S1 Visible-Light-Responsive Graphitic Carbon Nitride

1 downloads 0 Views 4MB Size Report
Ammonium oxalate (Sigma-Aldrich, 99+%), tert-butyl alcohol (Sigma-Aldrich, ... potassium phosphate monobasic (Fisher Scientific, 99.3%), acetonitrile (ACN) ...
Visible-Light-Responsive Graphitic Carbon Nitride (g-C3N4): Rational Design and Photocatalytic Applications for Water Treatment

Supporting Information

Qinmin Zheng,1 David P. Durkin,2 Justin E. Elenewski,3 Yingxue Sun,1, 4 Nathan A. Banek,3 Likun Hua,5 Hanning Chen,3 Michael J. Wagner,3 Wen Zhang,5 Danmeng Shuai,1*

1

Department of Civil and Environmental Engineering, The George Washington University,

Washington, DC 20052, United States 2

Department of Chemistry, Johns Hopkins University, Baltimore, MD 21218, United States

3

Department of Chemistry, The George Washington University, Washington, DC 20052, United

States 4

Department of Environmental Science and Engineering, Beijing Technology and Business

University, Beijing 100048, China 5

Department of Civil and Environmental Engineering, New Jersey Institute of Technology,

Newark, NJ 07102, United States * Corresponding Author: Phone: 202-994-0506, Fax: 202-994-0127, Email: [email protected], Website: http://materwatersus.weebly.com/

Environmental Science and Technology 3 Tables, 16 Figures, and 38 Pages

S1

I. Materials and Methods Reagents All chemicals were at least reagent grade and used as received. The synthesis of g-C3N4 involved urea (Sigma-Aldrich, 98%), melamine (Acros Organic, 99+%), cyanuric acid (SigmaAldrich 98%), barbituric acid (Sigma-Aldrich, 99%), etidronic acid monohydrate (Sigma-Aldrich ,95+%), and ethanol (Sigma-Aldrich, 99.5%). A buffer solution of 1 mM potassium phosphate monobasic (Fisher Scientific, 99.3%) was adjusted to pH 7.3 and used in photocatalytic experiments. Phenol (Sigma-Aldrich, 99+%), atrazine (Sigma-Aldrich, 98.8%), carbamazepine (Sigma-Aldrich, 98+%), and sulfamethoxazole (Sigma-Aldrich, 99+%) were used as probe contaminants in photocatalytic studies. Potassium nitrate (Fisher Scientific, 99.9%), humic acid sodium salt (Sigma-Aldrich), magnesium chloride (Fisher Scientific, 99.4%), calcium chloride dihydrate (Fisher Scientific, 74.3%), and sodium hydrosulfide hydrate (Sigma-Aldrich, NaHS 60+%) were selected as representative natural water constituents or foulants for photocatalytic reactions. Ammonium oxalate (Sigma-Aldrich, 99+%), tert-butyl alcohol (Sigma-Aldrich, 99+%), superoxide dismutase (from bovine erythrocytes, Sigma-Aldrich, > 3,000 units mg-1 of protein), catalase (from bovine liver, lyophilized powder, Sigma-Aldrich, 2,000-5,000 units mg-1 of protein), and L-histidine (Sigma-Aldrich, 99+%) were selected as scavengers for oxidative species in photocatalytic reactions. The eluents for high performance liquid chromatographic (HPLC) analyses of pollutants consisted of sodium acetate (Sigma-Aldrich, anhydrous), potassium phosphate monobasic (Fisher Scientific, 99.3%), acetonitrile (ACN) (Fisher Scientific, 99.9%), and methanol (Sigma-Aldrich, 99.9%). All solutions were prepared in ultrapure water (Millipore, Milli-Q, 18.2 M cm).

S2

Density Functional Theory (DFT) Simulations Electronic structure calculations were conducted using density functional theory (DFT)1, 2 as implemented in the CP2K suite.3 An initial cell relaxation was performed with the PerdewBurke-Ernzerhof (PBE) density functional,4 a moderately-sized double-zeta valence polarized (DZVP) basis set,5 and Goedecker-Teter-Hutter (GTH) pseudopotentials (grid cutoff of 300 Ry)6, 7

in a hybrid Gaussian-plane wave (GPW) framework.8, 9 This optimized cell geometry was then

used to define subsequent relaxations for undoped and doped materials via the range-separated, hybrid Heyd-Scuseria-Ernzerhof 06 (HSE06) functional.10, 11 Energy shifts due to solvation were quantified by performing a reference calculation on undoped g-C3N4 using an implicit model.12 All structures with nitrogen site substitutions were assumed to be uncharged, while interstitially doped systems and those with dopants at carbon sites were assigned a charge of +1. The relative stability of each dopant family was determined by comparing geometries optimized using undoped g-C3N4 cell parameters, excluding the nitrogen to carbon substitutions which used the average of optimized cell parameters for that series of dopants.

