Carbon Nanotube Chirality Determines Properties ... - ACS Publications

Subscriber access provided by University of Sussex Library

Communication

Carbon nanotube chirality determines properties of encapsulated linear carbon chain Sebastian Heeg, Lei Shi, Lisa V Poulikakos, Thomas Pichler, and Lukas Novotny Nano Lett., Just Accepted Manuscript • DOI: 10.1021/acs.nanolett.8b01681 • Publication Date (Web): 08 Aug 2018 Downloaded from http://pubs.acs.org on August 9, 2018

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 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 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.

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 22 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

Nano Letters

Carbon nanotube chirality determines properties of encapsulated linear carbon chain Sebastian Heeg,† Lei Shi,‡ Lisa V. Poulikakos,¶ Thomas Pichler,‡ and Lukas Novotny∗,† †ETH Zürich, Photonics Laboratory, 8093 Zürich, Switzerland ‡University of Vienna, Faculty of Physics, 1090 Wien, Austria ¶ETH Zürich, Optical Materials Engineering Laboratory, 8093 Zürich, Switzerland E-mail: [email protected]

Abstract Long linear carbon chains (LLCCs) encapsulated inside double-walled carbon nanotubes (DWCNTs) are regarded as a promising realization of carbyne, the truly onedimensional allotrope of carbon. While the electronic and vibronic properties of the encapsulated LLCC are expected to be influenced by its nanotube host, this dependence has not been investigated experimentally so far. Here we bridge this gap by studying individual LLCCs encapsulated in DWCNTs with tip-enhanced Raman scattering (TERS). We reveal that the nanotube host, characterized by its chirality, determines the vibronic and electronic properties of the encapsulated LLCC. By choice of chirality, the fundamental Raman mode (C-mode) of the chain is tunable by ∼ 85 cm−1 and its band gap by ∼ 0.5 eV, suggesting this one-dimensional hybrid system to be a promising building block for nanoscale optoelectronics. No length dependence of the chain’s C-mode frequency is evident, making LLCCs a close to perfect representation of carbyne.

1

ACS Paragon Plus Environment

Nano Letters 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

Keywords linear carbon chains, carbyne, carbon nanotubes, Raman spectroscopy, TERS

Main manuscript Carbyne by definition is an infinitely long linear carbon chain with sp 1 hybridization that forms the truly one-dimensional allotrope of carbon at the one-atom cross section limit. 1–3 Its anticipated stiffness, strength, and elastic modulus exceed that of any other known material. 4 In its most common form, carbyne is a polyyne with alternating single and triple bonds originating from a Peierls distortion. 5 This bond-length alternation (BLA) dominates the electronic and vibronic structure of carbyne. 6–8 It opens up a direct band gap that is sensitive to external pertubations thus offering tunability. Finite linear carbon chains are expected to exhibit the properties of carbyne if they consist of 100 or more atoms as the BLA saturates. 1 Exploring the fundamental properties of carbyne experimentally, however, has long been hindered by its extreme chemical instability and short chain lengths of up to 44 atoms. 9 The synthesis of linear carbon chains inside carbon nanotubes, sketched in Fig. 1(a), overcomes these obstacles. 10–16 The tubes act as nano-reactors, prevent chemical interaction of the chains with the environment, and allow for long chain lengths. A major leap forward in forming carbyne are long linear carbon chains (LLCCs) with lengths up to several hundreds of nanometers that have recently been synthesized inside double-walled carbon nanotubes (DWCNTs). 17–21 The local environment inside a nanotube, characterized by its chirality, is expected to affect encapsulated chains through interactions such as van-der-Waals (vdW) forces, charge transfer or dielectric screening, and may even limit the chain length. 10–20 These interactions vary with chirality, modify the BLA, and effectively mask the intrinsic properties of LLCCs. 22,23 =Recently, it was even suggested that there is a residual length dependence of the chain’s properties far beyond the currently accepted limit of 100 atoms. 17,18,22 This would mean that LLCCs are not the finite realization of carbyne and that the currently accepted 2


