical aspects ... 6
3 Experimental ... 6 3.1 Wet oxidation ... 7 3.2 Membrane filtration... 7 3.3 Analysis ... 8 3.3.1 DM ... 8 3.3.2 LC-MS ... 8
3.3.3 Organic and nitrogen compounds and elemental content analysis ... 8
4 Results ... 9
4.1 Nanofiltration ... 9
4.2 LC-MS ... 11
4.3 Distribution between permeate, retentate, feed and membrane ... 13
4.4 Concentrations ... 15 4.5 WO vs. NF vs. WO+NF ... 17 5 Discussion ... 18 5.1 Wet oxidation ... 18 5.2 Nanofiltration ... 19 5.2.1 Separation of HTC-water ... 19 5.2.2 Separation of WO-water ... 20
5.2.3 Flux and Fouling ... 21
5.3 Most effective reduction... 22
5.4 Contribution to society, industry and research ... 23
5.5 Future prospects ... 24
6 Conclusions and Outlook ... 25
1
1 Introduction
1.1 HTC, a solution to sludge problem
Wastewater treatment from municipal sewage and forest industries produces sludge with high levels of carbon, nitrogen, phosphorus and water. Due to the high water content and often presence of toxic compounds, the utilisation of sludge is problematic (1,2). Furthermore, due to a Swedish regulation (2001:512) that prevents sludge from sewage treatment plants in landfills, another use needs to be found. A newly released government inquiry, from 2020 (2), suggests a ban on spreading sewage sludge on/in soil, which would prevent use of sewage sludge as fertilizer or landfill cover. The ban aims to prevent spread of hazardous substances, pharmaceutical residues and microplastics into ecosystems. The inquiry also emphasized phosphorous extraction from the sludge, to secure phosphorous as a circular resource (2).
Treatment of sludge often involves dewatering steps, to ease handling. The origin and composition can vary between sludges, and water is bound in different ways to the suspended particles in different sludges. This makes it challenging to find one suitable dewatering process (3). By using hydrothermal carbonisation (HTC) the sludge can be converted into hydrochar, with properties similar to fossil charcoal and with increased energy density compared to the starting material (4). Thereafter, the water can easily be separated from the hydrochar (3–5). The HTC is performed in a closed reactor, with constant volume, at elevated temperature (180 – 250°C) and self-generated pressure (3,4,6). The HTC process is cost and energy effective, with possibilities to recycle nutrients, such as phosphorus, and create useful energy carrier (3).
The process water from the HTC is not pure, and contains toxic and organic matter (7), which prevents recirculation to internal of external wastewater treatment facilities. The HTC-water has been shown to contain solubilized and/or fragments of proteins, lipids and carbohydrate (8), C1-C2 carboxylic acids (7), aldehydes, alkenes, aromatic, furanic
and phenolic compounds (9).
1.2 Wet oxidation
Wet oxidation (WO) is a process which can degrade recalcitrant pollutants, generate volatile fatty acids (8), increase the biodegradability of dissolved organic carbon (DOC) (7) and reduce biochemical oxygen demand (BOD), total organic carbon (TOC) and chemical oxygen demand (COD) (5,7), which can contribute to oxygen depletion when released in nature (10). With an increased pressure of O2 and heating,
the molecules in the process water oxidises. This have been found to generate more NH4+, by oxidation of organically bound N (5,8). Some increase of NO2- and NO3
-occurs, as well, during the WO, but to a lesser extent than NH4+ (5). An increased O2
pressure also increase the concentration of organic acids (5,7). Most abundant is acetic acid (55-60% of DOC), formic acid (20-30% of DOC) and lactic, levulinic and propanoic acid (5-10% of DOC) (5,7,8).
The oxidation is exothermic (5,7), and once started no further heating is needed. The excess energy released can be used to drive the HTC reaction (7).
In order to prepare efficient WO, enough O2 needs to be applied. This is can be
2 Where 𝑝𝑂2 is the pressure of O2 in Pascal (Pa), 𝑐𝐶𝑂𝐷 is the concentration of COD in
g/l, 𝑀𝑂2 is the molar mass of O2 (32 g/mol), 𝑅 is the gas constant (8.3145 J·mol-1·K-1),
𝑇 is the temperature in Kelvin (K), 𝑉𝑟 is the volume of the reactor (0.42 l) and 𝑉𝑙 is the volume of the liquid in litres (l).
WO degrades a lot of organic matter, however, not everything is converted to CO2,
thus further purification of the water may be needed to meet the requirements of legislation (5,7,8).
1.3 Membrane filtration
Membrane filtration can be used for water purification. It uses membranes with a certain pore size and an applied pressure, to push solvent and certain solutes through the membrane (1).
Membrane filtration can be divided into four categories depending on their pore sizes and molecular weight cut off (MWCO); microfiltration (MF), ultrafiltration (UF), nanofiltration (NF) and reverse osmosis (RO). RO is the most selective membrane and only permeates water, with a pore size less than 0.001 µm. NF allows monovalent ions through its 0.001 – 0.01 µm pores and has a MWCO of 200 – 1 000 Da. UF can permeate small molecules with sizes up to 0.1 µm, or 2 000 – 100 000 Da. Lastly, MF has the largest pores sizes, 0.1 – 10 µm or > 100 000 Da, and can retain cells, cell fragments and DNA (1,11,12).
Pressure is applied over the membrane in order to push the permeate through the pores. By increasing the pressure more permeate can be obtained. With decreasing pore size, the pressure required to enable the highest possible flux increases (1,11,12). The flux will, as a consequence of the small pore size, also be lower. The flux is calculated according to:
𝐹𝑙𝑢𝑥 = 𝑉𝑝 𝐴𝑡
Where 𝑉𝑝 is the volume of the permeate in litre (L), 𝐴 is the active area of the membrane in square meter (m2) and 𝑡 is the time, in hours (h), required to obtain V
p
(1).
The material of the membrane will affect the performance of the filtration. The material choice will depend on the liquid that is to be filtrated. If the solvent is organic a hydrophobic membrane could be suitable, whereas water solutions require hydrophilic membranes, such as cellulose acetate, or polysulfonate and polyether-sulfone, which can be modified to become moderately hydrophilic (12,13). Additionally, there are other membrane materials that can serve other purposes. For instance, ceramic membranes are more resistant to water transport, temperature and a wide pH range, and are easier to maintain and clean. These are inorganically based, from sintered aluminium oxide, titanium oxide or zirconium dioxide (13). Mixed matrix membranes or composite membranes consist of a bulk layer, providing stability, and an active outer skin of a different material.
Thin film composites (TFC) have a thin (0.15 – 0.25 µm) active layer, e.g. cellulose acetate or polyamide, that is bound to a thicker, stability providing, porous substrate (1). Most TFC membranes carries a negative charge at neutral pH, thus rejecting negatively charged compounds due to electrostatic repulsion (12,14). The rejection of negatively charged organic acids have been confirmed to be greater, than what was expected from the MWCO of the filter and correlated with the degree of ionization of the compound. A study of Bellona et al. was done with a polyamide TFC NF
3 membrane, and by increasing the pH of the feed, the negative charge on the membrane increased, as well as the deprotonation of the organic acids, giving them a negative charge, thus increasing the rejection (15). Membranes with a stated MWCO of 200 – 400 Da have rejected compounds as small as 150 – 180 Da, even at acidic conditions (16).