g-C3N4 Synthesis The conventional g-C3N4 samples (i.e., U and M) were synthesized from urea or melamine. 10 g of urea or melamine powder was put into an alumina crucible with a cover (not sealed), heated at a rate of 2.3 °C min-1 and maintained at 550 °C for 4 h in a muffle furnace, and then cooled down naturally. The supramolecule-based g-C3N4 samples without non-metal doping, MC, were synthesized from melamine and cyanuric acid (mass ratio 1:1). Carbon-doped, supramolecule-based g-C3N4 samples, MCBx, were synthesized from melamine, cyanuric acid, and barbituric acid with different mass ratios (i.e., 2 g of melamine, (2-x) g of cyanuric acid, and

S3

x g of barbituric acid; x = 0-1.9). Phosphorus-doped, supramolecule-based g-C3N4 samples, MCEy, were synthesized from melamine, cyanuric acid, and etidronic acid with different mass ratios (i.e., 2 g of melamine, (2-y) g of cyanuric acid, and y g of etidronic acid; y = 0-0.5). Precursors were first dispersed in 40 mL of ethanol to form suspension. The suspension was next stirred at ambient temperature for 3 h, followed by sonication (Elamsonic P, 37 kHz, 100 W) at room temperature for an additional 3 h. The suspension was then dried on a hot plate at 70 °C until no obvious liquid was found, and white supramolecular aggregates were formed. Finally, the dried supramolecular aggregates were put into an alumina crucible with a cover (not sealed), heated at a rate of 2.3 °C min-1 and maintained at 550 °C for 4 h in a muffle furnace, and cooled down naturally.

g-C3N4 Characterization The crystal phase of g-C3N4 was determined by X-ray powder diffraction (XRD) analyses on a Rigaku Miniflex+ diffractometer with Cu Kα radiation. Sample morphologies were characterized with a scanning electron microscope (SEM, JEOL 6700F) and a transmission electron microscope (TEM, Philips CM300 FEG). For SEM, g-C3N4 samples dispersed in ethanol were cast and dried on a sample stub, and SEM micrographs were collected at 10 kV with a secondary electron detector. For TEM, g-C3N4 samples dispersed in ethanol were also cast and dried on a Cu grid with a carbon support, and TEM images were collected at 300 kV. The elemental distribution of MCE was characterized by SEM-energy dispersive spectroscopy (EDS) (SEM, TESCAN MIRA3 FEI; EDS, TEAM Octane SSD) at an accelerating voltage of 20 kV.

S4

The surface properties were investigated by X-Ray photoelectron spectroscopy (XPS). A PHI 5400 system was used for analysis under UHV conditions (pressure < 10-8 Torr). A Mg K source (1253.6 eV) was utilized, and ejected photoelectrons were measured with a hemispherical energy analyzer operating at 58.7 eV constant pass-energy. Peak positions were referenced to C1s, 284.5 eV, and CasaXPS was used to determine chemical composition and atomic concentrations at the surface (up to ca. 10 nm). Attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectra were collected using a Nicolet 6700 spectrometer from 4000-525 cm-1 (32 scans at 4 cm-1 resolution). Spectra were peak normalized by the C-N asymmetric stretch at 1231 cm-1, a region present in every sample tested with relative invariant spectral intensity. The Braunauer-Emmitt-Teller (BET) surface area and porosity analyses were performed by N2 adsorption/desorption using a Micromeritics TriStar 3000. Isotherm adsorption data for P0/P was recorded from 0.06-0.989. The volume of micropores was determined by t-plot analysis and the volume of mesopores was determined by BJH analysis. Samples were degassed at 140 °C for 12 h under dynamic vacuum (10-3 Torr) prior to analysis. Bulk carbon, hydrogen, and nitrogen were analyzed on Model CE 440 CHN Analyzer. The capsule containing g-C3N4 samples was injected into a high temperature (1000 °C) furnace and combusted in pure O2 under static conditions. To ensure the complete combustion, a dynamic burst of O2 was added at the end of the combustion period. The resulting combustion product contained CO2, H2O, and N2/NOx, and it next passed over Cu to scrub excess O2 and reduce NOx to N2. After scrubbing, the gases entered a mixing volume chamber to ensure a homogenous mixture at constant temperature and pressure, and were detected by high-precision thermal conductivity detectors. A H2O trap and a CO2 trap were used between the detectors, and the

S5

differential signal before and after the trap was proportional to the H2O concentration and CO2 concentration. Finally, N2 was measured against a He reference. The optical absorbance spectra of the photocatalysts and their band gap were determined by a Thermo Scientific Evolution 300 UV-vis spectrophotometer with a Praying Mantis diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) accessory. The absorbance of gC3N4 samples irradiated by light over a wavelength scan from 200 to 800 nm was measured, and the value was then converted into reflectance by equation S1:

1 A  log( ) R

(S1)

where A is the measured absorbance and R is the reflectance. The function , representative of optical absorption, which is equivalent to adsorption coefficient divided by scattering coefficient for g-C3N4, was calculated from the Kubelka-Munk formula:13-15



(1  R)2 2R

(S2)

