w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 2 6 2 e2 7 0
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/watres
Static and dynamic removal of aquatic natural organic matter by carbon nanotubes Gaurav S. Ajmani a, Hyun-Hee Cho b, Talia E. Abbott Chalew a, Kellogg J. Schwab a, Joseph G. Jacangelo a,c, Haiou Huang a,d,* a
Center for Water and Health, Johns Hopkins University, Baltimore, MD 21205, USA SD Center, KOLON Water & Energy Co., LTD., Seocho-3dong, Seocho-gu, Seoul 137-870, South Korea c MWH, Lovettsville, VA 20180, USA d State Joint Key Laboratory of Environmental Simulation and Pollution Control, School of Environment, Beijing Normal University, Beijing 100875, China b
article info
abstract
Article history:
Carbon nanotubes (CNTs) were investigated for their capability and mechanisms to
Received 22 January 2014
simultaneously remove colloidal natural organic matter (NOM) and humic substances from
Received in revised form
natural surface water. Static removal testing was conducted via adsorption experiments
2 April 2014
while dynamic removal was evaluated by layering CNTs onto substrate membranes and
Accepted 12 April 2014
filtering natural water through the CNT-layered membranes. Analyses of treated water
Available online 24 April 2014
samples showed that removal of humic substances occurred via adsorption under both static and dynamic conditions. Removal of colloidal NOM occurred at a moderate level of
Keywords:
36e66% in static conditions, independent of the specific surface area (SSA) of CNTs. Dy-
Carbon nanotube
namic removal of colloidal NOM increased from approximately 15% with the unmodified
Static adsorption
membrane to 80e100% with the CNT-modified membranes. Depth filtration played an
Membrane filtration
important role in colloidal NOM removal. A comparison of the static and dynamic removal
Natural organic matter
of humic substances showed that equilibrium static removal was higher than dynamic
Drinking water
(p < 0.01), but there was also a significant linear relationship between static and dynamic removal (p < 0.05). Accounting for contact time of CNTs with NOM during filtration, it appeared that CNT mat structure was an important determinant of removal efficiencies for colloidal NOM and humic substances during CNT membrane filtration. ª 2014 Elsevier Ltd. All rights reserved.
1.
Introduction
Membrane technology for water treatment has grown rapidly due to shrinking water supply, deteriorating source water quality, and increasing drinking water demands. Low pressure membrane (LPM) filtration, characterized by operating
pressures below 1e2 bar, has drawn particular attention due to low energy demands and high microbial removal efficiency (Furukawa, 2008). Dissolved natural organic matter (NOM), consisting of humic materials, lipids, polysaccharides, proteins, organic acids, and other organic materials (Cheng et al., 2005), is a major target of membrane water treatment because of its relevance to public health and water treatment plant
* Corresponding author. State Joint Key Laboratory of Environmental Simulation and Pollution Control, School of Environment, Beijing Normal University, Beijing 100875, China. Tel./fax: þ86 10 5880 5768. E-mail address: [email protected] (H. Huang). http://dx.doi.org/10.1016/j.watres.2014.04.030 0043-1354/ª 2014 Elsevier Ltd. All rights reserved.
