Graphene Oxide Membranes in Extreme Operating Environments

Subscriber access provided by UNIV OF CALIFORNIA SAN DIEGO LIBRARIES

Article

Graphene Oxide Membranes in Extreme Operating Environments: Concentration of Kraft Black Liquor by Lignin Retention Fereshteh Rashidi, Nikita S. Kevlich, Scott Sinquefield, Meisha L. Shofner, and Sankar Nair ACS Sustainable Chem. Eng., Just Accepted Manuscript • DOI: 10.1021/ acssuschemeng.6b02321 • Publication Date (Web): 09 Dec 2016 Downloaded from http://pubs.acs.org on December 11, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Sustainable Chemistry & Engineering is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Sustainable Chemistry & Engineering

Graphene Oxide Membranes in Extreme Operating Environments: Concentration of Kraft Black Liquor by Lignin Retention Fereshteh Rashidi1, Nikita S. Kevlich1,2, Scott A. Sinquefield2, Meisha L. Shofner2,3 and Sankar Nair1,2*

1

School of Chemical & Biomolecular Engineering, Georgia Institute of Technology, Atlanta, GA 311 Ferst Dr NW, Atlanta, GA 30332-0100 2

Renewable Bioproducts Institute, Georgia Institute of Technology, Atlanta, GA 500 10th Street NW, Atlanta, GA 30332

3

School of Materials Science and Engineering, Georgia Institute of Technology, Atlanta, GA 771 Ferst Drive NW, Atlanta, GA 30332-0245

*

Corresponding author: [email protected]

Keywords: black liquor, lignin, nanofiltration, graphene, pulp and paper, sustainability

Abstract We report a detailed study of graphene oxide (GO) membranes for concentration of Kraft black liquor (BL), which is a caustic (pH ~12), hot (80-95°C), and high-volume (~500 gal/min in a typical pulp mill) by-product of the papermaking process. Membrane-based concentration of BL is attractive as an energy-efficient alternative to thermally driven evaporation processes, but challenging due to the harsh operating conditions and high fouling potential of BL (15-18 wt% solids). We fabricate thin (< 300 nm) GO membranes supported on macroporous poly(ethersulfone) (PES) supports by vacuum filtration techniques, and discuss in detail their morphology, structure, thermmechanical stability, and chemical stability as characterized by several techniques. Furthermore, detailed permeation measurements at transmembrane pressures 1 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 27

(TMPs) up to 50 bar and temperatures up to 85°C show that the membranes have high performance in concentrating BL feeds containing high and low TS (total solids): high flux (in the range of 10-55 kg m-2 h-1), high lignin rejection (up to 98%), low fouling, and high stability throughout extended exposure (30 days) to BL at realistic operation conditions. The molecular weight cut-off (MWCO) of the membranes was determined to be ~625 Da by means of dye rejection experiments. The present membranes are also expected to have low cost due to the use of relatively inexpensive functional membrane and substrate materials (GO and PES).

Introduction The Kraft process is used worldwide for pre-treatment (pulping) of biomass to produce cellulose feedstock for papermaking or for conversion to chemicals and materials (biorefining). In the Kraft process, biomass is treated with a highly caustic solution of NaOH and Na2S to release the cellulose and delignify the wood.1-3 Weak black liquor (WBL) is the byproduct of this process, and is a corrosive fluid containing up to 18 wt% total solids (TS) at pH~12 and temperatures of 80-95°C. The solids present in WBL (Table S1, Supporting Information) are mainly lignin and inorganics, with significant concentrations of hemicelluloses and organic chemicals such as hemicelluloses/hydroxyacids. To recycle WBL, it is concentrated by water removal in a six-stage evaporator train to produce strong black liquor (SBL, up to 75-80 wt% TS4). SBL is used as fuel for generating heat and power. The inorganics are recovered in the form of ash from SBL combustion. More than 500 million tons/year of BL are generated worldwide, and the concentration of WBL to SBL by evaporation is highly energy-intensive (about 0.2 Quads/year in the US alone).5

2 ACS Paragon Plus Environment

Page 3 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Energy-efficient BL concentration is hence an important industrial and environmental issue affecting the sustainability of the forest products sector.5 Membrane-based BL concentration up to 30 wt% solids (thereby replacing the first two evaporators) is attractive for significantly reducing energy usage. A key challenge is the development of a long-lived, lowcost membrane that can operate with high flux and lignin rejection while withstanding the extreme combination of highly alkaline pH and the relatively high temperature of BL. Since the MW of lignin is in the range of 0.5-10 kDa, nanofiltration (NF) membranes with effective pore size ~ 1 nm are required. A number of reports (summarized in Refs.5-8 and references therein) have highlighted the challenges faced by membranes in BL processing. Conventional polymeric membranes can provide good lignin rejections (up to 97%)6 but are generally challenged by poor BL stability/low lifetimes. high fouling, and fairly high cost. Ceramic membranes are stable in BL but are challenged by high cost and lower lignin rejections. The best ceramic NF membranes currently lack a small enough pore size and cannot allow lignin rejections higher than 80%.6