The band gap was obtained by extrapolating the linear portion of (h)1/2 versus photon-energy plots at (h)1/2 = 0, h is Planck’s constant (4.14×10-15 eV s),  is the frequency of photons which can be obtained by dividing c, the speed of light (3.0×108 m s-1) by λ, the photon’s wavelength. The photoluminescence (PL) spectra were obtained from a home-made apparatus based on a Thermo Nicolet Nexus 670 rapid scan FTIR spectrometer. The PL spectra of the samples were obtained using a 354 nm diode pumped solid state laser from Teem Photonics as excitation source. The intensity on the samples was 30 µW focused to a spot of 100 µm in diameter. The emitted fluorescence was focused to a 300 mm path length monochromator and collected by a thermoelectrically-cooled charge-coupled device (CCD) camera, both from Princeton S6

Instruments. A long wavelength pass filter cutting at 420 nm was used to block the laser light into the monochromator. Zeta potential and hydrodynamic diameter of g-C3N4 aqueous suspension were determined using folded capillary cells (DTS 1061, Malvern) on a Zetasizer Nano ZS instrument (Malvern, ZEN3600). The temperature was maintained at 25 °C, and the scattering angle was 173° from the incident laser beam. The UV-vis spectra of g-C3N4 aqueous suspension was obtained on a Thermo scientific Evolution 201PC spectrophotometer. The absorption spectra of the catalyst suspension were measured in the region of 200 to 800 nm with at a resolution of 2 nm. The sample cell was a quartz cuvette (1 cm by 1 cm). The concentration of the prepared catalysts was 1 g L-1 in the aqueous solution, and the suspension was dispersed using a bath sonicator (100 W) for 10 min.

Light Source Description and Characterization Xenon arc lamp (1000 W) was selected as a light source in our study, similar to the one used in other research.16 A long-pass optical filter with a cut-off wavelength of 400 nm was used to simulate visible light irradiation at earth’s surface. Spectral irradiance, photon fluence, and optical powder density of the Xenon lamp were recorded by a spectroradiometer (AvaSpec ULS2048L). The spectral irradiance was plotted in Figure S1, and the photon fluence and the optical powder density were 601 (mol of photons) m-2 s-1 and 16.7 mW cm-2, respectively.

S7

Figure S1. The spectral irradiance of the Xenon lamp ( > 400 nm). g-C3N4 with a band gap of 2.72 eV only absorbs photons with a wavelength up to ca. 460 nm.

Photocatalytic Activity Tests in a Phosphate Buffer First, 15 mg of g-C3N4 was mixed with ultrapure water and dispersed via sonication (Elamsonic P, 37 kHz, 100 W) for 15 minutes. The dispersed g-C3N4 suspension and a probe pollutant (i.e., phenol, atrazine, carbamazepine, or sulfamethoxazole) were added to a phosphate buffer solution in a jacketed reactor, giving a total reaction volume of 15 mL, the g-C3N4 loading of 1 g L-1, the phosphate buffer concentration of 1 mM (pH 7.3), and the pollutant concentration of 100 µM. The reactor had a diameter of 4 cm and starting depth of suspension of 1.5 cm. The S8

surface of the suspension was centered 20 cm below the Xenon lamp, which had a beam width of 6 cm. The reaction temperature was maintained at 25 oC. The suspension was stirred for 20 minutes in the dark in order to ensure suspension homogeneity, and a sample of 300 μL was next taken to quantify the initial pollutant concentration. Pollutant adsorption onto g-C3N4 was negligible, based on the comparison of the measured initial concentration with the concentration in control reactors in absence of g-C3N4. The reactor was then irradiated under the Xenon lamp (> 400 nm, typically for 60 minutes), during which aqueous aliquots were withdrawn periodically. The aliquots were next centrifuged at 13,000 rpm (16,060 g) for 1 hour in the dark to allow the particles to settle down. The supernatant was transferred to a 1.5 mL amber autosampler vial for subsequent analysis via HPLC with photodiode array detection (Shimadzu LC-20AT Prominence HPLC-DAD). For atrazine degradation over multiple photocatalytic cycles, g-C3N4 was harvested by centrifugation after each cycle, rinsed with ultrapure water, and reused for the next run. The photocatalyst used after four cycles was characterized by SEM-EDS and ATR-FTIR.

Photocatalytic Activity Tests in Simulated Complex Water Matrices and Real Water Samples To explore the influence of water chemistries on photocatalytic performance, particularly water matrices representative of water treatment systems, the reactivity of MCB0.07 for atrazine degradation was explored in simulated water samples (1 mM phosphate buffer solution with the presence of 5 mM NaHCO3, 50 mM CaCl2, 50 mM MgCl2, 5 mg L-1 humic acid sodium salt, 5 mg L-1 NaNO3 as N, or 0.1 mM NaHS; pH 7.3 for all solutions except for NaHCO3 solution with pH 7.5). Ca2+ and Mg2+ were selected to represent the hardness species in water, NO3-, humic