263
w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 2 6 2 e2 7 0
operation. Small aromatic humic substances have been a major concern for drinking water safety because they can react with chlorine to form disinfection byproducts (DBPs) potentially dangerous to human health (Singer, 1999). Since LPMs cannot effectively remove humic substances as they are smaller than membrane pores, feedwater to LPM filtration is often pretreated by adsorption to activated carbon (AC) or coagulation before membrane filtration (Huang et al., 2009; Stoquart et al., 2012). Alternatively, LPM filtrate may undergo further treatment through nanofiltration/reverse osmosis (RO) membranes before subsequent chlorination (Schafer, 2001). Large DOC constituents (herein termed colloidal DOC), including organic biopolymers such as polysaccharides and their clusters (MWH, 2012), are of concern primarily because of their role in membrane fouling. LPM fouling is the loss of membrane permeability due to the buildup of aquatic material, mostly colloidal DOC, on the membrane surface or within the pores (Howe and Clark, 2002; Lee et al., 2004). As a good carbon source for bacteria involved in biofilm formation, biopolymers in colloidal DOC are also related to the biofouling of RO membranes used in water treatment (Ridgway and Flemming, 1996). Thus, removal of colloidal DOC by LPMs will be important in applications such as a pretreatment for RO, in order to prevent biofouling of downstream membranes. Colloidal DOC has also been found to be associated with toxic waterborne compounds, so its removal may also be significant for human health (Buffard and Leppard, 1995). Carbon nanotubes (CNTs) have emerged as a novel carbonaceous material and adsorbent for NOM because of their hydrophobicity and high specific surface area (SSA) (Wang et al., 2008; Upadhyayula et al., 2009; Yang et al., 2011; Apul et al., 2012). CNTs also have several advantages over traditional AC. Their surface can be functionalized with negativelyor positively-charged functional groups. They also can undergo desorption much more readily than ACs, because their surface does not contain deeply embedded pores (Lu and Su, 2007; Hyung and Kim, 2008; Upadhyayula et al., 2009). Therefore, many studies have considered the ability of CNTs to adsorb heavy metals, NOM, and other organic matter such as pharmaceuticals from water under static conditions (Hyung and Kim, 2008; Wang et al., 2008; Cho et al., 2011; Upadhyayula et al., 2009; Yang et al., 2011). However, little research has evaluated the ability of CNTs to remove different NOM fractions under dynamic conditions, as would be the case during a filtration application. This is important because CNTs will preferentially be used as a fixed media than as suspended materials in full-scale drinking water treatment due to their environmental mobility and unknown human toxicity. In the past decade, fabrication of membranes with CNTs has been intensively explored for their ability to improve membrane permeability, reduce membrane fouling, and enhance membrane rejection of contaminants. In such studies, the removal of aquatic NOM will occur possibly via adsorption (CNT-impregnated or layered membranes) or sieving (aligned CNTs). A few studies have considered the ability of CNT-membranes to reject NOM (Celik et al., 2011; Shawky et al., 2011), but none have compared removal under static and dynamic filtration conditions to assess CNT
removal efficiency or considered removal mechanisms. This comparison is important to make, in order to fully understand how CNTs will behave in membrane applications and how to optimize the design of next-generation CNT membranes. An ideal membrane application would maximize the CNT capability as a sorbent under static conditions. In this study, dynamic removal of NOM was studied by natural water filtration through CNT-layered microfiltration (MF) membranes, as a follow-up to our previous study investigating the antifouling properties of such CNT-membranes (Ajmani et al., 2012). The objectives of this study were to: 1) build upon previous work examining flux and fouling behavior of CNT-layered membranes and investigate the removal efficiencies and mechanisms for major NOM fractions during filtration, 2) evaluate the capabilities of CNTs with different physical structures and SSAs to remove different NOM fractions during static adsorption, 3) determine the relationship between the maximum adsorption capacity of the different CNTs under static conditions and the removal efficiency under the dynamic conditions, and 4) based on these results, identify key parameters for the design of CNT membranes for NOM removal.
2.
Materials and methods
2.1.
Carbon nanotubes
Pristine single-walled and multi-walled CNTs (SWCNTs and MWCNTs) with different diameters and SSAs were purchased from Cheap Tubes Inc. (Vermont, USA). According to the manufacturer, they were prepared using chemical vapor deposition and then further purified. Physical characteristics of all CNTs studied (as reported by the manufacturer) are presented in Table 1, along with the abbreviations that will be used to refer to different CNTs.
2.2.