By recirculating the retentate as feed, as seen in Figure 1, the retentate will be more concentrated and more permeate will be obtained. This is known as crossflow. During the filtration some material will allocate on the membrane, called fouling. Extensive fouling will eventually lower the flux and change the filtration effectiveness. Backwashing and eventually chemical cleaning are methods for removing fouling (12).
Figure 1. A schematic view of a crossflow filtration process. Feed is applied over the membrane and pushed through the pores of the membrane. The liquid obtained through the membrane is called permeate. The liquid that does not pass through the membrane is called retentate and is recirculated as feed.
1.4 Liquid chromatography-mass spectrometry
Liquid chromatography-mass spectrometry (LC-MS) is a technique used for separation and detection of molecules in a liquid sample. In the first part, liquid chromatography (LC), separation of the molecules in the sample is achieved in a column with a polar or non-polar stationary phase. The molecules in the sample will be separated by their polarity. In a column with a non-polar stationary phase, polar compounds will elute early, with a low retention time. As the retention time increases more non-polar molecules will elute. The sample travels through the column by a mobile phase, which can have different polarity and pH depending on the sample and column. This, and the flow rate of the mobile phase, will affect the separation of the molecules in the column. The elutes are often detected by ultraviolet visible spectrometry (UV-vis). UV-vis detectors measure the absorbance of the molecules at a certain wavelength. At wavelengths ≥ 210 nm only molecules with an adsorbing chromophore can be detected (17).
The second part of the LC-MS, mass spectrometry (MS), vaporizes the molecules and ionizes them. The ionized molecules are seldom stable and fragmentates into smaller ions. The ions travel through a charged chamber which is selective for a certain mass to charge ratio (m/z). By changing the selectivity for the m/z, the entire range of ions present in the chamber can be detected (17).
Every molecule provides a specific fragmentation pattern. An obtained mass spectrum can therefore be compared to spectra of molecules with known fragmentation patterns.
Feed
Permeate
4 Thus, an unknown molecule in a sample can be identified. This is facilitated by good separation in the LC (17).
1.5 Content of organic, nitrogen-, phosphorus- and sulphur-containing compounds
To determine how WO and NF alter the content of organic carbon, nitrogen-, phosphorus- and sulphur-compounds in the HTC-water, different types of analyses can be performed. The dry matter (DM) in a sample is determined by the ratio between the mass of the dried and wet material. A low DM indicates purer water, with less matter. However, it does not take volatile compounds into account. The majority of the molecules in the HTC-water are organic. TOC measures the total amount of organic carbon, which includes dissolved and undissolved organics. To degrade organic molecules oxygen is needed. To measure the amount of oxygen that is needed for biological degradation under aerobic conditions, BOD7 is measured. The total amount
that can chemically be oxidised is determined by COD (10). The ratio of BOD7:COD
can indicate the readiness of the degradation of organic compounds in the sample, were a higher ratio indicates higher readiness (19). Not everything in the process water is organic, nitrogen containing compounds are also present and are nutrients that contributes to eutrophication (10). NO2-, NO3-, NH4+ and total nitrogen (N),
phosphorus (P) and sulphur (S) content can be measured in water (1). 1.6 Aim of the diploma work
By WO and/or NF reduction of DM, TOC, COD, BOD7, P, S, N, NO2-, NO3- and NH4+
content in HTC-water, is hoped to be achieved. The aim is to evaluate whether, and how, WO and NF contribute to change in these parameters, for a safer handling and/or release of the HTC process water.
Two NF membranes, with different MWCO, denoted ‘NFW’ and ‘DL’ by the suppliers, will be used and evaluated. The chosen MWCOs are in the upper and lower end of the anticipated MW distribution range. Due to the electrostatic repulsion from the membrane to negatively charged molecules, the membrane with larger MWCO could perform equally well as the membrane with lower MWCO in rejecting acids, at a pH>7. By comparing the nitrogen-, phosphorus-, sulphur-, COD, BOD7 and TOC
content, before and after filtration, an indication can be given of which types of molecules are being separated and whether the choice of membrane matters.
Degradation of organic matter is anticipated from WO, which would conduce in lower MW, lower COD and BOD7, because of the previous oxidation, and lower TOC.
Proteins present in HTC-water, could after WO be degraded into small nitrogen compounds, which would show in increased amounts of NO2-, NO3-, NH4+ in the
filtration feed and consequently in the permeate. With the negative charge of the membrane taken into consideration, higher levels of NH4+ is expected in the permeate,
because of attraction, while the levels of NO2- and NO3- should be repelled and found
in higher concentrations in the retentate.
5
Figure 2. A view over the experimental set-up. NFW and DL are the two types of NF membranes that are being used in this study. Filtration of HTC-water with NFW and DL membranes are denoted with red and orange colour, respectively. Filtration of WO-water with NFW and DL membrane are denoted with purple and green colour, respectively. The analysis after the NF will be performed on both the permeate and retentate.
2 Popular scientific summary including social and ethical
aspects
2.1 Popular scientific summary
When treating wastewater from municipalities or industry, an anaerobic digested sludge is formed. There is a problem with finding an environmentally friendly use of this sludge, since it contains some hazardous substances. When it is treated at 200°C for 2 h, under pressure, it forms an energy-rich hydrochar, a process which takes thousands of years in nature. The water that was trapped in the sludge can now easily be filtered off. This water is not completely clean and needs treatment, to prevent heavy burdens on wastewater treatment facilities or release of hazardous substances. In a process called wet oxidation (WO), O2 is applied to the water in a closed reactor,
by stirring the water the contact with the oxygen is even for all the water. By heating the reactor to 230°C, the pressure also increases and the molecules in the water react with the oxygen. The reaction is the same as when burning something in air; the material is oxidised by the oxygen. Thus, wet oxidation is like burning molecules in water. By doing this the molecules in the HTC-water were degraded. The total amount of carbon-containing molecules was reduced by 54% and the amount of oxygen required to completely burn the carbon-containing molecules was reduced by 65% during the WO of HTC-water. The measure of how easily the organic compounds in the sample can be degraded, was 75% higher after the WO treatment. Some molecules contain nitrogen, and these were not reduced but transformed to NH4+, increasing the
NH4+ concentration by 164%.
Two membranes, one with 150 – 300 Da pores and one with 300 – 500 Da pores, were used to filter off the molecules in the water. Da is a unit where 1 Da ≈ 1 proton or neutron. These pores are approximately 100 – 300 times smaller than the diameter of human hair. When the pores are that small the process is called a nanofiltration and a pump is needed to put a pressure on the water to filter it through the membrane.
Filtrating after wet oxidation gave little effect, probably because the molecules remaining after the wet oxidation were too small to be rejected by the filter. Only filtering, without wet oxidation, was just as efficient, since the molecules were already in the right size for filtration. Using the filter with 150 – 300 Da pores was more
6 efficient than using the filter with 300 – 500 Da pores. That is because the pores of the membrane are smaller and could separate more of the compounds. Filtrating with the more efficient membrane (150 – 300 Da) was also 4 times faster, because a higher pressure could be applied on this membrane.