S9

acid, and HS- were selected as representative scavengers for oxidative species generated from photocatalytic reactions, and HCO3- was selected to represent alkalinity species that could also react with the oxidative species. Atrazine degradation was also conducted in real water samples collected form the Griffith Water Treatment Plant (GWP) and the Broad Run Wastewater Reclamation Facility (BRWRF) in Virginia. Samples were collected from the raw water of GWP (GWP-1, i.e., water from Occoquan reservoir, VA), the final effluent from GWP (GWP-4), the effluent after membrane bioreactor treatment in BRWRF (BRWRF-1), and the final effluent from BRWRF (BRWRF-3), and filtered through a 0.45 m polyvinylidene fluoride (PVDF) membrane before photocatalytic reactions. For the exploration of long-term effect of complex water matrices for photocatalytic activity, MCB0.07 was also added to the water samples and magnetically stirred in the dark for 24 hours to pre-foul the photocatalyst before the photocatalytic reactions. Atrazine was added to the reactor 20 minutes before irradiation to ensure suspension homogeneity, and again, there was no evidence of adsorption in any system. The flow chart of water and wastewater treatment process in GWP and BRWRF is shown in Figure S2. pH and alkalinity of these real water samples are listed in Table S1.

S10

Figure S2. The flow chart of (a) water treatment process in the Griffith Water Treatment Plant (GWP) and (b) wastewater treatment process in the Broad Run Wastewater Reclamation Facility (BRWRF).

Table S1. Characteristics of Water Samples GWP-1

GWP-4

BRWRF-1

BRWRF-3

pH

6.7

7.2

7.5

7.0

Alkalinity (mg L-1 as CaCO3)

56

61

87

N/A

S11

HPLC Analysis The Shimadzu LC-20AT Prominence HPLC System was equipped with a Shimadzu C18 column (4.6 × 50 mm, 1.8 μm particle size). The HPLC analysis method for phenol was based on the previous work and employed a mobile phase of 65:35 1 mM sodium acetate: ACN at pH 3, a flow rate of 0.75 mL min-1, an injection volume of 20 μL, and a 254 nm detection wavelength.17 The HPLC analysis method for atrazine was based on the previous work and employed a mobile phase of 50:50 ultrapure water: ACN, a flow rate of 1 mL min-1, an injection volume of 100 μL, and a 223 nm detection wavelength.18 The HPLC analysis method for carbamazepine was based on the previous work and had a mobile phase of 55:45 ultrapure water: ACN, a flow rate of 1 mL min-1, an injection volume of 25 μL, and a 213 nm detection wavelength.19 The HPLC analysis method for sulfamethoxazole was based on the previous study and had a mobile phase of 70:30 5 mM potassium phosphate monobasic (pH adjusted to 5.0): methanol, a flow rate of 1 mL min-1, an injection volume of 20 μL, and a 268 nm detection wavelength.16

S12

II. Results and Discussion Table S2. Band Gap, and the Energy Level of Conduction Band Minimum (CBM) and Valence Band Maximum (VBM) Calculated Based on Density Functional Theory (DFT) Simulations Band Gap (eV)

CBMa (V)

VBMa (V)

Undoped g-C3N4

2.75

-1.03

1.71

N1 → C doped g-C3N4

2.58

-1.01

1.57

N2 → C doped g-C3N4

2.08

0.05

2.13

N3 → C doped g-C3N4

2.94

-0.67

2.28

N1 → P doped g-C3N4

2.06

-0.71

1.36

N2 → P doped g-C3N4

1.69

-0.20

1.49

N3 → P doped g-C3N4

2.90

-0.70

2.20

C1 → P doped g-C3N4

1.93

0.12

2.05

C2 → P doped g-C3N4

1.86

0.19

2.05

Pore → P doped g-C3N4

1.14

-0.06

1.08

a

vs. standard hydrogen electrode (SHE)

S13

Influence of Mass Transfer Processes on Observed Reaction Rates The following calculations were performed to investigate the potential effects of aqueous/solid and intraparticle mass transfer processes on the observed reaction rates. The calculations followed a similar approach reported for N-nitrosodimethylamine (NDMA) reduction on a Pd/Al2O3 catalyst or a Ni catalyst.20 The results show that mass transfer was not expected to limit the observed reaction rates in our system.

Aqueous/solid mass transfer limitations Here we estimate the lower limit for the aqueous/solid mass transfer rate constant for atrazine degradation on MCB0.07 (kaq/s, a) and compare it with the measured reaction rate constant. Atrazine degradation on MCB0.07 was selected because it showed the fastest degradation kinetics compared to the other photocatalytic reactions in our study. Criteria for the evaluation of potential influence of aqueous/solid mass transfer limitations on the observed rate constants are following: 1. No aqueous/solid mass transfer limitations expected if calculated kaq/sa is much greater than the largest measured kobs, MCB value. 2. Significant potential for aqueous/solid mass transfer limitations is expected if kaq/sa is close to or less than the measured kobs, MCB value. The slip velocity method21 is used to calculate the mass transfer coefficient (kaq/s) for particles traveling at the slip velocity (ut) relative to the suspending liquid. Because the particle size of MCB0.07 was much smaller than 1 mm, Stokes’ law is assumed to be applicable and the particle’s slip velocity is calculated by:

