advances.sciencemag.org/cgi/content/full/5/11/eaax8275/DC1

Supplementary Materials for

Dissociating stable nitrogen molecules under mild conditions

by cyclic strain engineering

Gao-Feng Han, Xiang-Mei Shi, Seok-Jin Kim, Jeonghun Kim, Jong-Pil Jeon, Hyuk-Jun Noh, Yoon-Kwang Im, Feng Li, Young Rang Uhm, Chul Sung Kim, Qing Jiang*, Jong-Beom Baek*

*Corresponding author. Email: jbbaek@unist.ac.kr (J.-B.B.); jiangq@jlu.edu.cn (Q.J.)

Published 1 November 2019, Sci. Adv. 5, eaax8275 (2019)

DOI: 10.1126/sciadv.aax8275

This PDF file includes:

Supplementary Materials and Methods Supplementary Text Fig. S1. Comparison of the as-prepared samples in N2 and Ar. Fig. S2. GNPs after different nitrogenation times. Fig. S3. GNPs after different nitrogenation times. Fig. S4. Studies of the rotation speed of ball milling, the loading amount of graphite, the critical reaction pressure, and temperature of the ball-mill container. Fig. S5. Ball milling by ZrO2 balls. Fig. S6. The poisoning experiments. Fig. S7. The self-nitrogenating protection phenomenon of Fe balls. Fig. S8. The transfer experiments. Fig. S9. Extension to other carbon materials. Table S1. Elemental analysis of the GNP.

Supplementary Materials

Supplementary Materials and methods

Checkout pressure tightness of the container

In order to prevent any gas leakage during ball-milling, the pressure tightness of the container was

carefully checked by the following combined methods:

1. The working condition simulation. We checked the pressure tightness under actual

working conditions, except filled with inert Ar gas (9 bar). Due to the increase in surface

area during ball-milling, the as-prepared sample will adsorb gas, especially in the

beginning. Thus, this test was divided into two stages. In the first stage, ball-milling for 10

h, the change in surface area was negligible. After an additional 10 h ball-milling, the

volume change of charged gas should be less than 10 mL. This test was conducted in the

beginning.

2. Immersing the whole container in water. After being charged with N2 gas (8 bar), the

container was immersed in water, and held for 20 min to make sure no bubbles appeared.

We conducted this test two times, before and after ball-milling. The test after ball-milling

was to detect any leakage caused by the vigorous agitation. Unlike the previous working

conditions simulation test, this test was conducted for every experiment.

Calculation of consumed gas

The consumed gas (Vconsumed) was calculated by determining the difference in gas volume before

(Vcharged) and after ball-milling (Vremant).

Vconsumed = Vcharged – Vremant (S1)

The charged gas before ball-milling (Vcharged) was first measured by the water displacement

method, and then the container was filled again at the same pressure. The amount of remnant gas

after ball-milling (Vremant) was measured by the water displacement method.

The consumed gas (Vconsumed) was composed of two parts: one is the chemically fixed (Vchem); the

other is physically adsorbed (Vphys).

Vconsumed = Vchem + Vphys (S2)

It was noted that a small amount of the consumed gas was associated with physical adsorption

(Vphys). This was caused by an increase in the surface area of graphitic nanoplatelets (GNP) during

unzipping and delamination by ball-milling. The physically adsorbed gas (Vphys) was about 13.7

mL g−1

, which is approximately 3 times higher than the values determined with BET

measurements (4.6 mL g−1

). The possible reason is that there were many free edge sites before

exposure to air. Oxidation of the free edge sites would deteriorate the capacity for physical

adsorption.

Although the volume of consumed gas (Vconsumed) cannot be directly translated into the N content

in GNP, it can be used as an indirect indicator of N content. Since the preparation method is the

same, it results in the same amount of physical adsorption.

Characterizations

The material microstructures were characterized by high-resolution transmission electron

microscope (HR-TEM, JEM-2100). The surface morphology of the Fe balls was observed with

optical polarization microscopy (Leica DM 2500P). X-ray absorption spectroscopy was measured

at 6D UNIST-PAL beamline in the Pohang Accelerator Laboratory (Pohang, South Korea). X-

Ray diffraction data (XRD) were collected on a D/max2500V (Rigaku, Japan) using Cu-Kα

radiation (λ = 1.5418 Å) at a scan rate of 4° min−1

. Physical adsorption was analyzed by the

Brunauer-Emmett-Teller (BET) method using nitrogen adsorption-desorption isotherms on a

Micromeritics ASAP 2504N.