2.2 Social and ethical aspects
Wastewater treatment is an essential part in our society, with an extensive, never-ending production of sludge as its side stream. Because of the Swedish regulation (2001:512), that prevents sludge from sewage treatment in landfills, and the suggested ban hindering the use of sewage sludge as fertilizer and landfill cover, finding useful application of the sludge is acute. With HTC the sludge can be converted into useful products, but without treatment of the process water the problems with eutrophication and hazardous substances in the environment are only postpone from the sludge to the process water.
With successful WO and/or NF, reduction of organic- and nitrogen-containing compounds will provide cleaner water. By understanding how the distribution of the organic matter, nitrogen-, phosphorus- and sulphur-containing compounds changes during NF, the search for applications for permeate and retentate will be aided. With a society as ours, where the earth’s recourses are being overutilized, it is vital to complete the cycle of resources and not waste them.
It is important to ensure that all parts of residual steams are utilized to truly fulfil a sustainable process. This is in line with four of the UN sustainable development goals (SDG). Goal 6.3 Improve water quality, wastewater treatment and safe reuse, aims to improve water quality by reducing pollution and release of hazardous chemicals and increasing the portion of treated wastewater, recycling and safe reuse of water. Goal 9.4 Upgrade all industries and infrastructures for sustainability, correlates with this study, because it targets industrial upgrades that increase resource use efficiency and clean and environmentally sound technologies and processes. Goal 11.6 Reduce the
environmental impact of cities, can greatly benefit from improving waste management,
as can be done by HTC of sludge and treatment of the process water. Goal 14.1 Reduce
marine pollution, focuses on preventing and reducing marine pollution of all kind, in
particular from land-based activities, like nutrient pollution, which would happen if treatment of the HTC wastewater would not exist (20). Two of the Swedish environmental objectives can also be benefited; a non-toxic environment, by reducing hazardous compounds that are released in lakes, rivers and seas, and zero
eutrophication, by limiting the nitrogen that ends up in lakes and seas (21).
WO and NF as water treatment, as part of the HTC process, makes it possible to reuse the water within the industry, or release it into nature without causing eutrophication or harming the aquatic life. Thus, making the industry of the green HTC process, that already utilizes a large side stream from wastewater treatment, even more sustainable.
3 Experimental
7
3.1 Wet oxidation
220 g of HTC-water was loaded in a 420 ml stainless steel reactor. At an oxygen content of 150% of COD (6 400 kPa for 220 g HTC process water, obtained by equation 1) the reactor was heated to the reaction temperature; 230°C. The reaction time was 2 h. The reactor was cooled to room temperature and the oxidised material was obtained. 8 replicates were made, giving a pooled total of 1680.7 g WO water. See Figure 3 for set-up.
100 g of WO water was removed for analysis and the remaining was used in filtration.
Figure 3. The WO set-up. The reactor containing 220 g HTC-water is marked in orange. The stirrer motor, stirring with 500 rpm, is marked in purple. The O2 is let in through the blue pipe, to a
pressure of 6 400 kPa. In pink is a 20 MPa rupture disc. The reactor is placed in the mantel (red) for heating to 230°C. After the WO, and cooling, the pressure was released through the green vault.
3.2 Membrane filtration
Two polyamide TFC nanomembranes, ‘DL’ (GE Osmonics) with a MWCO of 150 – 300 Da and recommended pressure of 220 psi, and ‘NFW’ (Synder) with a MWOC of 300 – 500 Da and recommended pressure of 110 psi, were used for NF.
The pooled WO-water and the HTC-water were each diluted 3.75 times with distilled water, to yield 6 litres of each feed material for the filtrations. 1 litre of diluted feed water was used for each filtration, allowing three replicate runs of each feed for each membrane.
The membranes were installed in the filtration cell, SEPA CF II membrane element cell, (GE infrastructure water and process technology). A 300 – 400 psi pressure was applied over the membrane, to prevent leakage. The feed was pumped through the filtration cell, using a Busck MS 80A-4 B34 pump. See Figure 4 for set-up. The feed was filtrated through the DL membrane with a pressure 220 psi and the NFW membrane with a pressure of 110 psi, adjusted by a retentate flow control vault. The flux was measured by measuring the volume of permeate per unit time, at different times during the filtration. The retentate was circulated back to the feed, see Figure 4. The filtration was run until 400 ml of the retentate remained.
8
Figure 4. Filter set-up, showing the permeate catcher to the left (turquoise), the filter cell in the middle (red), feed flask in the back (orange), with feed inlet tube (blue) going through the pump at the right (purple), and retentate outlet tube (green) returning from the filter cell to the feed flask. The retentate pressure gauge and retentate flow control valve (pink) is located on the retentate outlet tube.
3.3 Analysis
3.3.1 DM
The DM of the HTC-water, WO-water, and all three replicates of permeate and retentate from the filtration was measured in a Mettler Toledo MJ33 moisture analyser.
3.3.2 LC-MS
The HTC-water, WO-water, and the pooled permeate and retentate samples were analysed with LC-MS, Agilent 1260 infinity II HPLC with incorporated UV-vis detector. The LC column was operating at 30.0°C, with a flow rate of 0.300 ml/min, with 0.5% formic acid (≥95%, Sigma Aldrich) in milliQ water (pH=4) as mobile phase. The injection volume was 2.50 µl. The column used was a Zorbax SB-C18 from Agilent.
The UV-vis detector detected compounds at a wavelength of 254 nm with a bandwidth of 4 nm. The MS was operating in positive mode, with 3000 V, a gas temperature at 250°C, drying gas flowrate of 12.0 l/min, nebuliser pressure of 35 psi and the quadrupole temperature at 0°C. The molecules were fragmented by 70.00 V and detected in the mass range of 50.00 – 2000.00 m/z.
The all samples were diluted to 1 g/l, based on the DM. Each sample was filtered with a 0.22 µm nylon filter prior addition to LC-vials. All samples were analysed with UV-vis and MS.
3.3.3 Organic and nitrogen compounds and elemental content analysis
The pooled samples of permeate and retentate, together with the feeds, HTC- and WO-waters, were diluted with distilled water prior sending the samples for analysis. The samples were sent to ALS Life Science, for analysis of TOC, COD, BOD7, NO2--N,
NO3--N, NH4+-N, N, P and S. The results of TOC, COD, BOD7, NO2--N, NH4+-N, P
9
4 Results
4.1 Nanofiltration
Before the WO the HTC-water was dark brown, close to black, see Figure 5. After WO the water had brighter colour than before the WO. Furthermore, the WO-water also had precipitated material.
Figure 5. The feeds for the filtrations. The left of the WO-bottles has been allowed to settle, while the right bottle has been stirred.
The permeates of the NFs were clear for all samples, regardless of the membrane or feed used see Figure 6. The permeate had darker colour when HTC-water had been filtered and when the NFW membrane had been used, see Figure 6. All of the retentates had the same dark colour as the HTC-water regardless of membrane and feed used. The retentates from the WO-water had sedimentation of particles, resulting in a layer of particles in the bottom of the flasks, with a clear, but coloured liquid.