In this work, we investigate the properties of graphene oxide (GO) membranes supported on appropriate macroporous polymer supports for BL concentration applications. Graphenebased membranes by a number of groups for applications such as nanofiltration (NF)9,10 and desalination by reverse osmosis (RO).11 Thin film graphene and GO membranes have unique attributes such as a pseudo-2D structure, high flexibility, and mechanical strength.12-14 GO is a chemical derivative of graphene with large number of oxidized functional groups on the surface, and has been great of interest for development of large surface area GO thin film membranes.15 Promising performance of GO membranes has been demonstrated in certain gas separation and fluid filtration (NF and RO) applications, in terms of high water fluxes and significant rejections



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 27

of organic and inorganic solids.16,17 However, nothing is known regarding the characteristics of GO membranes in extreme operating environments such as in BL concentration. We hypothesize that GO surfaces should be highly stable in BL conditions and that the nanoscale spacing between the GO sheets would allow high water flux and good lignin rejection. We then select macroporous poly(ethersulfone) (PES) as a support material for the functional GO layer, owing to its good acid and alkali resistance, thermal and mechanical stability, and low cost. We then fabricate GO membranes on macroporous PES supports by vacuum filtration, characterize their structure in detail, and evaluate their performance for the concentration of softwood Kraft BL under operating conditions of direct relevance to the Kraft process.

Experimental Methods Materials Fine grade synthetic graphite laminate powder (particle size < 20 µm), sulfuric acid (98% H2SO4), phosphoric acid), hydrochloric acid (HCl), potassium persulfate (K2S2O8), phosphorus pentoxide (P2O5) and potassium permanganate (KMnO4) were obtained from Sigma-Aldrich. Deionized (DI) water for GO synthesis was produced with a Thermo Scientific 7128 deionization system. For water permeation measurements, we used DI water purified in an RO plant and supplied to the building. All materials were used without further purification. Stock softwood BL (TS 30 wt% with pH~12.6 was donated by the Renewable Bioproducts Institute and obtained from industrial sources. Table S1 shows the composition of the BL feed.

GO Laminate Synthesis


Page 5 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


GO was synthesized from fine grade synthetic graphite laminates of < 20 µm size, by a modification18 of the Hummers method.19 In the original Hummers method, pre-oxidation of graphite was recommended in order to avoid the formation of incompletely oxidized graphitecore/GO-shell particles. For this pre-treatment, 3 g of graphite powder was added into 12 mL of 98% H2SO4 at 80°C while stirring. Then 2.5 g K2S2O8 and 2.5 g P2O5 were gradually added, and the mixture was maintained at 80°C for 4 h under stirring. It was then cooled to room temperature, diluted with 500 mL DI water, filtered through a filter paper with pore size of 2 µm, washed with DI water to remove excess acid and oxidizing agents, and dried at room temperature. The pre-treated graphite powder was added into 120 mL H2SO4 at 0°C. Then 15 g KMnO4 and 250 mL DI water were then gradually added, with an ice bath being used to keep the temperature below 50°C. After stirring for 2 h, another 700 mL of DI water was added. In the final step, 20 mL of 30% H2O2 was added gradually until the mixture turned yellow due to reduction of MnO4- to water-soluble Mn2+. The mixture was then filtered through a filter paper with pore size of 2 µm, washed several times with DI water, then washed with 1 M aqueous HCl, and finally washed with DI water until the pH became neutral (7). The final GO powder was obtained after drying at room temperature for 24 h. Figures 1a-1d illustrate the main steps of the above procedure.

Fabrication of GO Membranes Macroporous PES films (purchased from Sterlitech) of thickness 140 µm, pore size of 0.2 µm, and porosity of 75% were used as supports for fabricating GO membranes. First, the GO powder was dispersed in DI water (2 mg GO/mL water) and sonicated for 2 hr to exfoliate the GO laminates followed by centrifugation to remove unexfoliated GO. Then, GO membranes



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 27

were deposited on the PES supports from the aqueous GO suspensions using a vacuum filtration method (Figures 1a-1d). The PES support was placed on a Millipore filter (90 mm diameter and 160-250 µm pore size) with the permeate side connected to a vacuum pump, and the GO suspension was permeated through the assembly by the vacuum driving force resulting in trapping and deposition of the exfoliated GO laminates on the PES support. The thickness of the GO membrane can be controlled by the volume and concentration of the GO suspension. The asmade GO membranes were dried at 50°C for 24 h.

Figure 1. Schematic process for fabrication of a thin (< 300 nm) defect-free GO membranes on a macroporous poly(ethersulfone) support.