Fourier transform infrared spectra (FTIR) were recorded on a Perkin-Elmer Spectrum 100. The

samples were first mixed with potassium bromide (KBr, Sigma-Aldrich) by grinding, then were

pressed into pellets. The Raman spectra were obtained on a WITec Alpha300R with laser

wavelength of 532 nm.

The elemental analysis (EA) was performed on an Element Analyzer (TruSpec Micro CHNS,

LECO Corp.). All data were the average value of at least three analyses. The chemical

composition and bond natures were characterized by X-ray photoelectron spectra (XPS) on a

Thermo Fisher K-alpha XPS spectrometer. The deconvolution of the N 1s spectra was performed

by XPSPEAK 41, programmed by Raymund W.M. Kwok, the Chinese University of Hong Kong.

Time-of-flight secondary ion mass spectrometry (TOF-SIMS, TOF.SIMS5, IONTOF GmbH,

Germany) was used to analyze the trace composition of the Fe balls, and the as-prepared GNP.

The primary ion species was Bi with a dose of 2.0 × 109, and the raster area was about 100*100

µm2 for the Fe balls and 400*400 µm

2 for the as-milled GNP.

The Fe loading was determined by thermogravimetric analysis (TGA, STA8000, PerkinElmer

Instruments Co. Ltd, USA) at a heating rate of 5 °C min−1

in air. The thermal stability was

measured at a heating rate of 20 °C min−1

in N2. The sample loading was typically 30 µg. The

TGA was conducted by compensating buoyancy effect, which is caused by the change in gas

density. The buoyancy effect was eliminated by subtracting the background curve, which was

measured under the same conditions except without the sample.

Mossbauer spectra were taken in transmission mode with a conventional constant acceleration

spectrometer at room temperature. A 50 mCi 57

Co(Rh) was used as the γ-ray source. The isomer

shifts were calibrated by α-Fe foil. The Mossbauer spectra were fitted by using a least-squares

computer program. The isomer shift (IS), quadrupole shift (QS), hyperfine field at the site of the

Fe nucleus (H), and relative spectral area of Fe species were derived from the fitted spectra.

Supplementary Text

The method of calculating the thermal-equilibrium bulk concentration of N in the bulk

The thermal-equilibrium bulk concentration of N (wt.) at 1 bar of N2 is given by (15):

N (wt.) = 0.098 exp(− ΔHs ∕ RT) (S3)

where, R is the gas constant (1.987 cal mol−1

K−1

), the ΔHs (= + 7.2 kcal mol−1

) is the enthalpy for

the reaction:

1∕2 N2, gas ↔ N* (S4)

When T is room temperature (298 K, 25 °C), the mass percentage of N is calculated as 5.14 ×

10−7

. After transferring to atom percentage, the theoretical thermal-equilibrium bulk concentration

of Fe powders is 2.05 × 10−6

.

The method of calculating N content according to XRD patterns

The lattice constant has a relation with the doped content of N as follows (28):

a0, N = a0 + D (at.% N) (S5)

where the a0, N is the lattice constant after N doping, the a0 is the lattice constant of pure Fe, a0 =

2.8664 Å. D is the conversion factor.

The a0, N can be calculated with the combination of Bragg's law (Equation S6) and interplanar

spacing dhkl function (Equation S7):

n λ = 2 dhkl sinθ (S6)

dhkl = a0, N ∕ (h2 + k

2 + l

2)1∕2

(S7)

where the n is a positive integer, λ is the diffracted wavelength of Cu Kα, λ = 1.5418 Å. dhkl is the

distance between two adjacent and parallel planes of atoms, and h, k, and l are the Miller indices.

Here, we selected the strongest (110) facets in the XRD pattern. Since the sum of h, k, and l is

odd, thus n = 1. θ is the scattering angle, and the XRD pattern shows that θ is 22.052 °.

The a0, N is 2.9015 Å after calculation.

The D is derived from our DFT using Equation S5.

D = (a0, N, DFT − a0, DFT) ∕ (at.% N)

In our DFT model, the percentage of N is 1.8182%. The difference in lattice constant (a0, N, DFT −

a0, DFT) is 0.0180 Å. The D is calculated as 0.0099 Å.

Finally, in terms of Equation S5, the calculated N content is 3.5 at.%. This means that the N

atoms in the Fe powders are at thermal nonequilibrium.