Figure 6. Comparison between the permeates (top) and the retentates (middle and bottom) from all NFs. The feed for the bottles in the left half is HTC-water (red and orange dashed lines) and on the right, the feed is WO-water (purple and green dashed lines). The middle picture has unstirred bottles, where sedimentation can be seen for the WO retentate. The bottom picture has stirred flasks, where the sedimentation has not settled.
WO.NFW WO.DL
HTC.NFW HTC.DL
10
Figure 7. The membranes from the filtrations. More fouling is observed when WO-water is used as feed (right half). The fouling also has a more yellow tone of brown, compared to the HTC-water fouling. More fouling is seen in the DL filters, compared to the NFW filters. New filters were used in every filtration.
The flux of the membrane (calculated according to equation 2) increased slightly as the filtration proceeded, for both membranes and both types of feeds, see Figure 8. The pressure of the feed dropped occasionally during the filtration and was restored to the set pressure when noticed. The average pressure drop during the whole filtration is seen inTable 1 Table 1. During one of the filtrations of WO-water with the DL membrane, the pressure fluctuated between 180 – 220 psi during the entire filtration and could not be measured. The largest pressure drop was found in WO.DL filtration, 70 psi (Table 1). This filtration also had most visible amount of fouling (Figure 7). Less fouling was seen when the HTC-water was filtered. Additionally, by comparing the membranes, NFW seems to have lower amount of fouling.
Additional data from the filtration can be found in Appendix 1.
Table 1. The total pressure drops of the filtrations, presented as the mean of three replicates of each membrane-feed combination and the ±2 standard deviation, giving an approximately 95% confidence interval. During one filtration of WO-water with DL membrane the pressure fluctuated and could not be measured, thus no standard deviation could be calculated, since the mean is calculated from two replicates. The average absolute deviation is presented for WO.DL instead.
HTC.NFW HTC.DL WO.NFW WO.DL
Pressure drop/ psi 33 ± 11 33 ± 23 37 ± 11 70 ± 0
WO.NFW WO.DL
11
Figure 8. The mean flux, from three filtration replicates, changing with time of filtration and the change in pressure during the filtration. Error bars represent the ±2 standard deviation. The error bars of ‘Pressure WO.DL’ are the average absolute deviation. After a detected pressure drop the pressure was restored to its original pressure. The NFW membrane is depicted with red and purple lines and the DL membrane with orange and green lines. The purple and green lines are from NF with WO-water.
The filtration process was over four times faster with the DL membrane than the NFW membrane. Around 20 minutes compared to 85 minutes for 1 l feed, see Figure 8. The pH was measured before and after the filtrations with pH-paper. The pH of the feeds was 7 for HTC-water and 7.5 for WO-water. The pH of the permeate and retentate from the filtration of HTC-water was 7, and from WO-water 8 (Appendix 1).
4.2 LC-MS
From the LC-MS runs, UV-vis spectra were obtained together with MS signals and mass spectra. The UV-vis spectra of the HTC-water, the permeates and the retentates of the HTC-water can be seen in Figure 9. The software (Agilent Open Lab (Chemistry edition)) detected peaks in the UV-vis spectra and calculated the integrals of the peaks. Presented are the results of the software’s default determinations, in order to have a similar detection pattern for all samples. Therefore, could peaks appear in the spectra, that are not mentioned in the percental distribution charts, which displays all the detected peaks (Figure 10 and Figure 12).
12 (orange and dark orange). In Figure 10 the relative distribution of the area under the peaks can be seen.
Figure 9. UV-vis spectra after LC of the HTC samples from the NF. Sample names beginning with ‘P’ is the permeate of that filtration, and sample names beginning with ‘R’ are the retentate of that filtration. Sample names ending with ‘NFW’ or ‘DL’ shows if the sample was obtained from NF with NFW or DL membrane. HTC-water was the feed for the filtrations.
Figure 10. The percental distribution of the peak areas in the UV-vis spectra in Figure 9. Sample names beginning with ‘P’ is the permeate of that filtration, and sample names beginning with ‘R’ are the retentate of that filtration. Sample names ending with ‘NFW’ or ‘DL’ shows if the sample was obtained from NF with NFW or DL membrane. HTC-water was the feed for the filtrations.
13
Figure 11. UV-vis spectra after LC of the WO samples from the NF. The dashed lines show, from left to right: peaks at 1.0 minutes, 1.1 minutes, 2.4 minutes and 3.3 minutes. The box to the right shows the peaks at eluting at retention time 3.3 minutes. Sample names beginning with ‘P’ is the permeate of that filtration, and sample names beginning with ‘R’ are the retentate of that filtration. Sample names ending with ‘NFW’ or ‘DL’ shows if the sample was obtained from NF with NFW or DL membrane. WO-water was the feed for the filtrations.
Figure 12 The percental distribution of the peak areas in the UV-vis spectra in Figure 11. Sample names beginning with ‘P’ is the permeate of that filtration, and sample names beginning with ‘R’ are the retentate of that filtration. Sample names ending with ‘NFW’ or ‘DL’ shows if the sample was obtained from NF with NFW or DL membrane. WO-water was the feed for the filtrations.
The compounds eluting at 2 and 4.7 minutes of the HTC filtration and at 2.4 and 3.3 minutes of the WO filtration are not detected in the MS.
No acids were detected in the samples. An LC-MS analysis with basic mobile phase (pH=8, ammonia solution) in negative MS-mode did not result in any mass signals (Appendix 2). Under these conditions carboxylic acids are deprotonated and thus stable anions, hence commonly detected. Furthermore, the pH of feeds, permeates and retentates were above or equal to 7.
4.3 Distribution between permeate, retentate, feed and membrane
The permeate from the NF generally contained lower amounts of DM (25-70% lower), TOC (60-90% lower), COD (15-75% lower), BOD7 (35-80% lower), P (85-99%
lower), S (90-98% lower), NH4+-N (65-95% lower) and N (60-95% lower), than the
retentate, see Figure 13 (for the masses see Appendix 3). However, higher amounts of COD (80% higher) and NO2--N were found in the permeate of NFW membranes
(purple bars), along with higher amount of NO3--N (55-65 % higher) in the both
14
Figure 13. The distribution of the matter from the feed to the permeate and retentate. Sample names beginning with ‘P’ is the permeate of that filtration, and sample names beginning with ‘R’ are the retentate of that filtration. Sample names ending with ‘NFW’ or ‘DL’ shows if the sample was obtained from NF with NFW or DL membrane. The error bars of DM are the ±2 standard deviation and the error bars of the other analytes are the measurement uncertainty with a level of confidence of approximately 95%. A: HTC-water used as feed. Neither NO2- nor NO3
-were detected in the HTC-waters. B: WO-water used as feed.
The fouling on the membranes when filtering WO-water is mostly caused by BOD7,
sulphur- and nitrogen-containing compounds, as the differences between the amounts in the feed and the combined amounts in the permeate and retentate were positive, see purple and green bars in Figure 14. Filtration of HTC-water did not show a clear allocation of any compound. According to Figure 14 15% more NH4+ was detected in
the permeate and retentate than in the feed, of HTC filtration with DL membrane (orange bar).
A
15
Figure 14. The mass balance of the filtrations, calculated by subtracting the mass in the feed with the masses in the permeate and retentate. Mass balance from NF of HTC-water with NFW membrane are red, and orange with DL membrane. Mass balance from NF of WO-water with NFW membrane are purple, and green with DL membrane. The error bars are the measurement uncertainty with a level of confidence of approximately 95%.