Page 2 of 22

Page 3 of 22 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

Nano Letters

models on the chain’s length dependence fail.However, while the properties of confined long linear carbon chains are governed by their nanotube host and potentially by their length, the correspondence between the host tube’s chirality, the properties of the encapsulated chain, and the length of the chain has not been investigated experimentally. Raman spectroscopy is an excellent tool to study the properties of the encapsulated chain and the characteristics of the encasing CNT. The dominant Raman mode of carbyne (C-mode) reports the bond-length alternation of the chain. 1 The chirality specific Raman features of CNTs, in particular the radial breathing mode (RBM), are very well understood. 24,25 Accordingly, Raman measurements of nanotubes hosting LLCCs allow us to correlate the electronic and vibronic properties of the LLCC to the properties of the nanotube and help to identify the dominating interaction between host and guest. Importantly, such a correlation cannot be established through bulk measurements because it is impossible to verify that the Raman signal of both chain and tube arises from the same pair. While this can be achieved through confocal Raman measurements, the length of the investigated chain remains unknown. Tip-enhanced Raman scattering (TERS), on the other hand, offers the nanoscale spatial resolution necessary to find and characterize individual pairs of carbon nanotube and chain, measure the length of the chain, and unveil the correlation between the nanotube’s chirality and the chain’s properties. 17–19 In this letter, we present TERS measurements that quantitatively link the C-mode frequency of long linear carbon chains entirely to the chirality of the encapsulating carbon nanotubes. We observe that the frequency of the C-mode decreases as the inner nanotube diameter is reduced, pointing towards van-der-Waals forces as the dominating interaction between nanotube and chain. The chain’s properties show no length dependence and have hence converged to the limit of infinite length. This provides strong evidence that long linear carbon chains encapsulated in carbon nanotubes are the finite realization of carbyne in different local environments. Details of our experimental setup have been described in previous reports. 26–28 In short,

3


(a)

(b)

200 nm

4.8

120 kHz

50

0

0

C-mode

G

RBM x10

23 nm

800 1200 Raman shift (cm -1)

(d)

100 nm

1600

1

I(z) ~ e-z/τ

(ITERS-ICF)/(ITERS(z=0)-ICF)

(c)

Page 4 of 22

tip down - TERS tip up - CF

b

400

photons (kcts/ms)

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

Normalized Raman intensity (a.u.)

Nano Letters

τ = 2.5 nm

0 0

10

20

30

z(nm)

Figure 1: TERS characterization of long linear carbon chains. (a) Sketch of a linear carbon chain (red) encapsulated in a double-walled carbon nanotube (blue/yellow). (b) Confocal (black) and TERS (red) Raman spectra of the same double-walled carbon nanotube and encapsulated long linear carbon chain. The RBM arises from a (6,4) inner carbon nanotube. (c) TERS image of a linear carbon chain. The image contrast corresponds to the intensity of the C-mode centered at 1798 cm−1 . An intensity profile (inset) extracted along the white line shows a resolution of ∼ 23 nm. (d) Normalized TERS C-mode intensity as function of the separation z between tip and sample. a focused radially-polarized laser beam (λ = 633 nm) irradiates a gold pyramid fabricated by template-stripping. 28 The enhanced field at the tip apex defines a localized excitation source for Raman scattering. Raman scattered light is collected in backscattering configuration either by a combination of bandpass filters that transmit the spectral region of the linear carbon chain’s C-mode (1770 cm−1 to 1870 cm−1 ) 10–21 followed by a single-photon counting avalanche detector, or by a combination of spectrograph and charge-coupled device (resolution ∼ 3 cm−1 ). To avoid sample degradation due to heating, we used laser powers of 115 µW for TERS and 1.15 mW for confocal (CF) imaging. Integration times are 25 ms/pixel for imaging and up to 60 s for acquiring full Raman spectra. LLCCs were grown inside double-walled carbon nanotubes (DWCNTs) in a high-temperature and high vacuum process as described in Ref. 17. The tubes were then purified, individualized, and dispersed 4


Page 5 of 22 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

Nano Letters

on thin glass cover slides using chlorosulfonic acid in an oxygen free atmosphere. To locate individual carbon chains inside DWCNTs we first perform confocal Raman raster scans (not shown). Once the signature of a linear carbon chain is detected, we verify its presence by observing the C-mode of the chain in a confocal Raman spectrum (black) as shown in Fig. 1(b). We then position the tip near the surface (< 1 nm) and record high resolution topographical (see Supporting Information) and TERS images as shown in Fig. 1(c). The TERS image reveals that the DWCNT contains a linear carbon chain of length 200 nm. The inset in Fig. 1(c) shows a line profile (white) extracted from the TERS image indicating a resolution of 23 nm. The TERS spectrum (red), shown in Fig. 1(b), exhibits a peak at 1798 cm−1 originating from the chain’s C-mode, a longitudinal-optical phonon that arises from the in-phase stretching of the triple-bonds along the chain. It is the only Raman-active phonon in polyynic carbyne. The group of peaks with the dominant component at 1588 cm−1 are the G-modes, typical signatures of CNTs. 24 The strong RBM at 348 cm−1 belongs to the inner carbon nanotube that is in resonance with the excitation wavelength. 29,30 The RBM corresponds to a radial displacement of the carbon atoms forming the nanotube and enables us to determine its chirality. 24,25,31 We do not observe any signature of the defect-induced D-mode (∼ 1350 cm−1 ), indicating that the nanotubes are pristine and largely free of defects. 24 Upon increasing the tip-sample distance the Raman intensity drops exponentially as shown for the C-mode in Fig. 1(d), which confirms the local and enhanced nature of our TERS signal as compared to the confocal Raman spectra (black) in Fig. 1(b). Notably, the enhancement of the C-mode is smaller than that of the CNT Raman modes, which is currently being studied. The difference may be attributed i.e. to the effect of spatial coherence or screening effects that reduce the near-field intensity in the CNT’s interior where the chain is located. 24,32 We identified and characterized 14 different chain-nanotube systems that exhibit both a single C-mode frequency and a single RBM frequency. We present the corresponding TERS spectra with a focus on RBMs in Fig. 2(a) and C-modes in Figs. 2(b). As we are