ut 

gd p2 (  p   ) 18

(S3) S14

where g is the gravity constant, dp and 𝜌p are the hydrodynamic diameter and density of the photocatalyst particle, respectively, 𝜌 is the density of water, and μ is the absolute viscosity of water. The aqueous/solid mass transfer coefficient is then estimated by the following expression:

kaq / s 

Dmol D Sh  mol (2  0.6 Re0.5 Sc0.33 ) dp dp

(S4)

where Dmol is the molecular diffusion coefficient of the reacting solute (atrazine), Sh is the Sherwood number, Re is the modified Reynold’s number, and Sc is the Schmitt number. The last two parameters are calculated by the following expressions: Re 

d p ut

H O 2

Sc 

(S5)

H O 2

Dmol

(S6)

where 𝜈H2O is the kinematic viscosity of water. The molecular diffusion coefficient for atrazine in water is calculated using the method in Hayduk and Laudie:21

Dmol 

13.26 105 1.14 ( ' ) 0.589

(S7)

where Dmol is in the unit of cm2 s-1, μ is in the unit of g m-1 s-1, and ν’ is the molar volume of atrazine, which is 250.6 cm3 mol-1, calculated using the LeBas method.22 Plugging the results into equation S5, using 𝜇 = 1.002 g m-1 s-1 (at 20 °C), then

Dmol

13.26 105   5.111010 m2 s 1 1.14 0.589 1.002 (250.6)

The values and other physical constants summarized in the Table S3 are then used in equations S2-6 to determine kaq/s. S15

Table S3. Constants Used to Calculate the Mass Transfer Coefficient Constant

Value

Dmol of atrazine

5.11×10-10 m2 s-1

Gravity constant, g

9.81 m s-2

Photocatalyst particle diameter, dp

2.39×10-6 m, determined by dynamic light scattering

Kinematic viscosity of water, νH2O

1.003 × 10-6 m2 s-1 (at 20 °C)

Particle density,23 ρp

1.34 × 106 g m-3

Water density, ρ

9.98 × 105 g m-3 (at 20 °C)

Absolute viscosity of water, μ

1.002 g m-1 s-1 (at 20 °C)

m g g )(2.39 106 m) 2 (1.34 106 3  9.98 105 3 ) 2 s m m  1.07  106 m  s 1 ut   g 18 18(1.002 ) ms 6 6 m d p ut (2.39 10 m)(1.07 10 s ) Re    2.54 106 2  H 2O 6 m 1.003 10 s 2 6 m  H 2O 1.003 10 s Sc    1962 2 Dmol 10 m 5.1110 s m2 5.111010 D D s (2  0.6(2.54 10 6 ) 0.5 (1962) 0.33 )  4.30 10 4 m  s 1 kaq / s  mol Sh  mol (2  0.6 Re0.5 Sc 0.33 )  dp dp 2.39 106 m gd p2 (  p   )

(9.81

To calculate the mass transfer rate constant, the mass transfer coefficient should be multiplied by the geometric surface area of the catalyst per volume of solution, a: a

total surface area SAp  M 1   total volume  p Vp VR

(S8)

where SAp is the geometric surface area of one MCB0.07 aggregate, which is assumed to have a spherical structure; M is the mass of photocatalyst in the reactor; Vp is the volume of one S16

MCB0.07 aggregate; and VR is the volume of reaction suspension. For the fastest reaction measured, M = 0.015 g and VR = 0.015 L, so SAp  M

1 4  (2.39 106 m / 2) 2  (0.015 g ) 1 a     1871.1m1 6 3 6 3 5 3  p Vp VR 1.34 10 g / m  4 / 3  (2.39 10 m / 2) 1.5 10 m

kaq / s a  4.30 104 m  s 1 1871.1m1  0.8s 1

(S9)

This value is about 3 orders-of-magnitude larger than the measured kobs,MCB value (8.0×10-4 s-1). Thus the results indicate the aqueous/solid mass transfer is expected to have negligible limitation on the measured reaction rates.

Intraparticle mass transfer limitations The potential of intraparticle mass transfer resistance is also investigate by using the same approach in previous studies, the criteria used are following: 1. No resistance to pore diffusion if

𝑘𝑜𝑏𝑠 𝐿2 𝐷𝑒

< 1.

2. Significant resistance to pore diffusion if

𝑘𝑜𝑏𝑠 𝐿2 𝐷𝑒

(S10a) > 1.