4.4 Concentrations
For all membranes, and all types of feed, the permeates contained lower concentrations of the analysed compounds and elements compared to the retentates (Figure 15), aside from DM in the WO permeate and COD in the WO NFW permeate. Additionally, there is no notably difference between the membranes’ ability to increase the concentration in the retentate. Except for the higher enrichment of NH4+-N in the
retentate from HTC-water, there is no substantially affect of the feed used in the concentrations obtained in the retentates.
16
Figure 15. The distribution in permeate and retentate of the feeds’ concentrations for each parameter, DM, TOC, COD, BOD7, P, S, N NO2--N, NO3--N and NH4+-N. The concentrations of the feeds are 1 (blue lines). Sample names
beginning with ‘P’ is the permeate of that filtration, and sample names beginning with ‘R’ are the retentate of that filtration. Sample names ending with ‘NFW’ or ‘DL’ shows if the sample was obtained from NF with NFW or DL membrane. The error bars of DM are the ±2 standard deviation and the error bars of the other analytes are the measurement uncertainty with a level of confidence of approximately 95%. A: Permeate and retentate from HTC filtration. Neither NO2- nor NO3- were detected in the HTC samples. B: Permeate and retentate from WO filtration. The relationship between the BOD7 and the COD in the permeate can be seen in
Figure 16. The BOD7:COD-ratio of the HTC-water was 0.67. After WO, the
BOD7:COD-ratio increased by 75%, to 1.16, see the blue bars in Figure 16. The
BOD7:COD-ratio of HTC-, WO-water, all retentates and the permeates after WO were
all above 0.43, which indicates easy degradation of the compounds in the water. The larger the BOD7:COD-ratio, the larger the biodegradable fraction (i.e. biodegradable
within 7 days).
A
17
Figure 16.The ratio of the BOD7:COD for the feeds (HTC- and WO-water), the permeates and retentates of the
filtrations through the two membranes (NFW and WO). The limit of where a compound can be considered easily degraded is at 0.43 and is marked by the blue line. The error bars are the measurement uncertainty with a level of confidence of approximately 95%.
4.5 WO vs. NF vs. WO+NF
From the concentrations of the analysed compounds and elements (DM, TOC, COD, BOD7, P, S, N, NO2--N, NO3--N and NH4+-N) a comparison between the WO’s, and
the NFs’ ability to decrease the target analytes from the HTC-water can be made (Figure 17). This is useful for determining whether only WO, only NF or a combination of WO and NF performs better in decreasing the target compounds and elements. Figure 17 shows the concentrations of the analysed organic and nitrogen-containing compounds along with phosphorus- and sulphur-content, in WO-water and the NF permeates, as a fraction of the concentrations in the ingoing HTC-water. The orange- and green-coloured bars, representing the DL membrane, are, in most cases, notably lower than the red- and purple-, and blue-coloured bars, representing the NFW membrane and WO. The TOC, BOD7, P, S, and N fractions of the HTC concentrations
in the DL permeates were 35, 33, 1.5, 3.3 and 14%, respectively, of the WO fractions, and 49, 46, 9, 34 and 51%, respectively, of the NFW permeate fractions.
NF with DL alone is more efficient in removing the target analytes (P, NO2--N, NO3-
-N and -NH4+-N) than WO and NF combined. However no clear difference was seen for
DM, TOC, COD, BOD7 and S, see the orange and green bars in Figure 17, which also
18
Figure 17. A comparison between the concentrations of WO-water, the permeates of filtered HTC-water and filtered WO-water, depicted as the quota of the HTC-water. Bars above the dark blue line have larger concentrations than the HTC-water. No NO2-, nor NO3-, were present in the HTC-water. The error bars of DM are
the ±2 standard deviation and the error bars of the other analytes are the measurement uncertainty with a level of confidence of approximately 95%.
5 Discussion
5.1 Wet oxidation
The WO treatment changes the concentrations and composition of the compounds in the HTC-water. A change of colour, from dark brown to brown/orange, in the WO-water compared to the HTC-WO-water (Figure 5) implies that some form of change has occurred. By consulting the UV-vis spectra of the HTC- and WO-samples, a change in the composition of the HTC-water during the WO can be clearly seen (Figure 9, Figure 10, Figure 11 and Figure 12). In all of the HTC-samples elution is observed early, from 0.8 minutes, and late, at 4.3 minutes, which indicates that all samples contained polar as well as non-polar compounds. The majority of the compounds elute at the earlier retention times, thus being more polar.
19 The polar compounds, with retention times between 0.7 – 1.5 minutes in the LC-MS, provided several mass spectra. However, because of co-elution from the LC, the spectra contained, in many cases, several compounds. This, in combination with an unknown composition of the feeds, the compounds in the mass spectra could not be certainly assured. Nevertheless, a decrease in concentrations of DM 50%), TOC (-54%), COD (-65%), BOD7 (-40%), P (-26%) and N (-21%) from HTC-water to
WO-water could be seen (Figure 17). This is due to the oxidation and degradation of these organic, nitrogen- and phosphorus-containing compounds. This coincide with other studies, that have reported a decrease of TOC, COD, and BOD by 35, 50 and 20%, respectively, by WO (5,7).
The amount of oxidizable matter in the HTC-water was degraded to a larger extent than the biological oxidizable matter, thus giving a 70% higher BOD7:COD-ratio in
the WO-water compared to the ratio in HTC-water (Figure 16 and Figure 17). This indicates that the compounds remaining in the WO-water are more easily degraded than the compounds in the HTC-water. Other studies have also reported an increased BOD7:COD-ratio after WO; 44% for WO of HTC-water (5) and 200% for
Fe-catalysed WO of sucrose (7).
The small nitrogen containing compounds, NO2- and NO3-, were present in the
WO-water but not in the HTC-WO-water and the concentration of NH4+ increased with 165%
after WO, see Figure 17. This increase of the small nitrogen-containing compounds can be explained by oxidation of other nitrogen containing compounds to NO2-, NO3
-and NH4+. The concentration increase was very different; for the different compounds;
the NH4+ concentration increased approximately 2500 and 100 times more than the
concentrations of NO2- and NO3-, respectively. A difference in concentration change
between NH4+ coincide with the literature, which reported a 80% increase of NH4+,
and no detected increase of NO2- or NO3- (5). Since an oxidation is performed the
concentration increase would be anticipated to be largest for NO2- and NO3-. As this is
not the case, the oxidation of nitrogen is not complete in the conditions of the WO. Furthermore, the distribution of nitrogen in different compounds changes after WO (Appendix 3). Before WO the N in the HTC-water were divided between NH4+ and
other compounds, 30:70. After WO, the N distribution in the water changed, to NH4+
containing almost 100% of the N present in the water. This shows that the compounds containing N before the WO have been degraded and formed NH4+.
Several studies have reported an increase of organic acids after WO, where acetic acid and formic acid are the most abundant acids (5,7,8). However, the samples in this study did not show signals of containing acids. The MS was test-run in negative mode with a mobile phase with pH=8, in which carboxylic acids are prominent, since they appear in their deprotonated form and are stable negative ions. No signals were appearing with these conditions (Appendix 2), which is an indication that there is no presence of acids. Furthermore, the pH was not below 7 for any of the feeds, permeates or retentates.