Characterization The interlayer spacing (d-spacing), morphology, thickness, and lateral dimensions of the GO laminates and membranes were characterized and determined by a combination of X-ray 6 ACS Paragon Plus Environment

Page 7 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


diffraction (XRD, PAnalytical XPert Pro with a CuKα source with 0.154 nm wavelength and operating in Bragg-Brentano geometry), scanning electron microscopy (SEM, Hitachi SU FESEM operated at 5 keV, 10 µA), atomic force microscopy (AFM, ICON Dimension scanning probe microscope Bruker, operated under tapping mode with Mikromasch NSC14 silicon cantilevers of 8 nm tip radius, 5 N.m-1 force constant, and 47-76 kHz resonance frequency), and dynamic mechanical analysis (DMA, Mettler Toledo DMA/SDTA861). For the DMA measurements, a thermal sweep was performed from 30-150°C with 2°C/min heating rate, 1 Hz frequency and 1 µm amplitude. The amplitude was verified to be in the linear viscoelastic range of the sample at the low and high temperature limits used for the measurement. The porosity of the PES support was measured by a Micromeritics Autopore IV mercury porosimeter.

Permeation Measurements The membranes were cut into circular sections of 47 mm diameter with effective membrane area of 13.1 cm2. A Sterlitech high-pressure dead-end stirred cell (HP4750X, 2500 psi maximum pressure rating) with an impeller assembly (mechanical Stirrer with SS 316 stir bar and shaft) was used for permeation measurements. A collection vessel placed on a digital mass balance (Ohaus Pioneer PA313) was used to collect the permeate liquid. The permeate mass, system temperature, and feed pressure were monitored and recorded digitally in real time on a computer via a LabVIEW interface. Nitrogen gas was used to pressurize the liquid on the feed side. Figure S1 (Supporting Information) shows a schematic of the permeation apparatus. After the membrane was mounted in the permeation cell, the feed side was filled with 300 ml of the feed liquid and the system was heated to the desired temperature (20, 70, or 85°C) using heating tape (McMaster Carr Extreme-temperature heat cable with 10' Length, 520 Watts, 120 VAC with



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 27

a plug). Both the heating tape and the thermocouples were connected to a temperature controller. In addition, the cell along with the heating tape was sealed with fiberglass to inhibit heat losses. Upon equilibration of the temperature, the transmembrane pressure (TMP) in the range of 1-50 bar were created by applying nitrogen pressure to the feed side through a regulator. The permeation measurement at each TMP was conducted with a fresh feed liquid. Prior to conducting the measurements with BL, at least two pure water permeation runs were performed at 50 bar TMP and 20°C to condition the membrane. The permeate flux at the given TMP was calculated from the steady-state portion of the permeate mass-versus-time data. During the steady-state period, 3 mL of permeate was collected separately for composition measurements. Since measurements conducted at different temperatures at which there is significant variation in BL physical properties, we maintained the Reynolds number (Re) to be approximately constant since it has the greatest influence on external mass transfer effects. Literature correlations20,21 specifically developed for BL were used to estimate its viscosity and density at different concentrations and temperatures. For the stirred-cell permeation module used in this study, Re =

ρd2N/µ where ρ is the solution density, µ is the viscosity, d is the impeller diameter (2.94 cm), and N is the impeller angular velocity (rps). For pure water permeation at 20°C, the impeller speed was 1.33 rps, and it was appropriately scaled for different feed viscosities and densities using the condition of constant Re (~6000).

Determination of Solute Rejections The observed (apparent) rejection of lignin was determined using the equation R = (1Cp/Cf)×100%, where Cp and Cf denote the permeate and bulk feed concentrations respectively.2226

The concentrations of lignin in the feed and permeate were determined with an Agilent 8510


Page 9 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


UV-Vis Absorption Spectrophotometer, via the integrated intensity of the characteristic peak at 280 nm. To avoid lignin precipitation (which occurs at lower pH), all UV-Vis dilutions and calibration solutions were made in a pH 13 buffer (glycine/sodium hydroxide/sodium chloride solution, Sigma-Aldrich). A quartz cuvette was used for all measurements. Although the rejection of inorganic solutes is not the main focus of this report, we measured a related quantity – namely the % conductivity reduction of the permeate relative to the feed –

using a

conductivity meter (Oakton CON 150).27 This method was adopted because BL contains a complex mixture28-30 of inorganics (Table S2), and a precise compositional analysis (by ICP-MS or ICP-OES) for all the permeate samples is time-consuming and prohibitively expensive. Since lignin macromolecules also carry charge and may influence the conductivity, these experiments were instead performed with a simulated (model) BL that was formulated based upon the composition shown in Table S2.