5


Nano Letters

(b)

Normalized C-mode Raman intensity

(a)

Normalized RBM Raman intensity

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

250

300

350

400

450

1700 1750 1800 1850 1900

Raman shift (cm -1)

Raman shift (cm -1)

Figure 2: TERS spectra of 14 pairs of long linear carbon chains and encapsulating carbon nanotubes. (a) RBM spectra of the inner nanotube from highest (top) to lowest frequency. The RBMs correspond to the (6, 4), (6, 5), (7, 2), and (8, 0) chiralties, respectively, see main text and Supporting information. (b) C-mode Raman modes of the chains corresponding to the RBMs shown in (a). RBM and C-mode spectra are normalized independently to the same peak height. interested in peak positions only, the RBMs and C-modes are independently normalized to the same amplitude and offset for clarity with decreasing RBM frequency from top to bottom. For all spectra in Fig. 2, a single RBM can clearly be assigned to a corresponding single C-mode. Around 40 additional TERS measurements showed a single C-mode but lacked RBM signatures, which shows that the resonance condition of the CNTs is the bottleneck in our experiments. Further ∼ 20 measurements showed multiple RBM and/or C-modes – prohibiting the correlation between the nanotube’s chirality and the chain’s properties – and were discarded as they arose from bundles or CNT agglomerates. Some of the RBMs and C-modes in in Fig. 2(a) have slight shape asymmetries. We primarily assign these to the limited spectral resolution in our TERS experiments, as we sacrifice spectral resolution for the high signal throughput necessary to detect faint RBM signals even with TERS enhancement. A detailed analysis of the C-mode’s peak shape and

6


Page 6 of 22

Page 7 of 22

(b )

(a )

# c a rb o n a to m s 1 0 0 0 2 0 0 0 0

1 8 4 0

C -m o d e + R B M R e f. [1 7 ]

1 8 3 0

1 8 3 0

1 8 2 0

1 8 2 0

(6 ,5 )

1 8 1 0

1 8 1 0

1 8 0 0

1 8 0 0

(7 ,2 )

1 7 9 0

1 7 9 0

1 7 8 0

(6 ,4 )

1 7 8 0

C - m o d e R a m a n s h ift ( c m

-1

-1

)

)

1 8 4 0


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

Nano Letters

(8 ,0 ) 1 7 7 0

1 7 7 0 3 2 0

3 3 0

3 4 0

3 5 0

3 6 0

3 7 0

R B M

R a m a n s h ift ( c m

3 8 0 -1

)

3 9 0

0

1 0 0

2 0 0

L L C C le n g th ( n m )

Figure 3: Correlation between C-mode and RBM. (a) C-mode of encapsulated linear carbon chains vs radial breathing mode of the encasing inner carbon nanotubes extracted from the TERS spectra shown in Fig. 2. Labels indicate the the assignment of the RBMs to (n,m) chiralities, see main text and Supporting Information. (b) Length of LLCCs shown in (a) as observed by TERS in nm and as number of carbon atoms # forming the LLCC with 0.2558 nm lattice constant. 17 width with high spectral resolution is beyond the scope of this work and will be the subject of future studies. We now analyze in detail the correlation between chain and encasing nanotube by extracting the peak positions (single Lorentzian fits) from Fig. 2 and plotting the C-mode frequency as a function of RBM frequency in Fig. 3(a), including the singular data point from Ref. 19. The plot reveals three distinct characteristics. First, each RBM – representing the chirality of the encasing nanotube – uniquely corresponds to one specific C-mode frequency – representing the linear chain – to within ± 3 cm−1 . We refer to such a combination of RBM and C-mode from now on as ’pairing’. Second, specific pairings of RBM and C-mode occur multiple times. The pairing for the RBM at 348 cm−1 and the C-mode at 1798 cm−1 , dashed circle in Fig. 3(a), is particularly pronounced and applies to half of our measurements. These RBMs arise from the same inner nanotube chirality and the encapsulated chains show similar Raman frequencies. The length of the chains, depicted in Fig. 3(b) for all measured pairings, does not affect the chain’s C-mode frequencies. This is the core observation of 7