The method of calculating nitrogen fixation yield

The nitrogen fixation yield is defined by normalizing the total amount of fixed nitrogen (N in the

obtained carbon materials) to the surface of Fe balls and the milling time. The calculation method

is as follows:

(S8)

where, the φnorm is the N content normalized to C (at∕at%), the M is the mass loading of graphite

(M = 15 g), the mN is the standard atomic weight of nitrogen (14), the mC is the standard atomic

weight of carbon (12), the N is the total numbers of the steel ball (N = 4477 for Φ 3 Fe balls, N =

962 for Φ 5 Fe balls), the A is the surface area of Fe balls (A = 0.283 cm2 for Φ 3 Fe balls, A =

0.785 cm2 for Φ 5 Fe balls), the T is the ball-milling time (h).

When the ball milling time is 3 h, the calculated fixation yield is 272.3 and 459.7 μg cm−2

h−1

for

Φ 3 and Φ 5 Fe balls, respectively.

The method of calculating maximum nitrogen fixation efficiency

The maximum fixation efficiency could be given by the following equations:

Fixation efficiency = 𝑃−𝑃𝑐𝑟𝑖𝑡

𝑃 (S9)

where the P is the pressure of total N2 supplied, and the Pcrit is the critical reaction pressure in the

reaction (0.8 bar). If the initial pressure is 8 bar, the maximum fixation efficiency is 90 %.

Fig. S1. Comparison of the as-prepared samples in N2 and Ar. XPS survey spectra of the

samples, which were prepared using Fe balls with diameters of (A) 3 mm and (B) 5 mm. The

FTIR spectra of the samples, which were prepared using Fe balls with diameters of (C) 3 mm and

(D) 5 mm. The Raman spectra of the samples, which were prepared using Fe balls with diameters

of (E) 3 mm and (F) 5 mm. The nitrogenation method was ball-milling graphite (15 g) with Fe

balls (Φ = 3 mm or 5 mm) in N2 or Ar (8 bar) at a rotation speed of 500 rpm for 10 h.

800 700 600 500 400 300 200

O 1s N 1s

Inte

nsity (

a.u

.)

Binding energy (eV)

N2, 5 mm

Ar, 5 mmC 1s

800 700 600 500 400 300 200

O 1s N 1sIn

tensity (

a.u

.)

Binding energy (eV)

N2, 3 mm

Ar, 3 mm

C 1s

4000 3000 2000 1000

N2, 5 mm

Tra

nsm

itta

nce (

a.u

.)

Wavenumber (cm-1)

Ar, 5 mm

u C-O

u C=Cu OH

u OH u CºNu C=Cu C-N, C-O

500 1000 1500 2000 2500 3000 3500

N2, 3 mm

Inte

nsity (

a.u

.)

Raman shift (cm-1)

Ar, 3 mm

4000 3000 2000 1000

u OH

N2, 3 mm

u C-N, C-Ou C=C

u OH

Tra

nsm

itta

nce (

a.u

.)

Wavenumber (cm-1)

Ar, 3 mm

u CºN

u C=C

u C-O

500 1000 1500 2000 2500 3000 3500

N2, 5 mm

Inte

nsity (

a.u

.)

Raman shift (cm-1)

Ar, 5 mm

BA

DC

FE

Fig. S2. GNPs after different nitrogenation times. (A) The XPS survey spectra. (B) The XPS N

1s spectra. (C) The XPS C 1s spectra. (D) The FTIR spectra. (E) The Raman spectra. (F) The

XRD patterns. The nitrogenation method was ball-milling graphite (15 g) with Fe balls (Φ = 3

mm) in N2 (8 bar) at a rotation speed of 500 rpm.

292 290 288 286 284 282

C 1s

30 h10 h3 h2 h

Inte

nsity (

a.u

.)

Binding energy (eV)

1 h

C-N

800 700 600 500 400 300 200

30 h

10 h

3 h

2 hIn

tensity (

a.u

.)

Binding energy (eV)

1 h N 1sO 1s

C 1s

10 20 30 40 50

30 h

10 h

3 h

2 h

Inte

nsity (

a.u

.)

2q ()

1 h

500 1000 1500 2000 2500 3000 3500

0 h

1 h

2 h

3 h

30 h

10 h

Inte

nsity (

a.u

.)

Raman shift (cm-1)

4000 3000 2000 1000

1 h

2 h

3 h

30 h

10 hT

ransm

itta

nce (

a.u

.)

Wavenumber (cm-1)

u OHu CºN

u C=C

u C-N, C-O

406 404 402 400 398 396

30 h

10 h

3 h

2 h

Inte

nsity (

a.u

.)

Binding energy (eV)

1 hN 1sBA

DC

FE

Fig. S3. GNPs after different nitrogenation times. (A) The FTIR spectra. (B) The Raman

spectra. The nitrogenation method was ball-milling graphite (15 g) with Fe balls (Φ = 5 mm) in

N2 (8 bar) at a rotation speed of 500 rpm.