Determination of Molecular Weight Cut-Off (MWCO) The MWCO was determined as the molecular weight (MW) of a test dye molecule that has a rejection of 90%.23,31,32 This analysis was performed by measuring the rejection of several dye molecules with MWs in the range of 300-800 Da at 0.01 wt% dye concentration at room temperature over the same range of TMP used for BL concentration. The dyes used were Methyl Blue (800 Da, acidic), Congo Red (700 Da, basic), Amaranth (600 Da, acidic), Allura Red AC (500 Da, acidic), and Toluidine Blue O (300 Da, basic). Between all measurements, the membranes were cleaned with 1 N HCl solution after exposure to basic dyes and with 1 N NaOH solution after exposure to acidic dyes while stirring for 30 min at room temperature.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 27

Characterization of Stock BL The TS were determined by established standard protocols, specifically the TAPPI method T650.33 To determine inorganic content, a BL sample was heated at 105°C for at least 6 hrs to evaporate the water. The TAPPI method T625 was then used to remove the organic content by high-temperature (800-1000 °C) combustion and determine the inorganic content (as sulfated ash).34 The total organic content is the TS minus the inorganic content. Of the total organic content, the lignin concentration of a BL sample was determined by UV-Vis spectrophotometric measurement at 280 nm, pH > 13 using a 23.7 L g-1cm-1 extinction coefficient and 1 cm path length.35

Results and Discussion Several factors were considered in selecting an appropriate polymeric support material for the GO membranes. The physicochemical properties and the porosity of the polymeric support significantly affect the quality of the GO membrane as well as the overall stability of the supported membrane under BL exposure conditions. Polymers with aryl groups in their structural unit such as poly(sulfone) (PSF) and poly(ethersulfone) (PES) showed better thermal and chemical stability than those with alkyl groups such as poly(propylene) (PP) and poly(vinylidine fluoride) (PVDF). Although PSF and PES both have sulfone groups leading to an increase in stability under alkaline conditions, the repeat unit of (-O-aryl-SO2-aryl-O-aryl-SO2-aryl-)n in PES shows better stability than PSF whose repeating unit is (-O-aryl-SO2-aryl-O-aryl-alkanearyl-)n. PVDF is easily degraded in strong alkaline environments through interaction of the fluoride group with hydroxyl anions. PES (150°C) also has a higher thermal stability limit than PP (60°C) and PVDF (140°C). PES is more hydrophilic than PVDF, while PP is hydrophobic.


Page 11 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(a) 20 µm

(b)

(c)

Figure 2. Properties of the PES polymer support: (a) Pore size distribution by mercury porosimetry and (inset) SEM image, (b) dynamic mechanical analysis (DMA) at 30-150°C before and after immersion in 15 wt% BL at 85°C for 30 days, and (c) water permeation at 20°C before and after immersion in 15 wt% BL at 85°C for 30 days.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 27

The greater hydrophilicity of PES results in higher water flux and also better adhesion to the GO laminate membrane layer through interactions with the -COOH and -OH functional groups of the GO laminates. Figure 2 summarizes the measured properties of macroporous PES support films. The porosity was measured by mercury porosimetry, and shows an average pore size of ~500 nm with ~75% porosity (Figure 2a). The thermomechanical stability in the temperature range of 25-150°C was determined by dynamic mechanical analysis (DMA) on PES support samples before and after immersion in BL for 30 days at 85°C. Figure 2b shows that the PES support is thermomechanically stable, showing approximately constant storage modulus in the range of 30-150°C. In addition, chemical stability of the PES support was investigated by immersing it in 15 wt% TS stock BL at 85°C continuously for 30 days. Afterwards, the support was cleaned with a 1 M NaOH solution. The pure water flux of pristine and BL-immersed PES supports is shown in Figure 2c. It is seen that the permeation properties of the PES support are unaffected by BL exposure at higher temperatures. Similar measurements with PP and PVDF supports showed severe degradation upon BL exposure. As a result, we select macroporous PES supports for the fabrication of GO membranes in this study.

Figure 3 shows the permeate flux versus TMP through GO membranes at 70°C and 85°C using BL feeds with 0.3 wt% and 10 wt% TS. GO membranes allow high permeate fluxes (in the range of 20-50 kg m-2 h-1 at 35-50 bar TMP) comparable to the best commercial membranes evaluated in the literature.6 The maximum surface fluid velocity (defined as the product of the impeller angular velocity and the impeller radius) is in the range of 0.22-0.31 m/s, which is characteristic of the stirred-cell permeation apparatus used in this study. With the use of crossflow type modules that can achieve higher surface velocities, one may expect even higher fluxes.


Page 13 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. Permeate flux through GO membranes as a function of applied pressure (TMP) during BL-concentration under different operating conditions (0.3wt% and 10wt% TS BL at 70°C and 85°C). The maximum surface fluid velocity (defined as the product of the impeller rotational speed and the impeller radius) is in the range of 0.22-0.31 m/s.