Nano Letters 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

our experiments because it is in stark contrast to the pronounced length dependence of the C-mode frequencies for short linear carbon chains. 2,3 Our measurements reveal a unique correspondence between nanotube RBM and carbon chain C-mode frequencies, independent of the length of the carbon chain as observed by TERS. We conclude that the inner carbon nanotube chirality represented by this RBM uniquely determines the chain’s bond-length alternation and hence its vibronic and electronic structure. Before we assign RBMs to nanotube chiralities, we verify that the presence of a linear carbon chain does not modify the RBMs of inner and outer nanotubes, which could otherwise falsify our chirality assignment. We show this for one of the pairings (326 cm−1 /1835 cm−1 ), where we also observe the RBM of the outer nanotube at 187 cm−1 . This pairing is compared to an empty DWCNT formed by the same inner and outer nanotube chiralities, see Supporting Information. For both the empty and the filled DWCNT, the RBM frequencies of the inner and outer nanotube each agree to within 1 cm−1 . This uncertainty is negligible compared to the effect of the outer CNT discussed in the next paragraph, and will not be taken into account when assigning RBMs to nanotube chiralities. Assigning RBMs to chiralities is not straightforward for double-walled carbon nanotubes, because one particular inner tube can reside in different outer tubes. This variation in the diameter difference between inner and outer tube influences the wall-to-wall interactions. It renders the simple inverse relation between nanotube diameter and RBM frequency for singlewalled carbon nanotubes invalid. 24,31 Instead, the inner tube’s RBM frequency increases by up to 35 cm−1 depending on the outer nanotube’s diameter, while the diameters of both inner and outer tubes effectively remain unchanged. 29,30,33,34 As a consequence, DWCNT spectra show more RBMs than there are chiralities and RBMs of different frequencies may belong to the same inner tube chirality. This means that we cannot directly extract a linear trend from Fig. 3(a), and that each RBM frequency has to be analyzed and assigned to a nanotube chirality individually. We perform this assignment based on experimental Raman studies on DWCNTs and describe it in great detail in the Supporting Information. In ambigious cases,

8


Page 8 of 22

Page 9 of 22 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

Nano Letters

we additionally make use of the chiral angle dependence of the nanotube’s G− -mode for small tube diameters and arrive at a robust chirality assignment for all RBMs shown in Fig 3(a). Table 1 lists our tentative assignment of the pairings of RBM and C-mode in Fig. 3(a) to the chiralities of the inner carbon nanotubes and the corresponding tube diameters d. We find that the two RBMs around 327 cm−1 and 335 cm−1 belong to the (6, 5) inner nanotube, while the RBMs around 348 cm−1 and 364 cm−1 belong to the (6, 4) inner nanotube. These assignments intuitively make sense because chains with similar C-modes are allocated to inner tubes of the same chirality. The variation in the C-mode frequency for different pairings belonging to the same inner nanotube is then a measure for the effect of the outer nanotube on the encapsulated chain.We find this variation once for the (6, 5) nanotube, for which the difference in C-mode frequency amounts to 4 cm−1 , c.f. Table 1. This difference may arise from the fact that one outer nanotube is metallic while the other outer tube is very likely semiconducting, see Supporting Information for a detailed discussion. Our third observation from the plot in Fig. 3(a) is the decrease of the C-mode frequency with increasing RBM frequency. Making use of our assignment of RBMs to inner nanotube chiralities, c.f. Table 1, we plot in Fig. 4(a) the C-mode frequencies of the encapsulated chains as a function of the encasing inner carbon nanotube’s diameter. The data shows a clear decrease in the chain’s C-mode frequency ωC with decreasing diameter d and yields the linear relation ωC (d) = a + b · d,

(1)

with a = 1487 ± 36 cm−1 and b = 462 ± 41 cm−1 nm−1 . This relation excellently describes our observations and is plotted as the black line in Fig. 4(a). Note that we use the average C-mode frequency given in Table 1 instead of all 14 measured C-mode values for fitting Eq. 1 such that each pairing carries equal weight. The linear decrease of the encapsulated chain’s C-mode frequency with decreasing nanotube diameter is strong evidence that the properties of the encapsulated chains are dominated by vdW interactions with the encasing inner carbon nanotube. This general trend has re9


Nano Letters

(b)

C-mode vs. diameter Fit Eq. 1 C-mode only

1830

1840 1830

(6,5)

1820

1820

1810

1810

1800

1800

(7,2) (6,4)

1790

1790

1780

1780

(8,0)

1770 0.60

0.65

C-mode Raman shift (cm-1)


(a) 1840

1770 0.70

0.75

inner CNT diameter (nm)

(d)

2.3 2.2 2.1 2.0

EG Ref. [18] Fit EG Ref. [18] EG this work from fit

2.3 2.2

LLCC band gap EG (eV)

(c) LLCC band gap EG (eV)

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 22

1.9 1.8 1.7 1780 1800 1820 1840 1860 C-mode Raman shift (cm -1)

2.1

EG vs d (Table I) Fit Eq. 2

2.0 1.9 1.8 1.7 0.60

0.65

0.70

0.75

0.80

inner CNT diameter (nm)

