1

Supplementary Information

A Phenomenological Model for Selective Growth of

Semiconducting Single-Walled Carbon Nanotube Based on

Catalyst Deactivation

Shunsuke Sakurai, Maho Yamada, Hiroko Nakamura, Atsuko Sekiguchi, Fumiaki Tanaka,

Don N. Futaba, and Kenji Hata*

National Institute of Advanced Industrial Science and Technology (AIST), Central 5, 1-1-1,

Higashi, Tsukuba, Ibaraki, 305-8565, Japan.

*Author to whom correspondence should be addressed.

1. Measurement and minimization of the background gas level in the furnace

In this study, the background oxidative gases has been minimized mainly by the

following three measures. First, the gas exchange chamber (Fig. 3a) was equipped to prevent the

contamination of oxygen and water vapor from outside of the furnace. Second, to avoid the

generation of carbon monoxide (CO) via the reaction between carbon dirt impurities in the

furnace and water vapor during the catalyst conditioning process (water gas reaction: C + H2O

→ CO + H2), furnace tube and sample holders made of quartz were heated (800 °C, 10 minutes)

in an O2 atmosphere (diluted to about 20% with He) before each CVD process. Using Fourier

transform infrared spectroscopy (Thermo scientific, Nicolet6700), we confirmed that CO

concentration in the emitted gas from the furnace during CVD processes is less than 0.1 ppm, it

reached up to 20 ppm without O2 cleaning. Third, after O2 cleaning, the furnace and sample

holders were annealed in a hydrogen atmosphere (100%) for 5 minutes to decrease oxygen atoms

adsorbed on the furnace. During this hydrogen annealing, water concentration in the exhausted

gas typically reached to 100 ppm. By inserting this hydrogen annealing process, the increase in

water vapor concentration during the catalyst formation process was less than 2 ppm.

2. Supplemental Raman spectra

Electronic Supplementary Material (ESI) for Nanoscale.This journal is © The Royal Society of Chemistry 2015

2

Raman shift (cm−1)

Inte

nsity

(a.u

.)

532 nm

Selective CVD

1700 1600 1500 1400 1300 1200

Raman shift (cm−1)

Inte

nsity

(a.u

.)

532 nm

Non-selective CVD

1700 1600 1500 1400 1300 1200

Raman shift (cm−1)

Inte

nsity

(a.u

.)

532 nm

300 250 200 150 100Raman shift (cm−1)

Inte

nsity

(a.u

.)

532 nm

300 250 200 150 100

Raman shift (cm−1)

Inte

nsity

(a.u

.)

633 nm

300 250 200 150 100Raman shift (cm−1)

Inte

nsity

(a.u

.)

633 nm

300 250 200 150 100

Raman shift (cm−1)

Inte

nsity

(a.u

.)

785 nm

300 250 200 150 100

Raman shift (cm−1)

Inte

nsity (

a.u

.)

785 nm

300 250 200 150 100

(a)

(b)

(c)

(d)

AverageAverage

Average

Average

Average

Average

AverageAverage

Figure S1. Raman spectra of selective (left) and non-selective CVD (right) collected from

more than 25 positions. Red spectra show the average of spectra from all positions. (a)

Spectra showing G-band region (excitation wave length: 532 nm). (b) Spectra showing

RBM peaks (excitation wave length: 532 nm). (c) Spectra showing RBM peaks (excitation

wave length: 633 nm). (d) Spectra showing RBM peaks (excitation wave length: 785 nm).

3

H2O 0.25 % 0.5 % 1 % 1.5 % 2 %

0 ppm

2 ppm

10 ppm

20 ppm

30 ppm

40 ppm

50 ppm

Raman shift (cm-1)

H2

300 200 100 300 200 100 300 200 100 300 200 100 300 200 100

H2O 0.25 % 0.5 % 1 % 1.5 % 2 %

0

2 ppm

10 ppm

20 ppm

30 ppm

40 ppm

50 ppm

Raman shift (cm-1)

H2

300 200 100 300 200 100 300 200 100 300 200 100 300 200 100

Figure S2. Raman spectra showing RBM peaks of SWCNTs synthesized with different

water and hydrogen amounts during catalyst conditioning process. Upper: excitation

wavelength of 532 nm, lower: 785 nm.

4

300 200 100300 200 100

Raman shift (cm-1) Raman shift (cm-1)

785 nm532 nm

S22

M11

3. Reference experiment feeding water vapor during SWCNT growth process

To exclude the possibility that the selective etching of m-SWCNT, previously reported

mechanism of selective growth, worked mainly in the current work, a reference SWCNT growth

experiment was conducted by the following process. Catalyst formation process was conducted

as same as described in the manuscript. Then, without catalyst conditioning process, water vapor

(50 ppm) and hydrogen (0.25%) was introduced into the furnace during SWCNT growth process

with ethylene flow (0.25%). Raman spectra (Figure S4) shows a significant RBM peaks

corresponding to m-SWCNT, resulting in a greatly decreased selectivity of Psmall = 53% from

CVD process with catalyst conditioning process using same amount of water and hydrogen.

M11

S33

M11

S22

785 nm

532 nm

Si

Raman shift (cm−1)

Inte

nsity

(a.

u.)

100200300

Figure S4. Raman spectra of SWCNTs synthesized in the reference experiment. Upper: 532

nm excitation wavelength, lower: 785 nm. Raman shift regions indicated by red and blue

squares represent the signals from m-SWCNT and sc-SWCNT, respectively.

Figure S3. Raman spectra of HiPco. Left: 532 nm excitation wavelength, right: 785 nm.