As expected, operation with real BL leads to a significantly lower flux than with a diluted (0.3 wt% TS) BL, due to the additional concentration polarization (“fouling”) resistance created by the solids at the membrane surface. The chemical and physical properties of the membrane (pore size, pore morphology and hydrophilicity/hydrophobicity) affect the fouling/concentration polarization behavior.36 A high fouling resistance is usually expected to occur during BL concentration due to the high solids concentration. In previous literature using polymeric or ceramic membranes, the flux rapidly saturates as TMP is increased.6 However, the present GO membranes show an approximately linear increase in flux up to high TMP values. Our results for BL concentration are in good agreement with previously observed “anti-fouling” properties of 13 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 27

GO membranes during filtration of different organic dye molecules (methyl red, methylene blue, methyl orange, orange G, Tris (bipyridine) runthenium (II) chloride, Rhodamine B, Rose Bengal, methylene blue, brilliant blue) at concentrations of ~1 wt%.16 The hydrophilic surface of GO likely reduces interactions with relatively hydrophobic species such as lignin. The anti-fouling properties of the GO membranes also allow full recovery of flux after periodic cleaning, which is a key factor to increase the membrane lifetime. Finally, it is noted in Figure 3 that the temperature change from 70°C to 85°C does not have a significant effect on the flux, and that the GO membranes continue to operate with a high flux.

Figure 4 shows the lignin rejection properties of the GO membranes. Under all operating conditions investigated, including the most realistic conditions of 85°C and 10 wt% TS, the lignin rejection stabilizes at TMPs of 25-30 bar and reaches high values in the 95-98% range. Figure S2 (Supporting Information) shows photographs of the 10 wt% TS BL feed and the clarified permeate liquid obtained at a TMP of 35 bar and operating temperature of 70°C. Also, we estimated the inorganic reduction properties of these GO membranes. Figure S2 (Supporting Information) shows the conductivity reduction of the GO membranes using both stock BL feeds with 0.3wt% and 10wt% TS as well as a model (simulated) BL feed containing only inorganic solids (3.2 wt%) as shown in Table S2. The conductivity reduction properties are strongly dependent on the feed solids concentration and TMP. The conductivity reduction reaches 75% for the diluted 0.3 wt% BL feed whereas it is much lower (~20%) at high solids concentrations for both the real and simulated BL feeds. The latter is behavior is expected, since the present GO membranes perform as NF membranes with pore size appropriate for rejecting lignin (0.5-10 kDa) and not the small inorganic solutes that are expected to contribute the most towards the


Page 15 of 27

conductivity.

Since it is unlikely that a single membrane can efficiently retain lignin,

hemicelluloses, and small inorganic solutes, a second membrane stage (e.g., an RO membrane) would be desirable if these components are to be individually separated and recycled. Additionally, the diluted BL is at a lower pH (8-9), wherein a larger fraction of inorganic anions remain in the retentate, due to the association of their cationic counterparts with the anionic lignin. Also, the pure water flux of pristine and BL-immersed GO membrane is shown in Figure S3. It is seen that the permeation properties of the GO membranes are unaffected by BL exposure at higher temperatures.

(c)

(b)

BL-Feed (10wt%)

BL-Feed (0.3wt%)

(e)

(d)

Permeate 99.5% Lignin

(a)

Permeate 98% Lignin

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4. (a) Lignin rejection versus TMP during BL-concentration with GO membranes under different operating conditions (0.3wt% and 10wt% TS at 70°C and 85°C); photographs of the GO membrane (b) before and (c) after BL exposure over an operational time of 96 hr; and photographs of the BL-feed and permeate samples collected at (d) high (10 wt%) and (e) low (0.3 wt%) concentration feed tests.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 27

The permeation data in Figures 3-4 were collected by using fresh BL feed for each data point at a different TMP. This allows a reliable determination of the fundamental membrane characteristics. However, in preliminary experiments with the dead-end membrane cell (data not shown here), we filtered more than 2/3rd of the initial volume (~300 mL) of the 10 wt% TS BL feed over several days, using progressively increasing TMPs up to 50 bar. At the end of the experiment, the BL on the feed side was concentrated to more than 25 wt% TS due to water removal. The membrane continued to permeate liquid with no sign of degradation or severe concentration polarization. Although this measurement is not useful to determine membrane characteristics (since the feed concentration changed continuously), it indicates that the membrane performance is likely to be maintained for industrial BL feeds which have somewhat higher solids concentration (15-18 wt% TS) than the 10 wt% TS BL used in this work.

Figure 5 shows the determination of the molecular weight cut-off (MWCO) for the GO membranes. The rejection of five dye molecules (Experimental Methods) with different MWs in the range of 300-800 Da was determined at room temperature and under the same TMPs used for BL concentration. Figure 5a shows that the rejections stabilize at TMPs around 35 bar, similar to the lignin rejection behavior. Therefore, the rejections at 35 bar were used for MWCO determination. Based upon Figure 5b, the GO membranes reach 90% rejection at an MW of about 625 Da, which is then taken as the MWCO. The membranes also reach nearly 100% rejection at 800 Da, which is consistent with the high overall lignin rejections (95-98%) observed in Figure 4 and its known MW distribution (0.5-100 kDa).


Page 17 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(a)

(b)

Figure 5. Dye rejection by GO membrane as a function of (a) TMP for different dyes, and (b) dye MW at TMP of 35 bar. The MWCO is determined from the dashed lines shown. The solid curve is only a guide for the eye.