Figure 4: Correlation between diameter of encasing inner carbon nanotube vs C-mode frequency and bandgap of encapsulated long linear carbon chain. (a) C-mode frequency of LLCC as function of the encasing inner tube’s diameter. (b) C-mode frequencies of LLCCs for which no RBM from the inner tube could be detected. (c) Band gap EG as a function of C-mode from Ref. 18 (triangles) together with a linear fit that provides EG for the C-modes in this work (stars). (d) LLCC Band gap EG as function of inner CNT diameter d. cently been postulated qualitatively by Wanko et al. 22 We assign the small deviations from this linear trend, which amount to maximally 9 cm−1 for the (7, 2) inner nanotube, to charge transfer between the host nanotube and the encapsulated chain. By measuring chains encapsulated in different DWCNTs, our TERS measurements randomly probe different levels of charge transfer that vary with the four different inner CNT chiralities, labeled in Fig. 4(a), that are in resonance with the excitation. The overall effect of charge transfer, however, remains small compared to the effect of vdW interactions. In the following we use Eq. 1 to assign CNT diameters/chiralities to the tubes for which the RBM is not observed. This is possible because inverting Eq. 1 provides a robust estimate of the inner nanotube’s diameters based on the C-mode frequency alone. This estimate is relevant for the ∼ 40 TERS measurements without RBM signatures, as only a few of

10


Page 11 of 22 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

Nano Letters

the inner nanotube chiralities present in the sample are in resonance with our excitation wavelength, c.f. Fig. 3(a) and Table 1, and the resonance window of the RBM only spans around 50 meV. 24,25,31 Very rarely this is compensated by TERS enhancement. Figure 4(b) lists a representative set of measured C-mode frequencies for which no RBM appeared in the corresponding TERS spectrum. The blue arrows illustrate the assignment to diameters following Eq. 1. We expect the C-mode at 1808 cm−1 , for instance, to reside inside a tube with d = 0.694 nm, very close to the (7, 3) chirality with d = 0.696 nm, the resonance of which is at ∼ 2.5 eV and hence far from our excitation (1.96 eV). 29,35 Additional details on assigning C-modes to CNT diameters/chiralities are presented in the Supporting Information. We now compare our observations to recent wavelength-dependent Raman measurements on dense samples of LLCCs in DWCNTS. 18 In these bulk measurements, the Raman spectra were dominated by seven C-mode frequency bands at 1793 cm−1 to 1865 cm−1 . Three of these bands are within our measurement range (1793 cm−1 , 1802 cm−1 , and 1832 cm−1 ) and agree well with data, c.f. Fig 3(a) and (b), where we also find a strong clustering of the C-modes around 1800 cm−1 and 1830 cm−1 . It suggests that Eq. 1 connecting the diameter of the inner carbon nanotube to the C-mode frequency also holds for higher C-mode frequencies not observed in our TERS experiments. The highest observed C-mode at 1865 cm−1 corresponds to an inner tube diameter of d = 0.818 nm. This value is in very good agreement with recent transmission electron microscopy studies, which found d = 0.81 nm as the largest CNT diameter hosting long linear carbon chains. 15 Overall, our experimental data suggests that Table 1: Average RBM frequencies, chiralities, and diameters of the inner nanotubes together with the associated average C-mode frequencies and band gaps of the encapsulated chains. For the pairings measured multiple times, the errors are given by the standard deviation of the average values. For individual pairings, diameter d √ the experimental error is given. The 24 2 2 of a (n, m) nanotube is given by d = a/π n + nm + m with a = 0.246 nm. RBM (cm−1 ) (n,m) CNTin d (nm) C-m. (cm−1 ) Eg (eV) 18

327 ± 1 335 ± 3 348 ± 2 364 ± 3 (6,5) (6,4) 0.747 0.683 1835 ± 1 1831 ± 3 1798 ± 2 1798 ± 3 2.11 2.08 1.87 11


371 ± 1 (7,2) 0.641 1792 ± 1 1.83

391 ± 3 (8,0) 0.626 1772 ± 3 1.70

Nano Letters 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 22

LLCCs with C-mode frequencies from 1772 cm−1 to 1865 cm−1 grow inside nanotubes with diameters from 0.626 nm to 0.818 nm. Accordingly, equations 1 and 2 (see next paragraph) hold true for this diameter range. Combining the linear relationship between C-mode frequencies and band gaps EG of encapsulated linear carbon chains establish in Ref. 18 with Eq. 1 allow us to express EG as a function of the encasing inner nanotube’s diameter. We extract EG corresponding to a particular C-mode frequency from Ref. 18 and plot the data (triangles) in Fig. 4(c) together with a linear fit, see Supporting Information. This allows us to obtain EG corresponding the C-modes observed here, stars in Fig. 4(c) and Table 1, and to plot EG as a function of the inner nanotube’s diameter in Fig. 4(d). We find the band gap of encapsulated chains to increase linearly with the diameter of the inner nanotube as