(S10b)

where L is characteristic diffusion path length for the photocatalyst and De is the effective diffusivity of the reacting solute. L and De are estimated using the following equations:24

1 L  dp 6 De 

(S11)

Dmol



(S12)

where 𝜃 is the porosity of the photocatalyst particle (typically ranging from 0.2-0.7) and τ is the tortuosity factor (typically ranging from 2-10). The most conservative values from each range are used to maximize the possibility that the criterion in equation S10a is met (i.e., we chose the S17

smallest θ (0.2) and largest τ (10) in order to obtain the smallest possible De value), and, in turn, the largest possible value for equations 10a and b:

1 1 L  d p   2.39 106 m  3.99 107 m 6 6 2 10 m 5.1110  0.2 D  s De  mol   1.02 1011 m2  s 1  10 The largest observed reaction rate constant was 8.0×10-4 s-1 for atrazine degradation on MCB0.07. This rate constant and the values of L and De calculated above are then used to test the criteria outlined in equations 10a and b:

kobs

L2 8.0 104 s 1  (3.99 107 m)2   1.25 105 m  1 11 2 1 De 1.02 10 m  s

Because the value is several orders-of-magnitude less than one, intrapaticle mass transfer resistance is negligible on the time scale over which the photocatalytic degradation was observed in our system.

S18

Figure S3. Phenol degradation on MCBx (x = 0.01-1.9) under simulated visible sunlight irradiation. Experimental conditions: photocatalyst loading of 1 g L-1; phenol initial concentration of 100 𝜇M; phosphate buffer of 1 mM, pH 7.3; and Xenon lamp irradiation,  > 400 nm). Error bars represent uncertainties of duplicates.

S19

(a)

S20

(b)

S21

(c)

Figure S4. Scanning electron microscopic images and elemental mapping analyses of (a) MCE0.01, (b) MCE0.07, and (c) MCE0.5.

S22

Figure S5. (a) Phenol and (b) atrazine degradation on MCEy (y = 0.01-0.5) under simulated visible sunlight irradiation. Experimental conditions: photocatalyst loading of 1 g L-1; phenol or atrazine initial concentration of 100 𝜇M; phosphate buffer of 1 mM, pH 7.3; and Xenon lamp irradiation,  > 400 nm). S23

Figure S6. XRD pattern of different g-C3N4 samples (U, M, MC, and MCB0.07). a.u. represents arbitrary units. The peak at 13° corresponding to in-plane ordering of tri-s-triazine units was less pronounced for supramolecule-based g-C3N4 (i.e, MC and MCB0.07) compared to M, indicating that more defects may be generated in the growth of g-C3N4 with the addition of cyanuric acid and/or barbituric acid.15,

25

The other peak around 27.6°, representative of (002) interlayer

stacking, became broader for supramolecule-based g-C3N4 and U, suggesting their crystallinities are lower than that of M. The results may indicate that polymer-like growth was dominated in U, MC, and MCB0.07 compared to the growth of a graphitic structure in M.15, 25, 26

S24

Figure S7. ATR-FTIR spectra of different g-C3N4 samples (U, M, MC, and MCB0.07). Six bands in the 1200-1650 cm-1 region (1630, 1530, 1450, 1395, 1313, and 1235 cm-1) characterize stretching modes of C-N heterocycles,27 and one band at 805 cm-1 characterizes triazine.23

S25

Figure S8. XPS stack plots of carbon 1s (left) and nitrogen 1s (right) spectra for different g-C3N4 samples (U, M, MC, and MCB0.07). a.u. represents arbitrary units. The lower C(1s) binding energy region (284.5 eV) was attributed to alkyl carbon (i.e. -C-C- or -C-H), most likely either adventitious carbon adsorbed on the surface or any sp3 graphitic carbon formed during pyrolysis.27 The peak at ca. 288 eV was attributed to sp2 carbon (i.e., -C-N=C-) present in the backbone of g-C3N4.28 The broad region at ca. 293 eV was attributed to π-π* excitations between the graphitic layers.29 Analysis of the N(1s) region revealed three primary regions characteristic of g-C3N4. The peak at ca. 398.5 eV (±0.3) represented sp2 nitrogen (i.e., -C-N=C-) in the backbone of g-C3N4.28 The peak at 400.6 eV was attributed to nitrogen (i.e., N-C(3)) at the structural edges of g-C3N4,30, 31 and the broad peak at ca. 404 eV resulted from π-π* excitations between the stacked layers.29, 32 S26

Figure S9. SEM images of supramolecular complex (MCB0.07 precursor) prepared in different solvents: (a) ethanol, (b) water.

S27

Figure S10. (a) N2 adsorption-desorption isotherms of different g-C3N4 samples (U, M, MC, and MCB0.07) and (b) corresponding pore-size distribution. STP represents standard temperature (273 K) and pressure (1 atm). Isotherms and pore volumes are highlighted in the insets.

S28

Figure S11. Band gap analyses of different g-C3N4 samples (U, M, MC, and MCB0.07).

S29

Figure S12. Optical absorbance of g-C3N4 suspensions (1 g L-1 of U, M, MC, and MCB0.07 in a phosphate buffer, 1 mM, pH 7.3).

S30

Figure S13. Photoluminescence emission spectra of g-C3N4 samples U, M, MC, and MCB0.07. The spectra of U, MC, and MCB0.07 are highlighted in the inset. a.u. represents arbitrary units.

S31

Figure S14. (a) Photocatalytic rate constants of atrazine degradation on MCB0.07 under visible light irradiation over multiple cycles. (b) ATR-FTIR spectra and (c) SEM images of fresh and recovered MCB0.07 before and after reactions. Error bars represent 95% confidence intervals. Experimental conditions: photocatalyst loading of 1 g L-1; atrazine initial concentration of 100 𝜇M; phosphate buffer of 1 mM, pH 7.3; and Xenon lamp irradiation,  > 400 nm.