Figures 6a-6b show an AFM image of dispersed GO laminates obtained from an aqueous suspension and the height trace of the out-of-plane dimension. The GO laminates are 200-1000 nm in lateral dimensions and 5-6 nm in thickness. Figures 6c-6e show top surface and cross-section SEM views of the PES-supported GO membrane in pristine (as-made) form. A thin, uniform, and continuous GO membrane of thickness 250±31 nm is seen on the support. Figures 6f-6h show corresponding SEM images of the GO membrane after about 4 days of use in BL concentration experiments with 10 wt% TS BL at 70°C. The membrane was not cleaned 17 ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 27

after use. The visual appearance of the membrane surface is not significantly different from that of the pristine membrane, and maintains the thin, uniform, and continuous GO layer without any visible degradation. The PES support also does not show any significant visible degradation.

(b)

(a)

0.0

1.3 µm

(c)

(f)

(d)

(g)

(e)

(h)

Figure 6. (a) AFM image showing the lateral dimensions of the GO laminates, and (b) the height trace showing the out-of-plane dimensions; Top surface and cross-section SEM views of GO membrane supported on PES before (c-e) and after (f-h) BL concentration measurements.


Page 19 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(a)

(b

(c)

(d

(e)

Figure 7. XRD patterns of PES-supported GO membranes: (a) dry and wet as-made membranes before BL exposure, (b) exposed to 0.3 wt% TS BL at 20°C for 3 days, (c) exposed to 0.3 wt% TS BL at 70°C for 3 days, (d) exposed to 10 wt% TS BL at 20°C for 3 days, and (e) exposed to 10wt% TS at both 70°C (4 days) and 85°C (3 days). For all the BL-exposed membranes, XRD patterns were collected on dried membranes before and after cleaning the membrane surfaces.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 27

Figure 7 shows the XRD patterns of the PES-supported GO membranes before and after black liquor (BL) exposure at different temperatures (20, 70, and 85°C) and different BL concentrations (TS = 0.3 wt% and 10 wt%), and also before and after cleaning the BL-exposed membranes. All the patterns show the broad peak characteristic of the PES support centered at a 2θ position of 18.5°, and corresponding to a d-spacing of 0.49 nm (packing repeat distance of the glassy PES chains). The GO membranes show broad but well-defined peaks corresponding to the d-spacings between the oxidized sheets with sp3 hybridization and also between pristine/partially oxidized sheets with mainly sp2 hybridization.15 The broader XRD peaks of GO in comparison with the sharp peaks in graphite is because the layer oxidation (to form hydroxyl, carboxylic acid, and epoxide functional groups) is accompanied by stretching and disordering of the distance between the layers.37 The XRD patterns of all the dry as-made GO membranes in Figures 7a-7e present interlayer d-spacings in the regions of ~0.6 nm and ~0.8 nm. The former corresponds to pristine or partially oxidized regions, whereas the latter corresponds to fully sp3 hybridized domains with a monolayer of water present between the sheets16,38 even in the ‘dried state’ under ambient air. Figure 7a shows that after soaking the GO membrane in water, the interlayer d-spacing of the oxidized regions swells to ~1.3 nm corresponding to three water layers, in good agreement with the behavior of previously reported GO laminates15 prepared by vacuum filtration. In Figures 7b-7e, the XRD patterns of all the dried GO membranes after exposure to BL (regardless of the temperature and BL solids concentration) shows that the previous d-spacing of ~0.8 nm becomes very broad and is centered at a slightly larger d-spacing of ~0.9 nm. This indicates that exposure to BL has the effect of slightly swelling the interlayer spacing and reducing the registry between the oxidized layers. The swelling of the layers may be caused by the intercalation of low-MW lignin or other organic species between the layers. The


Page 21 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


functional d-spacing of ~0.9 nm in the GO membranes exposed to a BL environment is fully consistent with its high performance as a lignin-rejecting NF membrane and its MWCO of ~625 Da. Also, BL-exposed membranes after cleaning generally exhibit more well-defined XRD peaks than before cleaning, presumably due to the removal of surface foulants that reduce the XRD intensity from the GO membranes. However, the smaller interlayer d-spacing of ~0.6 nm is not affected by BL exposure as expected for pristine and hydrophobic graphene sheets.39

Conclusion We have demonstrated that GO membranes supported on low-cost macroporous PES supports can be used for high-performance (up to 98% rejection, and fluxes in the range of 20 kg m-2 h-1 at surface fluid velocities of 0.2-0.3 m/s) lignin rejection to concentrate 10wt% to 20 wt% BL under realistic process conditions for the purpose of Kraft black liquor concentration. While GO membranes have been previously demonstrated to work well for conventional filtration of organic molecules at low concentrations, to our knowledge this is the first work that studies in the detail their performance under extreme operating conditions (highly caustic pH and relatively high temperature) and with more concentrated feeds. The GO membranes are fabricated by a simple and low-cost vacuum filtration method and are likely amenable to scale-up. Characterization by combined XRD, SEM, AFM, and dye rejection measurements reveals the presence of thin (250 nm), uniform GO membranes on the PES supports. The interlayer dspacing (~0.9 nm) of the functional GO membranes is consistent with the MWCO of ~625 Da. The present GO membranes also are robust upon prolonged exposure to BL.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 27

Supporting Information Tables of BL composition, schematic of membrane permeation apparatus, conductivity reduction and water flux data for GO membranes. This material is available free of charge at http://pubs.acs.org.