Water samples
The natural surface water used in the study was collected from the source water for the North Bay Regional Water Treatment Plant in California, USA. It was shipped to the lab overnight, prefiltered through a 1.2 mm glass fiber membrane (Whatman GF/C) and stored in the dark at 4 C. Important characteristics of this water have been previously presented (Ajmani et al., 2012). Most important for the purposes of this
Table 1 e Properties of single-walled carbon nanotube (SWCNT) and multi-walled carbon nanotubes (MWCNTs) used in this study, as reported by manufacturer. CNT CNT Outer Length Specific surface abbreviation structure diameter (mm) area (SSA) (m2/g) (nm) SWCNT 8CNT 10CNT 20CNT 30CNT 50CNT
SWCNT MWCNT MWCNT MWCNT MWCNT MWCNT
1e2
Available online at www.sciencedirect.com
ScienceDirect journal homepage: www.elsevier.com/locate/watres
Static and dynamic removal of aquatic natural organic matter by carbon nanotubes Gaurav S. Ajmani a, Hyun-Hee Cho b, Talia E. Abbott Chalew a, Kellogg J. Schwab a, Joseph G. Jacangelo a,c, Haiou Huang a,d,* a
Center for Water and Health, Johns Hopkins University, Baltimore, MD 21205, USA SD Center, KOLON Water & Energy Co., LTD., Seocho-3dong, Seocho-gu, Seoul 137-870, South Korea c MWH, Lovettsville, VA 20180, USA d State Joint Key Laboratory of Environmental Simulation and Pollution Control, School of Environment, Beijing Normal University, Beijing 100875, China b
article info
abstract
Article history:
Carbon nanotubes (CNTs) were investigated for their capability and mechanisms to
Received 22 January 2014
simultaneously remove colloidal natural organic matter (NOM) and humic substances from
Received in revised form
natural surface water. Static removal testing was conducted via adsorption experiments
2 April 2014
while dynamic removal was evaluated by layering CNTs onto substrate membranes and
Accepted 12 April 2014
filtering natural water through the CNT-layered membranes. Analyses of treated water
Available online 24 April 2014
samples showed that removal of humic substances occurred via adsorption under both static and dynamic conditions. Removal of colloidal NOM occurred at a moderate level of
Keywords:
36e66% in static conditions, independent of the specific surface area (SSA) of CNTs. Dy-
Carbon nanotube
namic removal of colloidal NOM increased from approximately 15% with the unmodified
Static adsorption
membrane to 80e100% with the CNT-modified membranes. Depth filtration played an
Membrane filtration
important role in colloidal NOM removal. A comparison of the static and dynamic removal
Natural organic matter
of humic substances showed that equilibrium static removal was higher than dynamic
Drinking water
(p < 0.01), but there was also a significant linear relationship between static and dynamic removal (p < 0.05). Accounting for contact time of CNTs with NOM during filtration, it appeared that CNT mat structure was an important determinant of removal efficiencies for colloidal NOM and humic substances during CNT membrane filtration. ª 2014 Elsevier Ltd. All rights reserved.
1.
Introduction
Membrane technology for water treatment has grown rapidly due to shrinking water supply, deteriorating source water quality, and increasing drinking water demands. Low pressure membrane (LPM) filtration, characterized by operating
pressures below 1e2 bar, has drawn particular attention due to low energy demands and high microbial removal efficiency (Furukawa, 2008). Dissolved natural organic matter (NOM), consisting of humic materials, lipids, polysaccharides, proteins, organic acids, and other organic materials (Cheng et al., 2005), is a major target of membrane water treatment because of its relevance to public health and water treatment plant
* Corresponding author. State Joint Key Laboratory of Environmental Simulation and Pollution Control, School of Environment, Beijing Normal University, Beijing 100875, China. Tel./fax: þ86 10 5880 5768. E-mail address: [email protected] (H. Huang). http://dx.doi.org/10.1016/j.watres.2014.04.030 0043-1354/ª 2014 Elsevier Ltd. All rights reserved.