S32

S33

Figure S15. (a) The reactivity inhibition of phenol and atrazine degradation on U and MCB0.07 in the presence of scavengers (10 mM of tert-butyl alcohol as a ⋅OH scavenger, 10 mM of ammonium oxalate as a hole scavenger, and 10 mM of L-histidine as a 1O2 scavenger). (b) The reactivity inhibition of phenol degradation on U in the presence of various scavengers (10 mM of tert-butyl alcohol as a ⋅OH scavenger, 10 mM of ammonium oxalate as a hole scavenger, 10 mM of L-histidine as a 1O2 scavenger, 2 U mL-1 of superoxide dismutase as a O2-⋅ scavenger, and 200 U mL-1 of catalase as a H2O2 scavenger). The reactivity inhibition (%) was calculated by dividing the reactivity difference without and with the scavenger by the reactivity without the scavenger. Experimental conditions: photocatalyst loading of 1 g L-1; phenol or atrazine initial concentration of 100 𝜇M; phosphate buffer of 1 mM, pH 7.3; and Xenon lamp irradiation,  > 400 nm.

S34

Figure S16. Atrazine degradation by (a) freshly prepared MCB0.07 suspension in simulated water samples, (b) 24 h-aged MCB0.07 suspension in simulated water samples, (c) freshly prepared MCB0.07 suspension in real water samples, and (d) 24 h-aged MCB0.07 suspension in real water samples under simulated visible sunlight irradiation (Xenon lamp, > 400 nm). The initial concentration of atrazine was 100 M in simulated water samples and 20 M in real water samples. Atrazine degradation was faster with the lower initial concentration of 20 M. Control experiments were conducted under the same experimental conditions in a phosphate buffer (pH 7.3, 1 mM) prepared from ultrapure water. The foulants or natural water constituents were amended into the phosphate buffer for simulated water tests. No phosphate buffer was used for the real water tests. Photocatalyst loading was 1 g L-1. S35

References 1. Hohenberg, P.; Kohn, W. Inhomogeneous electron gas. Phys. Rev. 1964, 136, B864-B871. 2. Kohn, W.; Sham, L.J. Self-consistent equations including exchange and correlation effects. Phys.Rev. 1965, 140, A1133-A1138. 3. Hutter, J.; Iannuzzi, M.; Schiffmann, F.; VandeVondele, J. CP2K: Atomistic simulations of condensed matter systems. WIREs Comput. Mol. Sci. 2014, 4, 15-25. 4. Perdew, J.P.; Burke, K.; Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 1996, 77, 3865-3868. 5. Godbout, N.; Salahub, D.R.; Andzelm, J.; Wimmer, E. Optimization of Gaussian-type basis sets for local spin density functional calculations. Part I. Boron through neon, optimization technique and validation. Can. J. Chem 1992, 70 (2), 560-571. 6. Goedecker, H.; Teter, M.; Hutter, J. Separable dual-space Gaussian pseudopotentials. Phys. Rev. B 1996, 54, 1703-1710. 7. Krack, M. Pseudopotentials for H to Kr optimized for gradient-corrected exchange-correlation functionals. Theo.Chem.Acc. 2005, 114, 145-152. 8. Lippert, G.; Hutter, J.; Parrinello, M. A hybrid Gaussian and plane wave density functional scheme. Mol. Phys. 1997, 92 (3), 477-487. 9. VandeVondele, J.; Krack, M.; Mohamed, F.; Parrinello, M.; Chassaing, T.; Hutter, J. QUICKSTEP: Fast and accurate density functional calculations using a mixed Gaussian and plane waves approach. Comput. Phys. Commun. 2005, 167 (2), 103-128. 10. Heyd, J.; Scuseria, G.E. Assessment and validation of a screened Coulomb hybrid density functional. J. Chem. Phys. 2004, 120, 7274-7280. 11. Guidon, M.; Hutter, J.; VandeVondele, J. Robust periodic Hartree-Fock exchange for largescale simulations using Gaussian basis sets. J. Chem. Theory Comput. 2009, 5 (11), 3010-3021. 12. Bani-Hashemian, M.H.; Brück, S.; Luisier, M.; VandeVondele, J. A generalized poisson solver for first-principles device simulations. J. Chem. Phys. 2016, 144 (4), 044113. 13. Greene, E.F.; Tauch, S.; Webb, E.; Amarasiriwardena, D. Application of diffuse reflectance infrared fourier transform spectroscopy (DRIFTS) for the identification of potential diagenesis and crystallinity changes in teeth. Microchem. J. 2004, 76 (1-2), 141-149. 14. Kim, Y.I.; Atherton, S.J.; Brigham, E.S.; Mallouk, T.E. Sensitized layered metal-oxidesemiconductor particles for photochemical hydrogen evolution from nonsacrificial electrondonors. J. Phys. Chem. 1993, 97 (45), 11802-11810. S36