EG (d) = c + f · d,

(2)

with c = −0.14 ± 0.25 eV and f = 2.98 ± 0.37 eV nm−1 . Our observation of fixed pairings of RBM and C-mode frequencies indicates that we only observe chains for which the length does not affect the C-mode. Commonly accepted theoretical models predict that the length dependence of the chain’s properties wear off for chains that consist of more than 100 atoms (length ≥ 13 nm). 8 As this is below our spatial resolution, we are unable to probe this limit. From the shortest chain measured by TERS in this work (length ∼ 30 nm, c.f. Fig. 3(b)), however, we obtain a lower experimental limit of ∼ 230 atoms for which no length dependence of the chain’s properties is evident. Long linear carbon chains inside double-walled carbon nanotubes as measured by us can therefore be regarded as the finite realization of carbyne. It follows that the carbyne C-mode frequency and band gap predominantly depend on the local environment given by the chirality of the inner CNT. We would like to stress that Eqs. 1 and 2 will no longer hold for very short carbon chains

12


Page 13 of 22 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

Nano Letters

since the C-mode frequency becomes dependent on the number of atoms for shorter chains. This does not exclude the existence of shorter chains in CNTs – we merely expect them to occur at higher excitation energies and C-mode frequencies, i.e., as reported in Ref. 36. With few inner tube chiralities hosting LLCCs, the number of different local environments that affect the properties of the encapsulated chains is limited and distinct in nature. This fact is reflected in all experimental studies on linear carbon chains encapsulated in carbon nanotubes known to the authors. 10–18 The observed C-modes are well separated into different frequency bands that vary around comparable mean frequencies. In very good agreement with our measurements no C-modes between 1810 cm−1 and 1825 cm−1 are reported. According to Eq. 1, only the C-mode of chains encapsulated in the metallic (9, 0) and (8, 2) nanotubes fall in this range. One may speculate that this apparent lack of LLCCs in metallic nanotubes is due to a less efficient growth process as compared to semiconducting tubes, which would prohibit us and others from observing the corresponding C-modes. Alternatively, the interaction between a chain and a metallic tube may differ from the interaction with a semiconducting nanotube, leading to a deviation from the linear trend in Eq. 1. Our results suggest that tunability of carbyne of up to ∼ 95 cm−1 in C-mode frequency, c.f. Fig. 4(a), and 0.6 eV in band gap, c.f. Fig. 4(c) and (d), can be achieved by choosing the right nanotube host, a property that can have interesting applications in low-dimensional optoelectronics. Long linear carbon chains encapsulated in nanotubes with a narrow diameter distribution, for instance, may serve as absorptive elements with a uniform bandgap in nanoscale photodetection despite the strong variations of the host nanotube’s optical resonances with chirality. This work shows (i) which inner nanotube diameters and chiralities should generally be aimed for and (ii) which C-mode frequency should be used to monitor the growth of carbyne with the desired properties. In conclusion, using tip-enhanced Raman spectroscopy we have investigated isolated pairs of double-walled carbon nanotubes and encapsulated long linear carbon chains. The chain’s Raman frequency is governed by the chirality of the inner tube and reduces with decreasing

13


Nano Letters 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

inner tube diameter. This points towards van-der-Waals forces as the dominating interaction mechanism between the host nanotube and the encapsulated chain. No length dependence of the chain’s Raman mode frequency is evident for the long linear carbon chains investigated, suggesting that they can be viewed as a close to perfect representation of carbyne. Our experimental data establishes a firm link between the local environment of the host nanotube and properties of the encapsulated linear carbon chain – thereby explaining the variation in the chain’s Raman mode frequency – and identifies parameters for tuning the phononic and electronic properties of carbyne for device applications.

Acknowledgement The authors thank M. Frimmer, M. Parzefall and S. Reich for fruitful discussions. This research was supported by the Swiss National Science Foundation (grant no. 200021_165841). S. Heeg acknowledges financial support by ETH Zürich Career Seed Grant SEED-16 17-1.

Supporting Information Available Topography recorded during TERS imaging; effect of encapsulated chains on the RBM of DWCNT; assigning chiralities to the RBMs of inner carbon nanotubes; on the effect of the outer NCT on the encapsulated LLCC; assigning chain band gaps to pairings of RBM and C-mode; assigning CNT chiralities to C-mode frequencies without RBMs. This material is available free of charge via the Internet at http://pubs.acs.org/.

References (1) Heimann, R. B.; Evsyukov, S. E.; Kavan, L. Carbyne and Carbynoid Structures; Springer Science & Business Media, 2012.