Author Information Corresponding Author * S. Nair: [email protected] Notes The authors declare no competing financial interest.

Acknowledgements We acknowledge the following individuals (all Georgia Tech): M. Orr for assistance with DMA, J. Buchanan for BL composition characterization, D.-Y. Koh for assistance with Hg porosimetery, Prof. V. Breedveld for access to BL viscosity and density measurements, and Prof. E. Reichmanis for access to UV-Vis spectrophotometry. This work was supported by a consortium of companies (International Paper, Kapstone, Georgia-Pacific, SAPPI, and WestRock) and the Renewable Bioproducts Institute.


Page 23 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


References (1)

Fallavollita, J. A.; Mujumdar, A. S.; Avedesian, M. M. Kraft black liquor recovery in a

fluidized bed: Part I - A review. Canadian J. Chem. Eng. 1987, 65, 812-817. (2)

Reeve, D. W. The Kraft recovery cycle. TAPPI Press: Atlanta, 2002.

(3)

Adams, T. N.; Frederick, W. J. Kraft recovery boiler physical and chemical processes.

American Paper Inst. Inc., New York, 1988. (4)

Cardoso, M.; de Oliveira, E. D.; Passos, M. L. Chemical composition and physical

properties of black liquors and their effects on liquor recovery operation in Brazilian pulp mills. Fuel 2009, 88, 756-763. (5)

Forest Products Industry Technology Roadmap: Agenda 2020 www.agenda2020.org.

(6)

Arkell, A.; Olsson, J.; Wallberg, O. Process performance in lignin separation from

softwood black liquor by membrane filtration. Chem. Eng. Res. Des. 2014, 92, 1792-1800. (7)

Kong, L.B.; Hasanbeigi, A.; Price, L. Assessment of emerging energy-efficiency

technologies for the pulp and paper industry: a technical review. J. Cleaner Prod. 2016, 122, 528. (8)

De, S.; Bhattacharya, P.K. Recovery of water and inorganic chemicals from kraft black

liquor using membrane separation processes. TAPPI J. 1996, 79, 103-111. (9)

Hegab, H. M.; Zou, L. Graphene oxide-assisted membranes: Fabrication and potential

applications in desalination and water purification. J. Membr. Sci. 2015, 484, 95-106. (10)

Singh, V.; Joung, D.; Zhai, L.; Das, S.; Khondaker, S. I.; Seal, S. Graphene based

materials: Past, present and future. Prog. Mater. Sci. 2011, 56, 1178-1271. (11)

Cohen-Tanugi, D.; Grossman, J. C. Nanoporous graphene as a reverse osmosis

membrane: Recent insights from theory and simulation. Desalination 2015, 366, 59-70.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(12)

Page 24 of 27

Dikin, D. A.; Stankovich, S.; Zimney, E. J.; Piner, R. D.; Dommett, G. H. B.;

Evmenenko, G.; Nguyen, S. T.; Ruoff, R. S. Preparation and characterization of graphene oxide paper. Nature 2007, 448, 457-460. (13)

Zhu, Y.; Murali, S.; Cai, W.; Li, X.; Suk, J. W.; Potts, J. R.; Ruoff, R. S. Graphene and

graphene oxide: synthesis, properties, and applications. Adv. Mater. 2010, 22, 3906-3924. (14)

Bunch, J. S.; Verbridge, S. S.; Alden, J. S.; van der Zande, A. M.; Parpia, J. M.;

Craighead, H. G.; McEuen, P. L. Impermeable atomic membranes from graphene sheets. Nano Lett. 2008, 8, 2458-2462. (15)

Joshi, R. K.; Carbone, P.; Wang, F. C.; Kravets, V. G.; Su, Y.; Grigorieva, I. V.; Wu, H.

A.; Geim, A. K.; Nair, R. R. Precise and ultrafast molecular sieving through graphene oxide membranes. Science 2014, 343, 752-754. (16)

Akbari, A.; Sheath, P.; Martin, S. T.; Shinde, D. B.; Shaibani, M.; Banerjee, P. C.; Tkacz,

R.; Bhattacharyya, D.; Majumder, M. Large-area graphene-based nanofiltration membranes by shear alignment of discotic nematic liquid crystals of graphene oxide. Nat. Commun. 2016, 7, art. # 10891. (17)

Huang, K.; Liu, G.; Lou, Y.; Dong, Z.; Shen, J.; Jin, W. A graphene oxide membrane

with highly selective molecular separation of aqueous organic solution. Angew. Chem. Intl. Ed. 2014, 53, 6929-6932. (18)

Kovtyukhova, N. I.; Ollivier, P. J.; Martin, B. R.; Mallouk, T. E.; Chizhik, S. A.;

Buzaneva, E. V.; Gorchinskiy, A. D. Layer-by-layer assembly of ultrathin composite films from micron-sized graphite oxide sheets and polycations. Chem. Mater. 1999, 11, 771-778. (19)

Hummers, W. S.; Offeman, R. E. Preparation of graphitic oxide. J. Am. Chem. Soc. 1958,

80, 1339-1339.