15. Zhang, J.; Chen, X.; Takanabe, K.; Maeda, K.; Domen, K.; Epping, J.D.; Fu, X.; Antonietti, M.; Wang, X. Synthesis of a carbon nitride structure for visible-light catalysis by copolymerization. Angew. Chem. Int. Ed. 2010, 49 (2), 441-444. 16. Nalbandian, M.J.; Greenstein, K.E.; Shuai, D.; Zhang, M.; Choa, Y.; Parkin, G.F.; Myung, N.V.; Cwiertny, D.M. Tailored synthesis of photoactive TiO2 nanofibers and Au/TiO2 nanofiber composites: Structure and reactivity optimization for water treatment applications. Environ. Sci. Technol. 2015, 49 (3), 1654-1663. 17. Arana, J.; Dona-Rodriguez, J.M.; Portillo-Carrizo, D.; Fernandez-Rodriguez, C.; Perez-Pena, J.; Gonzalez Diaz, O.; Navio, J.A.; Macias, M. Photocatalytic degradation of phenolic compounds with new TiO2 catalysts. Appl. Catal., B 2010, 100 (1-2), 346-354. 18. Newcombe, D.A.; Crowley, D.E. Bioremediation of atrazine-contaminated soil by repeated applications of atrazine-degrading bacteria. Appl. Microbiol. Biotechnol. 1999, 51 (6), 877-882. 19. Borisover, M.; Sela, M.; Chefetz, B. Enhancement effect of water associated with natural organic matter (NOM) on organic compound-NOM interactions: A case study with carbamazepine. Chemosphere 2011, 82 (10), 1454-1460. 20. Frierdich, A.J.; Shapley, J.R.; Strathmann, T.J. Rapid reduction of N-nitrosamine disinfection byproducts in water with hydrogen and porous nickel catalysts. Environ. Sci. Technol. 2008, 42 (1), 262-269. 21. Harriott, P. Mass transfer to particles: Part I. Suspended in agitated tanks. AIChE J. 1962, 8 (1), 93-101. 22. Mackay, D.; Shiu, W.Y.; Ma, K.C., Illustrated Handbook of Physical-Chemical Properties of Environmental Fate for Organic Chemicals.; Taylor & Francis: 1997. 23. Khabashesku, V.N.; Zimmerman, J.L.; Margrave, J.L. Powder synthesis and characterization of amorphous carbon nitride. Chem. Mater. 2000, 12 (11), 3264-3270. 24. Butt, J.B., Mass Transfer in Heterogeneous Catalysis.; Massachusetts Institute of Technology Press: Cambridge, Mass, 1970. 25. Shalom, M.; Inal, S.; Fettkenhauer, C.; Neher, D.; Antonietti, M. Improving carbon nitride photocatalysis by supramolecular preorganization of monomers. J. Am. Chem. Soc. 2013, 135 (19), 7118-7121. 26. Dong, F.; Wu, L.; Sun, Y.; Fu, M.; Wu, Z.; Lee, S.C. Efficient synthesis of polymeric g-C3N4 layered materials as novel efficient visible light driven photocatalysts. J. Mater. Chem. 2011, 21 (39), 15171-15174.

S37

27. Li, X.; Zhang, J.; Shen, L.; Ma, Y.; Lei, W.; Cui, Q.; Zou, G. Preparation and characterization of graphitic carbon nitride through pyrolysis of melamine. Appl. Phys. A: Mater. Sci. Process. 2009, 94 (2), 387-392. 28. Liu, J.; Zhang, T.; Wang, Z.; Dawson, G.; Chen, W. Simple pyrolysis of urea into graphitic carbon nitride with recyclable adsorption and photocatalytic activity. J. Mater. Chem. 2011, 21 (38), 14398-14401. 29. Thomas, A.; Fischer, A.; Goettmann, F.; Antonietti, M.; Muller, J.; Schlogl, R.; Carlsson, J.M. Graphitic carbon nitride materials: Variation of structure and morphology and their use as metal-free catalysts. J. Mater. Chem. 2008, 18 (41), 4893-4908. 30. Kundu, S.; Xia, W.; Busser, W.; Becker, M.; Schmidt, D.A.; Havenith, M.; Muhler, M. The formation of nitrogen-containing functional groups on carbon nanotube surfaces: A quantitative XPS and TPD study. Phys. Chem. Chem. Phys. 2010, 12 (17), 4351-4359. 31. Raymundo-Piñero, E.; Cazorla-Amorós, D.; Linares-Solano, A.; Find, J.; Wild, U.; Schlögl, R. Structural characterization of N-containing activated carbon fibers prepared from a low softening point petroleum pitch and a melamine resin. Carbon 2002, 40 (4), 597-608. 32. Guo, Q.; Xie, Y.; Wang, X.; Zhang, S.; Hou, T.; Lv, S. Synthesis of carbon nitride nanotubes with the C3N4 stoichiometry via a benzene-thermal process at low temperatures. Chem. Commun. 2004, (1), 26-27.

S38