500 1000 1500 2000 2500 3000 3500

0 h

1 h

3 h

30 h

10 h

Inte

nsity (

a.u

.)

Raman shift (cm-1)

4000 3000 2000 1000

1 h

30 h

10 h

3 h

Tra

nsm

itta

nce (

a.u

.)

Wavenumber (cm-1)

u OH u CºN

u C=C

u C-N, C-O

B

A

Fig. S4. Studies of the rotation speed of ball milling, the loading amount of graphite, the

critical reaction pressure, and temperature of the ball-mill container. (A) The nitrogen

content in as-prepared nitrogenated GNP as a function of rotation speed. The rotation speed was

selected at 500, 400, 300, and 200 rpm, respectively. The nitrogenation method was ball-milling

graphite (15 g) with Fe balls (Φ = 3 mm) in N2 (8 bar) for 6 h. (B) The nitrogenation method was

ball-milling graphite (5, 10, 15, and 20 g) with Fe balls (Φ = 5 mm) in N2 (8 bar) at a rotation

speed of 500 rpm for 3 h. (C) The void in the container was calculated by the slope of the charged

gas volume (Vcharged) as a function of charged pressure. The N2 volume was measured by water

displacement. The obtained void was about 242.4 mL. (D) Results of the critical reaction pressure

study, after ruling out the physical adsorption. The graphite was first pretreated by bal-milling for

6 h in N2 (8 bar) to make the surface area stable. The further nitrogenation method was ball-

milling graphite (15 g) with Fe balls (Φ = 5 mm) in N2 (2.0 bar, true pressure) at a rotation speed

of 500 rpm for 3 h. The N2 volume and N2 pressure are the truth volume and pressure in the

container. (E) The normalized content of N to C, determined by EA. The container temperature

was adjusted by changing the ball-milling and waiting (cooling) time. The nitrogenation method

was ball-milling graphite (15 g) with Fe balls (Φ = 5 mm) in N2 (8 bar) at a rotation speed of 500

rpm for 3 h.

200 300 400 5000

3

6

9

Norm

. N

to C

(at

at%

)

Rotation speed (rpm)

R2 = 0.996

0

2

4

6

mill 10 min, wait 30 min

Norm

. N

to C

(at

at%

)

mill 60 min, wait 10 min

0

200

400

600

N2 V

olu

me (

mL)

Before nitrogenation

After nitrogenation

0.0

0.5

1.0

1.5

2.0

2.5

N2 p

ressure

(bar)

Independent experiments

2 4 6 8500

1000

1500

2000

Vcharg

ed (

mL)

Pressure (bar)

Slope = 242.4

R2 = 0.9992

0

30

60

90

120

150

Norm

. V

consum

ed (

mL g

-1)

5 10 15 20

0

2

4

6

8

10

Loading of graphite (g)

Fe c

onte

nt

in G

NP

(w

t%)

A C

B

D

E

Fig. S5. Ball milling by ZrO2 balls. (A) The nitrogen content in as-prepared nitrogenated GNP

for Fe and ZrO2 balls. The nitrogenation method was ball-milling graphite (15 g) with Fe or ZrO2

balls (Φ = 3 mm) in N2 (8 bar). For fair comparison, the input kinetic energy (MFe ω2

Fe = MZrO2 ω2

ZrO2) and total rotation number (TFe ωFe = TZrO2 ωZrO2, 180,000 cycles) are same. The rotation

speed are 300 rpm and 353 rpm for Fe and ZrO2 balls, respectively. (B) The Zr3d XPS spectrum

of the worn-off ZrO2 debris in GNP.

190 188 186 184 182 180 178

ZrNOx

Counts

(a.u

.)

Binding Energy (eV)

Zr3dB

0

1

2

3

4

ZrO2N

orm

. N

to C

(at

at%

)Fe

A

Fig. S6. The poisoning experiments. A small amount of triethanolamine (TEA), glass powders,

KCl, and Na2S were separately added with graphite before ball-milling. The nitrogenation method

was ball-milling graphite (15 g) with Fe balls (Φ = 5 mm) in N2 (8 bar) at a rotation speed of 500

rpm for 3 h.