14


Page 14 of 22

Page 15 of 22 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

Nano Letters

(2) Casari, C. S.; Tommasini, M.; Tykwinski, R. R.; Milani, A. Nanoscale 2016, 8, 4414– 4435. (3) Casari, C. S.; Milani, A. MRS Communications 2018, 1–13. (4) Liu, M.; Artyukhov, V. I.; Lee, H.; Xu, F.; Yakobson, B. I. ACS Nano 2013, 7, 10075– 10082. (5) Peierls, E. R. Quantum Theory of Solids; Clarendon Press, 1955. (6) K¯ urti, J.; Magyar, C.; Balázs, A.; Rajczy, P. Synthetic Metals 1995, 71, 1865–1866. (7) Milani, A.; Tommasini, M.; Fazzi, D.; Castiglioni, C.; Zoppo, M. D.; Zerbi, G. J. Raman Spectrosc. 2008, 39, 164–168. (8) Yang, S.; Kertesz, M. J. Phys. Chem. A 2006, 110, 9771–9774. (9) Chalifoux, W. A.; Tykwinski, R. R. Nature Chemistry 2010, 2, 967–971. (10) Zhao, X.; Ando, Y.; Liu, Y.; Jinno, M.; Suzuki, T. Phys. Rev. Lett. 2003, 90, 187401–4. (11) Fantini, C.; Cruz, E.; Jorio, A.; Terrones, M.; Terrones, H.; Van Lier, G.; Charlier, J.; Dresselhaus, M. S.; Saito, R.; Kim, Y. A.; Hayashi, T.; Muramatsu, H.; Endo, M.; Pimenta, M. A. Phys. Rev. B 2006, 73, 193408–4. (12) Jinno, M.; Ando, Y.; Bandow, S.; Fan, J.; Yudasaka, M.; Iijima, S. Chemical Physics Letters 2006, 418, 109–114. (13) Nishide, D.; Wakabayashi, T.; Sugai, T.; Kitaura, R.; Kataura, H.; Achiba, Y.; Shinohara, H. J. Phys. Chem. C 2007, 111, 5178–5183. (14) Shi, L.; Sheng, L.; Yu, L.; An, K.; Ando, Y.; Zhao, X. Nano Res. 2011, 4, 759–766. (15) Andrade, N. F.; Vasconcelos, T. L.; Gouvea, C. P.; Archanjo, B. S.; Achete, C. A.; Kim, Y. A.; Endo, M.; Fantini, C.; Dresselhaus, M. S.; Filho, A. G. S. Carbon 2015, 90, 172–180. 15


Nano Letters 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

(16) Andrade, N. F.; Aguiar, A. L.; Kim, Y. A.; Endo, M.; Freire, P. T. C.; Brunetto, G.; Galvão, D. S.; Dresselhaus, M. S.; Souza Filho, A. G. J. Phys. Chem. C 2015, 119, 10669–10676. (17) Shi, L.; Rohringer, P.; Suenaga, K.; Niimi, Y.; Kotakoski, J.; Meyer, J. C.; Peterlik, H.; Wanko, M.; Cahangirov, S.; Rubio, A.; Lapin, Z. J.; Novotny, L.; Ayala, P.; Pichler, T. Nature Materials 2016, 15, 634–639. (18) Shi, L.; Rohringer, P.; Wanko, M.; Rubio, A.; Wasserroth, S.; Reich, S.; Cambre, S.; Wenseleers, W.; Ayala, P.; Pichler, T. Phys. Rev. Materials 2017, 1, 075601–7. (19) Lapin, Z. J.; Beams, R.; Cançado, L. G.; Novotny, L. Faraday Discussions 2015, 184, 193–206. (20) Rohringer, P.; Shi, L.; Ayala, P.; Pichler, T. Adv. Funct. Mater. 2016, 26, 4874–4881. (21) Neves, W. Q.; Alencar, R. S.; Ferreira, R. S.; Torres-Dias, A. C.; Andrade, N. F.; San-Miguel, A.; Kim, Y. A.; Endo, M.; Kim, D. W.; Muramatsu, H.; Aguiar, A. L.; Filho, A. G. S. Carbon 2018, 133, 446–456. (22) Wanko, M.; Cahangirov, S.; Shi, L.; Rohringer, P.; Lapin, Z. J.; Novotny, L.; Ayala, P.; Pichler, T.; Rubio, A. Phys. Rev. B 2016, 94, 195422. (23) Bonabi, F.; Brun, S. J.; Pedersen, T. G. Phys. Rev. B 2017, 96, 155419–8. (24) Reich, S.; Thomsen, C.; Maultzsch, J. Carbon Nanotubes: An Introduction to the Basic Concepts and Physical Properties; Wiley-VCH New York, 2004. (25) Maultzsch, J.; Telg, H.; Reich, S.; Thomsen, C. Phys. Rev. B 2005, 72, 205438. (26) Hartschuh, A.; Sánchez, E.; Xie, X.; Novotny, L. Phys. Rev. Lett. 2003, 90, 095503. (27) Hartschuh, A. Angew. Chem. Int. Ed. 2008, 47, 8178–8191.