Page 25 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


(20)

Frederick, J. W.: Kraft recovery boilers; black liquor properties. TAPPI Press: Atlanta,

1997. (21)

Venkatesh, V., and Nguyen, X. N. Evaporation and concentration of black liquor, in

Chemical recovery in the alkaline pulping process. TAPPI Press: Atlanta, 1992. (22)

De, S.; Bhattacharya, P. K. Recovery of water and inorganic chemicals from kraft black

liquor using membrane separation processes. TAPPI J. 1996, 79, 103-111. (23)

Zabkova, M.; da Silva, E. A. B.; Rodrigues, A. E. Recovery of vanillin from

lignin/vanillin mixture by using tubular ceramic ultrafiltration membranes. J. Membr. Sci. 2007, 301, 221-237. (24)

Rajnish; Bhattacharya, P. K. Ultrafiltrative solute rejection behavior of black liquor. ACS

Symp. Ser. 1985, 281, 313-323. (25)

Poddar, T. K.; Singh, R. P.; Bhattacharya, P. K. Ultrafiltration flux and rejection

characteristics of black liquor and polyethylene-glycol. Chem. Eng. Commun. 1989, 75, 39-56. (26)

Cheryan, M.: Ultrafiltration and microfiltration handbook; CRC Press: Lancaster, PA,

1998. (27)

Li, J.; Ohagan, T.; MacLeod, J. M. Using ultrafiltration to analyze the molar mass

distribution of kraft lignin at pH 13. Canadian J. Chem. Eng. 1996, 74, 110-117. (28)

Sjöström, E.; Alén, R. (Eds.) Analytical methods in wood chemistry, pulping and

papermaking; Springer, 1999. (29)

Niemelä, K. Characterization of pulping liquors, in: Analytical methods in wood

chemistry, pulping and papermaking; Springer, 1999. (30)

J. Gullichsen (Ed.) Chemical pulping: papermaking science and technology series - 6 B;

Fapet Oy, 2000.



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(31)

Page 26 of 27

Foley, G.: Membrane filtration: a problem solving approach with MATLAB®;

Cambridge University Press, 2013. (32)

Causserand, C.; Rouaix, S.; Akbari, A.; Aimar, P. Improvement of a method for the

characterization of ultrafiltration membranes by measurements of tracers retention. J. Membr. Sci. 2004, 238, 177-190. (33)

Solids content of black liquor, TAPPI/ANSI Test Method T 650 om-15.

(34)

Analysis of soda and sulfate black liquor, TAPPI/ANSI Test Method T 625 cm-14.

(35)

Dong, D. J.; Fricke, A. L. UV-visible response of kraft lignin in softwood black liquor.

Mater. Res. Soc. Proc. 1990, 197, 77-86. (36)

van der Marel, P.; Zwijnenburg, A.; Kemperman, A.; Wessling, M.; Temmink, H.; van

der Meer, W. Influence of membrane properties on fouling in submerged membrane bioreactors. J. Membr. Sci. 2010, 348, 66-74. (37)

Li, D.; Muller, M. B.; Gilje, S.; Kaner, R. B.; Wallace, G. G. Processable aqueous

dispersions of graphene nanosheets. Nat. Nano. 2008, 3, 101-105. (38)

Tsou, C.-H.; An, Q.-F.; Lo, S.-C.; De Guzman, M.; Hung, W.-S.; Hu, C.-C.; Lee, K.-R.;

Lai, J.-Y. Effect of microstructure of graphene oxide fabricated through different self-assembly techniques on 1-butanol dehydration. J. Membr. Sci. 2015, 477, 93-100. (39)

Nair, R. R.; Wu, H. A.; Jayaram, P. N.; Grigorieva, I. V.; Geim, A. K. Unimpeded

permeation of water through helium-leak-tight graphene-based membranes. Science 2012, 335, 442-444.


Page 27 of 27

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


For Table of Contents Use Only Graphene Oxide Membranes in Extreme Operating Environments: Concentration of Kraft Black Liquor by Lignin Retention Fereshteh Rashidi, Nikita S. Kevlich, Scott A. Sinquefield, Meisha L. Shofner and Sankar Nair

Graphene oxide membranes supported on poly(ethersulfone) can separate lignin from Kraft black liquor at high pH and temperature, using much less energy than a conventional evaporative process.


Graphene Oxide Membranes in Extreme Operating Environments

Recommend Documents