0

300

600

900

1200

KClNa2

STEA Glass

Vconsum

ed (

mL)

N/A

Fig. S7. The self-nitrogenating protection phenomenon of Fe balls. (A) Digital photos of the

Fe balls. (B) Optical microscope image after ball-milling in N2 gas. (C) Optical microscope image

after ball-milling in Ar gas. The darkened zones are the remnant graphitic nanoplatelets, or

oxidized Fe particles. Even after careful rinsing, the surface cannot be completely cleaned. The

anti-abrasion ability was derived from the Fe content in nitrogenated GNP. (D) The Fe content in

GNP. (E) The enhanced factor. The Fe content was determined by TOF-SIMS for Φ 3 mm, and

TGA for Φ 5 mm. The nitrogenation method was ball-milling graphite (15 g) with Fe balls (Φ = 3

mm or 5 mm) in N2 or Ar (8 bar) at a rotation speed of 500 rpm for 10 h.

10-5

10-4

10-3

10-2

10-1

100

Fe c

onte

nt

in G

NP

(w

t/w

t)

N2 Ar

F 5*

F 3

F 5*

F 3

0

1

2

3

Enhanced f

acto

r (v

s in N

2)

F 3 F 5*

B

A

E

D

C

Fig. S8. The transfer experiments. (A) XPS survey spectra of the mixture Fe and GNP (Fe +

GNP) in well transferred GNP. (B) The high-resolution C 1s of the Fe + GNP mixture in well

transferred GNP. (C) The FTIR of the transferred GNP. (D) The TGA of the transferred GNP in

N2. The buoyancy effect was eliminated by subtracting the background during TGA

measurements.

200 400 600 80080

85

90

95

100

Weig

ht

(%)

Temperature (°C)

-0.05

-0.04

-0.03

-0.02

-0.01

0.00

Transferred GNP

Derivative w

eig

ht

(%/°

C)

4000 3000 2000 1000

u organic carbonate

u C-H

d C-H

u C-N

r C-H

Tra

nsm

itta

nce (

a.u

.)

Wavenumber (cm-1)

u C=C

u CºN

u OH

296 294 292 290 288 286 284 282

Fe + GNP

Inte

nsity (

a.u

.)

Binding energy (eV)

Transffered GNP

800 700 600 500 400 300 200

C 1sFe 2p O 1s

N 1s

Inte

nsity (

a.u

.)

Binding energy (eV)

Fe + GNP

Transferred GNP

BA

DC

Fig. S9. Extension to other carbon materials. The nitrogenation method was ball-milling

carbon black (10 g) or graphite (15 g) with Fe balls (Φ = 5 mm) in N2 (8 bar) at a rotation speed

of 500 rpm for 30 h. The N2 was recharged every 10 h.

0 5 10 15 20 25 300

50

100

150

200

Norm

. V

consum

ed (

mL g

-1)

Milling time (h)

Graphite Carbon black

Table S1. Elemental analysis of the GNP. The content of O and Fe are not provided here

because they could not be detected by our EA device. The O originates from the oxidation of free

edge sites after exposure to air. The Fe content was measured using TGA or TOF-SIMS.

No. Total

time

(h)

Ball

diameter

(mm)

Gas Elemental content (wt%) Norm. N

(at∕at%)

Comments

C H N

1 1∕3 3 N2 93.9 - - -

Φ 3 mm, different

time in N2

2 1 3 N2 85.5 - 2.7 2.7

3 2 3 N2 72.3 0.3 4.0 4.7

4 3 3 N2 80.0 0.3 5.5 5.9

5 6 3 N2 78.3 0.3 7.8 8.6

6 10 3 N2 74.6 0.5 9.7 11.1

7 30 3 N2 71.7 0.2 12.0 14.3

8 10 3 Ar 94.0 - 0.7 0.65

9 1∕3 5 N2 93.0 - 0.9 0.79

Φ 5 mm, different

time in N2

10 1 5 N2 86.9 0.7 2.6 2.5

11 2 5 N2 84.5 0.7 4.5 4.6

12 3 5 N2 83.3 0.7 5.8 6.0

13 6 5 N2 83.3 0.7 8.3 8.5

14 10 5 N2 78.4 0.3 10.4 11.4

15 30 5 N2 75.4 0.7 14.4 16.3

16 3 5 N2 84.7 0.8 5.6 5.7 Long cooling time

17 10 5 Ar 80.4 2.7 4.7 5.0 Transferred GNP

18 3 5 N2 85.5 0.5 5.7 5.7 TEA

19 3 5 N2 84.7 0.4 5.6 5.6 Glass

20 3 5 N2 74.3 0.4 2.4 2.8 KCl

21 3 5 N2 77.1 0.6 1.5 1.6 Na2S

22 6 3 N2 86.6 0.5 6.0 6.0 400 rpm

23 6 3 N2 89.9 0.3 3.5 3.3 300 rpm

24 6 3 N2 97.4 0.2 1.5 1.3 200 rpm

25 6 3 N2 92.4 0.3 1.8 1.7 ZrO2, 353 rpm