16


Page 16 of 22

Page 17 of 22 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

Nano Letters

(28) Johnson, T. W.; Lapin, Z. J.; Beams, R.; Lindquist, N. C.; Rodrigo, S. G.; Novotny, L.; Oh, S.-H. ACS Nano 2012, 6, 9168–9174. (29) Pfeiffer, R.; Simon, F.; Kuzmany, H.; Popov, V. N. Phys. Rev. B 2005, 72, 161404–4. (30) Pfeiffer, R.; Pichler, T.; Kim, Y. A.; Kuzmany, H. Carbon Nanotubes; Springer, Berlin, Heidelberg: Berlin, Heidelberg, 2007; pp 495–530. (31) Thomsen, C.; Reich, S. In Raman scattering in carbon nanotubes; M, C., Merlin, R., Eds.; Topics in Applied Physics; Springer, Berlin, Heidelberg: Heiderlberger Platz 3, D-14197 Berlin, Germany, 2007; Vol. 108. (32) Beams, R.; Cançado, L. G.; Oh, S.-H.; Jorio, A.; Novotny, L. Phys. Rev. Lett. 2014, 113, 186101–8. (33) Popov, V.; Henrard, L.; Lambin, P. Phys. Rev. B 2005, 72 . (34) Dobardžić, E.; Maultzsch, J.; Milošević, I.; Thomsen, C.; Damnjanović, M. phys. stat. sol. (b) 2003, 237, R7–R10. (35) Pfeiffer, R.; Kuzmany, H.; Kramberger, C.; Schaman, C.; Pichler, T.; Kataura, H.; Achiba, Y.; Kürti, J.; Zólyomi, V. Phys. Rev. Lett. 2003, 90, 162–4. (36) Zhao, C.; Kitaura, R.; Hara, H.; Irle, S.; Shinohara, H. J. Phys. Chem. C 2011, 115, 13166–13170.

17


photons (kcts/ms)

1 2 3 4 5 6 7 8 4.8 (c) 9 10 11 12 13 14 15 16 0

Nano Letters Page C-mode 18 of 22 tip down - TERS b tip up - CF

x10

400

120 kHz

50 0

G

RBM

23 nm

800 1200 Raman shift (cm -1)

(d)

1600

1

I(z) ~ e-z/τ

(ITERS-ICF)/(ITERS(z=0)-ICF)

(b)

Normalized Raman intensity (a.u.)

(a)

τ = 2.5 nm

ACS Paragon Environment 100Plus nm 0

0

10

20 z(nm)

30

(b) Nano Letters

Normalized RBM Raman intensity

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Normalized C-mode Raman intensity

(a) Page 19 of 22

ACS Paragon Plus Environment 250

300

350

400

Raman shift (cm -1)

450

1700 1750 1800 1850 1900 Raman shift (cm -1)

(b ) Nano Letters

(a ) 1 8 4 0

Page 20 of 22 1 8 4 0

C -m o d e + R B M R e f. [1 7 ] 0

1 8 3 0

0

1 8 2 0

(6 ,5 ) 0

1 8 1 0 0

1 8 0 0

(7 ,2 ) 0

1 7 9 0 0

(6 ,4 )

1 7 8 0 (8 ,0 )

0

1 7 7 0 3 2 0

3 3 0

3 4 0

3 5 0

3 6 0

3 7 0

R B M

R a m a n s h ift ( c m

3 8 0 -1

)

3 9 0

0

1 0 0

2 0 0

L L C C le n g th ( n m )




-1

-1

)

)

11 8 3 21 8 2 31 8 1 41 8 0 5 61 7 9 71 7 8 81 7 7 9 10 11 12 13

# c a rb o n a to m s 1 0 0 0 2 0 0 0 0

1830 1820

Fit Eq. 1 C-mode only

1840 1830

(6,5)

1820



(a)Page 1840 21 ofC-mode 22 vs. diameter Nano Letters (b)



11810 1810 2 1800 1800 (7,2) 3 (6,4) 1790 1790 4 1780 51780 (8,0) 1770 61770 7 0.60 0.65 0.70 0.75 inner CNT diameter (nm) 8 9 (d) 2.3 (c)102.3 EG Ref. [18] EG vs d (Table I) 2.2 2.2 Fit EG Ref. [18] 11 Fit Eq. 2 EG this work 2.1 122.1 from fit 2.0 132.0 1.9 141.9 1.8 151.8 ACS Paragon Plus Environment 1.7 161.7 17 1780 1800 1820 1840 1860 0.60 0.65 0.70 0.75 0.80 inner CNT diameter (nm) C-mode Raman shift (cm -1) 18 19 20 21

Nano Letters 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 22

Carbon Nanotube Chirality Determines Properties ... - ACS Publications

Recommend Documents