5.2 Nanofiltration
5.2.1 Separation of HTC-water
20 the percentage distribution charts of the UV-vis spectra from the LC (Figure 10). The permeate and retentate contains both polar and non-polar compounds, which eluted between 0.7 – 4.7 minutes. However, the peak areas of the permeates from the HTC-filtrations differ greatly and there is no pattern between the relationship of permeate and retentates, thus nothing can be said for the sizes of the polar and non-polar compounds.
On the other hand, size predictions of P-, S- and N-containing compounds can be made. The difference between the concentrations of P-, S- and N-containing compounds in the permeate and retentate is very large. The P-containing compound were 18 times higher in the NFW retentate and 260 times higher in the DL retentate compared to the permeate, 26 respectively 86 times higher for the S-containing compounds, and 13 respectively 31 times higher for the N-containing compounds. Thus, most P-, S- and N-containing compounds are larger than 500 Da. However, rejection of the sugars glucose, xylose and arabinose have been reported for the DL membrane and membranes similar to the NFW membrane (16). The molecular weight of these compounds ranges from 150 – 180 Da, thus rejection of molecules this small can be achieved by the membranes. Hence, compounds found in the retentate could be as small as 150 Da.
Mass spectra were only obtained for compounds eluting at a retention time between 0.7 – 1.5 minutes. In these mass spectra, reduction of peaks can be observed in the permeate, compared to spectra of the feed and retentate. Where peaks are missing one or more compounds are apparently removed by the membranes. Since the peaks have low retention time, these compounds are polar.
5.2.2 Separation of WO-water
NF of WO-water with DL membrane resulted in an apparent separation of all analytes, between the permeate and retentate (Figure 15). However, during filtration with the NFW membrane the concentrations of DM and COD were not substantially lower in the permeate compared to the retentate. Thus, the WO degrades the COD compounds, that in a greater extent are in the range of 300 – 500 Da and permeate through the NFW membrane, which contribute to the higher DM. However, with a ±2 standard deviation, the concentrations of DM in the permeate of the DL membrane is not substantially lower than the retentate. Even though all the other analytes show lower concentrations for the permeate than the retentate. For a lower level of confidence of the DM, approximately 68% with a ±1 standard deviation (Appendix 5), there is a notable lower concentration in the permeate compared to the retentate from the DL membrane. This coincides with the low permeate concentrations of the other analytes. Moreover, differences can be seen between the membranes’ separation abilities in the percental distribution of the integrated peaks of the UV-vis spectra (Figure 12). After WO the separation is most prominent for separating compounds with different polarities, where only polar compounds are found in the permeate, thus the membrane rejects the less polar compounds. Additionally, the integral for the P.WO.NFW peak is largest at a retention time of 1.0 (Figure 12). Thus, it is likely that the higher amount of the COD found in the permeate from the NFW membrane, as stated in the previous paragraph, is found at this retention time. However, the UV-vis only detects chromophore compounds, which not all COD compounds necessarily are.
21 After WO less nitrogen-containing compounds are rejected to the retentate, compared to before WO, see the striped bars of the N-category in Figure 15. Since the nitrogen-containing compounds are degraded into NH4+, and to a small extent NO2- and NO3-,
during the WO, a higher total nitrogen content is observed in the permeate, due to a higher amount of NH4+, NO2- and NO3- permeating.
Although there is no notably difference of the BOD7:COD-ratio between the
membranes and the feeds used (Figure 16), the ratio in the permeates increased after WO to be above the limit of 0.43. Where sample with a ratio below 0.43 is considered to be difficult to degrade (19). The NFW and DL membranes gave ratios of 0.66 and 0.60, respectively, for the permeates from the WO, similar to the ratio found by Larsson et al. after WO of sludge (5). Both feeds and retentates, from all filtrations, provided ratios higher than 0.43 which implies easy degradation. However, the concentrations of COD and BOD7 were higher in these samples compared to the
permeates (Figure 15), which still demands more from the environment for degradation.
5.2.3 Flux and Fouling
For all filtrations, the pressure of the feed decreased as the filtration proceeded (Figure 8, Table 1). The highest decrease was found when filtering WO-water through the DL membrane (70 psi drop from 220 psi). As seen in Figure 7, this membrane also has the highest amount of fouling of the membranes. When consulting the mass balance chart, (Figure 14) there is only a clear accumulation on the membranes of BOD7 and
nitrogen-containing compounds for filtration of WO-water. There is no observable difference between the membranes. Thus, the decrease in pressure when filtering WO-water through the DL membrane is likely caused by fouling on the membrane, at least by BOD7 and nitrogen-containing compounds. Considering the visibly higher amount
of foulant on the DL membrane (Figure 7), it is likely that more compounds are present on the membrane, than what can be concluded from Figure 13.
At the prevailing pH the membranes have a negative charge, thus positively charged compounds should be attracted to the membrane and permeate easier. The positively charged NH4+ is not found in higher concentrations in the permeate than in the
retentate (Figure 15). Additionally, a larger fraction of the negatively charged nitrogen compounds, NO2- and NO3-, were found in the permeate, compared with the fraction of
NH4+. Thus, it is questionable whether the membranes were negatively charged, as that
should have attracted NH4+ and repelled NO2- and NO3-. However, the mass balance of
the permeate and retentate to the feed shows that for NF of WO-water, some NH4+ is
found on the membrane, while the mass balance for the DL membrane in the HTC-filtration, do not show any fouling of NH4+. The much larger amount of NH4+ in the
retentate might be caused by an attraction of NH4+ to some other, negative, compound,
22 increasing the pressure of the feed, while an unrestored pressure would allow decrease of the flux.
The flux was higher for the DL membrane, compared to the NFW membrane, 100 – 150 l·m-2·h-1 and 20 – 35 l·m-2·h-1, respectively (Figure 8). This contradicts the literature, which states that the flux of a membrane with lower MWCO is lower, compared to a membrane with higher MWCO (12). In one study (14), the flux through the DL membrane was found to be 100 l·m-2·h-1 at a pressure of 220 psi. Thus, if the flux through the DL membrane should be lower than the NFW membrane, the flux trough the NFW membrane should be higher than 100 l·m-2·h-1.
Due to the difference in flux, the NF with the two membranes took different time. The DL membrane finished filtering 1 litre after 20 minutes, while the same amount of sample took 85 minutes with the NFW membrane. The membranes had different recommended working pressure, where the membrane with the lower MWCO (DL) had higher pressure (220 psi) and the membrane with higher MWCO (NFW) had lower pressure (110 psi). This is believed to be the cause of the flux difference and consequently the time difference.
5.3 Most effective reduction
WO reduced the concentrations of DM, TOC, COD, BOD7, P and N, and increased the
ratio of BOD7:COD from the HTC-water (Figure 17), as did the membranes. However,
the membranes performed better than only WO, in decreasing the concentrations of compounds and elements in the water. The reduction obtained with the DL membrane was notably better compared to the NFW membrane.