263
w a t e r r e s e a r c h 5 9 ( 2 0 1 4 ) 2 6 2 e2 7 0
operation. Small aromatic humic substances have been a major concern for drinking water safety because they can react with chlorine to form disinfection byproducts (DBPs) potentially dangerous to human health (Singer, 1999). Since LPMs cannot effectively remove humic substances as they are smaller than membrane pores, feedwater to LPM filtration is often pretreated by adsorption to activated carbon (AC) or coagulation before membrane filtration (Huang et al., 2009; Stoquart et al., 2012). Alternatively, LPM filtrate may undergo further treatment through nanofiltration/reverse osmosis (RO) membranes before subsequent chlorination (Schafer, 2001). Large DOC constituents (herein termed colloidal DOC), including organic biopolymers such as polysaccharides and their clusters (MWH, 2012), are of concern primarily because of their role in membrane fouling. LPM fouling is the loss of membrane permeability due to the buildup of aquatic material, mostly colloidal DOC, on the membrane surface or within the pores (Howe and Clark, 2002; Lee et al., 2004). As a good carbon source for bacteria involved in biofilm formation, biopolymers in colloidal DOC are also related to the biofouling of RO membranes used in water treatment (Ridgway and Flemming, 1996). Thus, removal of colloidal DOC by LPMs will be important in applications such as a pretreatment for RO, in order to prevent biofouling of downstream membranes. Colloidal DOC has also been found to be associated with toxic waterborne compounds, so its removal may also be significant for human health (Buffard and Leppard, 1995). Carbon nanotubes (CNTs) have emerged as a novel carbonaceous material and adsorbent for NOM because of their hydrophobicity and high specific surface area (SSA) (Wang et al., 2008; Upadhyayula et al., 2009; Yang et al., 2011; Apul et al., 2012). CNTs also have several advantages over traditional AC. Their surface can be functionalized with negativelyor positively-charged functional groups. They also can undergo desorption much more readily than ACs, because their surface does not contain deeply embedded pores (Lu and Su, 2007; Hyung and Kim, 2008; Upadhyayula et al., 2009). Therefore, many studies have considered the ability of CNTs to adsorb heavy metals, NOM, and other organic matter such as pharmaceuticals from water under static conditions (Hyung and Kim, 2008; Wang et al., 2008; Cho et al., 2011; Upadhyayula et al., 2009; Yang et al., 2011). However, little research has evaluated the ability of CNTs to remove different NOM fractions under dynamic conditions, as would be the case during a filtration application. This is important because CNTs will preferentially be used as a fixed media than as suspended materials in full-scale drinking water treatment due to their environmental mobility and unknown human toxicity. In the past decade, fabrication of membranes with CNTs has been intensively explored for their ability to improve membrane permeability, reduce membrane fouling, and enhance membrane rejection of contaminants. In such studies, the removal of aquatic NOM will occur possibly via adsorption (CNT-impregnated or layered membranes) or sieving (aligned CNTs). A few studies have considered the ability of CNT-membranes to reject NOM (Celik et al., 2011; Shawky et al., 2011), but none have compared removal under static and dynamic filtration conditions to assess CNT
removal efficiency or considered removal mechanisms. This comparison is important to make, in order to fully understand how CNTs will behave in membrane applications and how to optimize the design of next-generation CNT membranes. An ideal membrane application would maximize the CNT capability as a sorbent under static conditions. In this study, dynamic removal of NOM was studied by natural water filtration through CNT-layered microfiltration (MF) membranes, as a follow-up to our previous study investigating the antifouling properties of such CNT-membranes (Ajmani et al., 2012). The objectives of this study were to: 1) build upon previous work examining flux and fouling behavior of CNT-layered membranes and investigate the removal efficiencies and mechanisms for major NOM fractions during filtration, 2) evaluate the capabilities of CNTs with different physical structures and SSAs to remove different NOM fractions during static adsorption, 3) determine the relationship between the maximum adsorption capacity of the different CNTs under static conditions and the removal efficiency under the dynamic conditions, and 4) based on these results, identify key parameters for the design of CNT membranes for NOM removal.
2.
Materials and methods
2.1.
Carbon nanotubes
Pristine single-walled and multi-walled CNTs (SWCNTs and MWCNTs) with different diameters and SSAs were purchased from Cheap Tubes Inc. (Vermont, USA). According to the manufacturer, they were prepared using chemical vapor deposition and then further purified. Physical characteristics of all CNTs studied (as reported by the manufacturer) are presented in Table 1, along with the abbreviations that will be used to refer to different CNTs.
2.2.
Water samples
The natural surface water used in the study was collected from the source water for the North Bay Regional Water Treatment Plant in California, USA. It was shipped to the lab overnight, prefiltered through a 1.2 mm glass fiber membrane (Whatman GF/C) and stored in the dark at 4 C. Important characteristics of this water have been previously presented (Ajmani et al., 2012). Most important for the purposes of this
Table 1 e Properties of single-walled carbon nanotube (SWCNT) and multi-walled carbon nanotubes (MWCNTs) used in this study, as reported by manufacturer. CNT CNT Outer Length Specific surface abbreviation structure diameter (mm) area (SSA) (m2/g) (nm) SWCNT 8CNT 10CNT 20CNT 30CNT 50CNT
SWCNT MWCNT MWCNT MWCNT MWCNT MWCNT
1e2