In the cases where there was a clear difference between the membranes in permeate concentrations, the permeate from the NFW membrane always contained higher concentrations, independent of the feed used (Figure 15). This implies that the DL membrane provides cleaner permeate than the NFW membrane. This is not unexpected, considering the lower MWCO of the DL membrane; 150 – 300 Da. When comparing the concentrations in the permeate of the best performing membrane (DL) with HTC-water treated only by WO, the filtration gives lower concentrations, see Figure 17 (and Appendix 4). There is only an apparent difference in the reduction of nitrogen-containing compounds between the use of HTC- and WO-water as feeds for the NF. By using HTC-water directly in NF, without previous WO treatment, the reduction of NO2-, NO3-, NH4+, N was 100, 100, 60 and 50% better than NF,
respectively compared with previous WO treatment. For the other analysis (DM, TOC, COD, BOD7, P and S) no obvious difference was found between performing WO
before NF with the DL membrane.
Since the WO oxidises and degrades the molecules in the HTC-water, there is a higher amount of smaller molecules in the WO-water, which are not rejected by the membranes. Thus, the concentrations in the permeates do not decrease after WO. This is supported by the BOD7:COD-ratio (Figure 16), which indicates the difficulty in
degrading the compounds in the sample. The ratio increases after WO, which indicates that the compounds that were not available for biodegradation before WO, has subsequently been partly degraded and after WO treatment more available for biodegradation.
23 DL membrane, than to only conduct WO treatment of the HTC-water. Using the DL membrane without WO is slightly better or equal to using the DL membrane after WO.
5.4 Contribution to society, industry and research
Both membranes resulted in lower concentrations of the target analytes, were the DL membrane was most efficient. Since their function as a water cleaning technique was efficient, the use of NF, with or without WO, contributes to the SDG: Any organic pollutions that are degraded by WO or rejected by the membranes contributes to goal 6.3; Improve water quality, wastewater treatment and safe reuse and the Swedish environmental objective of A non-toxic environment. The SDG of Upgrade all
industries and infrastructures for sustainability (goal 9.4), can also be contributed to
by WO and NF. The operation of the NF technology is clean for the environment, as it is a mechanical, non-chemical, technology, that does not in itself cause any pollutions. Only electricity is needed to drive the pump. However, the production of the membranes and the equipment do entail environmental impact. Nonetheless, the membranes and the equipment can be reused, the environmental impact per filtration is lower. The WO only consumes O2 and release CO2, although being a greenhouse gas it
is not an environmental hazardous compound. Furthermore, if no treatment of the HTC-water took place, the degradation of the organic matter in the HTC-water would be shifted to occur in lakes, rivers or seas, where it also would produce CO2.
Incorporating the WO and/or NF technology into the HTC process, would upgrade the sustainability of that process. The SDG 11.6 Reduce the environmental impact of cities is already contributed by the HTC process of the sewage sludge, but needs itself further processing to be clean and have a process water that is safe to release into the environment. The additional steps of WO and NF can provide that. Lastly, this study targets wastewater, that without treatment and release to nature, could end up in seas and with a content of organic, nitrogen- and phosphorous-containing compounds contribute to nutrient pollution and eutrophication. Therefore, the treatment with WO and/or NF contributes greatly to this cause, as the levels of organic and phosphorous-containing compounds decreases with both treatments, and nitrogen-phosphorous-containing compounds (mostly NH4+) decrease especially with NF treatment. This targets the
SDG 14.1 Reduce marine pollution and the Swedish environmental objective of Zero
eutrophication.
If an industrial, or municipal wastewater treatment plant, have, or were to consider, incorporating HTC as a upgrading method for their sludge and do not have a solution for the treatment of the toxic and organic matter in HTC-process water (7) before sending it to internal or external wastewater treatment facilities, NF, with or without previous WO, is a technique that can provide reduction of TOC, COD, BOD7 and
nitrogen-, phosphorous-, and sulphur-containing compounds. By performing WO treatment, energy to drive the HTC process can be obtained, as stated in the literature (5,7). However, a previous WO does not improve the effect of the membranes’ ability reject TOC, COD, BOD7 and nitrogen-, phosphorous-, and sulphur-containing
compounds. The DL membrane, with the lower MWCO, provided both better rejection and faster filtration, which would be favourable in an industrial context.
24 studies of this field. Furthermore, more data on the aspects treated in this study would contribute to more certain conclusions, as this study was rather small, and analysis of all replicate were not performed for most of the analysis.
5.5 Future prospects
In order to further understand which compounds that are degraded through the WO and which are separated by the NF, LC-MS with higher resolution of the peaks, could support this. That would help avoid co-elution and MS spectra with several compounds present in one MS spectrum. Thus, enabling library search to identify the compounds the peaks represent. Furthermore, LC with size exclusion chromatography would provide information about the MW of the molecules that are separated by the membranes and how much the compounds in the WO are degraded.
To further improve WO, catalysts, such as Fe, could be used. By comparing the compositions before and after the WO, an understanding in the process occurring in the WO can be obtained.
The WO caused formation of precipitation. This was not specifically analysed, but by analysing its content a better understanding about what happens chemically in the WO could be gained. Furthermore, the precipitates were not removed before the NF. With remaining precipitate in the feed, a faster fouling may occur, as the membrane would be covered with it. Moreover, a layer of foulant can contribute to higher rejection by the membrane, thus it would be interesting to see whether removal of precipitate would provide a poorer separation by the membrane and/or a longer operation time. Depending on the content of the precipitate, different applications could be found for it. For instance, as fertilizer if it has high concentrations of NH4+ and phosphorous.
The NF was only run with 1 l feed at the time, always with a new membrane. In order to get a better understanding of the membranes’ effectiveness, filtrations with a larger amount of feed than 1 l and run until large enough fouling occurred to hinder the permeation would provide information of which membrane is the most preserved. With a longer filtration period, analysis of the permeate and retentate during the filtration could provide information about potential changes in the concentration, which might occur as a consequence of fouling. If concentration changes of the charge compounds occur, such as NO2-, NO3- and NH4+, a better understanding of the
membrane’s charge and its effect on the filtration could be revealed.
As a step towards large scale filtration, new feed could continuously be added to the retentate. This would mimic a continuous filtration and can be compared against batch filtration, as performed in this study. To further gain knowledge of industrial importance, treatment of fouling needs to be investigated. Such as backwashing and chemical cleaning.
25
6 Conclusions and Outlook
WO of HTC-water oxidises and degrades organic, nitrogen-, and phosphorous-containing compounds. The degradation of organic compounds results in lower DM, TOC, COD and BOD7 concentrations, with a reduction of 50, 54, 65 and 40%
respectively. The overall biodegradability increases as well, with an increased BOD7:COD ratio of 75%. The degradation of nitrogen-containing compounds led to,
almost exclusively, NH4+ formation. The NF with polyamide TFC flat sheet
membranes further reduced the concentrations of all analytes (DM, TOC, COD, BOD7,
P, S, N, NO2--N, NO3--N, NH4+-N). Both membranes’ (NFW and DL) reduction of the
concentrations in the original HTC-water, were, in most cases, without substantial difference between previous WO of the HTC-water. This is due to degradation of the compounds in the HTC-water, which in their degraded form more easily permeates through the membrane. The largest reduction of the HTC-water concentrations was obtained by the DL membrane, with a reduction of 45-63% for DM, 83-85% for TOC, 72-77% for COD, 80% for BOD7, 99% for P, 96-97% for S, 86-93% for N and
64-87% for NH4+, depending on whether HTC-water or WO-water was used as feed for
the filtration. From LC-MS analysis it was found that mostly polar compounds remained after WO, and to a higher extent was found in the permeate, than in the retentate. Filtration without WO showed no division between polar and more non-polar compounds in the permeate and retentate.
Acknowledgement
I would like to offer my special thanks to both my supervisors Gunnar Westin at RISE Processum AB, for asking the right questions to help me move forward, and Associate Professor Stina Jansson at Umeå University, for the eminent help with the academical formality, from aim to conclusion. The generous time they both have spent on me and this work has been very much appreciated. I would also like to thank the staff at RISE Processum AB; Jonna Almqvist for the excellent teaching of the WO process and the operation of the reactor, to Emma Johansson for all the advice about the nanofiltration, and to Jonas Fahrni and Shubhankar Bhattacharyya for their valuable help with the liquid chromatography, together with the rest of the staff, who always had time for my questions and made me feel right at home. I would like to extend my appreciation to Peter Axegård, Fredrik Öhman and Fredrik Björnerbäck at C-Green Technology AB, for sharing their knowledge in the subject of HTC and WO, it has been tremendously helpful. My family and friends, old and new, deserves my gratitude, for their support and in keeping me sane.
References
1. Tchobanoglous G, Burton FL, Stensel HD, editors. Wastewater engineering: treatment and reuse. 4th ed. Boston, USA: McGraw-Hill; 2003
2. Miljö- och energidepartementet. Hållbar slamhantering (SOU: 2020:3). Stockholm, Sweden: Nordstedts Juridik; 2020
26 4. Titirici M-M, editor. Sustainable carbon materials from hydrothermal processes.
Chichester, West Sussex, United Kingdom: Wiley; 2013.
5. Larsson S, Grimm A, Norlander J, Westin G, Gilenstam N, Söderberg C, et al. Effective carbonization of pulp and paper industry waste products - Phase 2. Umeå, Sweden: Skogens biomaterial och teknologi, Sveriges Lantbrukuniversitet; 2019. Report No.: 42372–2.
6. Funke A, Ziegler F. Hydrothermal carbonization of biomass: A summary and discussion of chemical mechanisms for process engineering. Biofuels, Bioprod Bioref. 2010; 4(2):160–77. doi: 10.1002/bbb.198
7. Riedel G, Koehler R, Poerschmann J, Kopinke F-D, Weiner B. Combination of hydrothermal carbonization and wet oxidation of various biomasses. Chem. Eng. J. 2015;279:715–24. doi: https://doi.org/10.1016/j.cej.2015.05.086
8. Suárez-Iglesias O, Urrea JL, Oulego P, Collado S, Díaz M. Valuable compounds from sewage sludge by thermal hydrolysis and wet oxidation - A review. Sci. Total Environ. 2017;584–585:921–34. doi: http://dx.doi.org/10.1016/j.scitotenv. 2017.01.140 0
9. Berge ND, Ro KS, Mao J, Flora JRV, Chappell MA, Bae S. Hydrothermal Carbonization of Municipal Waste Streams. Environ. Sci. Technol. 2011;45(13):5696–703.
10. Svärd M, Johansen T, Lewander M, editors. Havet: om miljötillståndet i svenska havsområden. Sweden: Naturvårdsverket, Havsmiljö-institutet; 2014. ISSN: 1654-6741.
11. Duke M, editor. Functional nanostructured materials and membranes for water treatment. Weinheim: Wiley-VCH GmbH & Co. KGaA; 2013.
12. Zhang TC, editor. Membrane technology and environmental applications. Reston, USA: American Society of Civil Engineers; 2012.
13. Staff A. Microfiltration and Ultrafiltration Membranes for Drinking Water (M17). Denver, USA: American Water Works Association; 2005.
14. Wang Z, Liu G, Fan Z, Yang X, Wang J, Wang S. Experimental study on treatment of electroplating wastewater by nanofiltration. J. Membr. Sci. 2007;305(1–2):185–95. doi: 10.1016/j.memsci.2007.08.011
15. Bellona C, Drewes JE. The role of membrane surface charge and solute physico-chemical properties in the rejection of organic acids by NF membranes. J. Membr. Sci. 2005;249(1–2):227–34. doi: 10.1016/j.memsci.2004.09.041
16. Nguyen N, Fargues C, Guiga W, Lameloise M-L. Assessing nanofiltration and reverse osmosis for the detoxification of lignocellulosic hydrolysates. J. Membr. Sci. 2015;487:40–50. doi: http://dx.doi.org/10.1016/j.memsci.2015.03.072
17. Harris DC, Lucy CA. Quantitative chemical analysis. 9th edition. New York, USA: Macmillan Education; 2016.
27 [cited 2020 Apr 16]. Available from: https://utslappisiffror.naturvardsverket.se/ en/Substances/Other-organic-substances/Total-organic-carbon/
19. Bingman I. Biologisk-kemisk karakterisering av industriavloppsvatten: tillämpning vid prövning och tillsyn av miljöfarlig verksamhet. Solna, Sweden: Statens naturvårdsverk; 1989. AR; 89:5.
20. United Nations. About the Sustainable Development Goals [Internet]. United Nations Sustainable Development; d.u. [cited 2020 Apr 16]. Available from: https://www.un.org/sustainabledevelopment/sustainable-development-goals/ 21. Sveriges miljömål. Sveriges miljömål [Internet]. Sveriges miljömål; d.u. [cited
28
Appendix
Appendix 1
The data of the filtrations. The data from samples HTC.NFW, HTC.DL, WO.NFW and WO.DL are mean values with standard deviation of the three performed replicates. The data about the HTC-water is the mean and standard deviation of 6 measurements. The pH of WO is the mean and standard deviation of 3 measurements and the 6 measurements for the volume. The value of pressure drop of WO.DL is the mean and average absolute deviation of 2 samples.
HTC HTC.NFW HTC.DL
Permeate Concentrate Permeate Concentrate
pH 7 ± 0 7 ± 0 7 ± 0 7 ± 0 7 ± 0
V/ ml 986.5 ± 14.5 668 ± 3 260 ± 14 660 ± 10 345 ± 65
WO WO.NFW WO.DL
Permeate Concentrate Permeate Concentrate
pH 7.33 ± 0.29 8 ± 0 8.17 ± 0.29 8 ± 0 8 ± 0
V/ ml 995.7 ± 4.6 595 ± 20 280 ± 33 570 ± 51 330 ± 49
HTC.NFW HTC.DL WO.NFW WO.DL
Time/ minute 85 ± 13 19 ± 2 81 ± 7 18.5 ± 0.5
29
Appendix 2
The UV-vis spectrum and the mass signal graph, of the negative mode MS, with an ammonia solution as the mobile phase (pH=8).
Appendix 3
30
Appendix 4
The concentrations of the analysed compounds and elements in the WO-water and the permeates of the DL membrane with and without previous WO. Where filtration provided notably better result the background is green, not a notably difference yellow and a notably worse result red. For filtrations with and without WO of HTC-water, that are better than only WO the HTC-water, the notably best concentrations are bold.
