Production of Highly Monolayer Enriched Dispersions of Liquid-Exfoliated Nanosheets by Liquid Cascade Centrifugation

Subscriber access provided by ORTA DOGU TEKNIK UNIVERSITESI KUTUPHANESI
Article
Production of Highly Monolayer Enriched Dispersions of Liquid-
Exfoliated Nanosheets by Liquid Cascade Centrifugation
Claudia Backes, Beata M Szydlowska, Andrew Harvey, Shengjun Yuan, Victor Vega-
Mayoral, Ben R. Davies, Pei-liang Zhao, Damien Hanlon, Elton Santos, Mikhail I
Katsnelson, Werner Josef Blau, Christoph Gadermaier, and Jonathan N. Coleman
ACS Nano, Just Accepted Manuscript • DOI: 10.1021/acsnano.5b07228 • Publication Date (Web): 05 Jan 2016
Downloaded from http://pubs.acs.org on January 9, 2016
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
ACS Nano 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 31 ACS Nano
1
2
3 Production of Highly Monolayer Enriched Dispersions of Liquid-Exfoliated Nanosheets
4
5 by Liquid Cascade Centrifugation
6
7 Claudia Backes,1,2 Beata M. Szydłowska,1,2 Andrew Harvey,1,2 Shengjun Yuan,3 Victor
8
9 Vega-Mayoral,4,5 Ben R. Davies,1,2 Pei-liang Zhao,6 Damien Hanlon,1,2 Elton Santos,7
10 Mikhail I. Katsnelson,3 Werner Josef Blau,1,2 Christoph Gadermaier,4,5 and Jonathan N.
11
12 Coleman1,2*
13
14 1
CRANN & AMBER, Trinity College Dublin, Dublin 2, Ireland
15
16 2
School of Physics, Trinity College Dublin, Dublin 2, Ireland
17
18 3
19 Institute for Molecules and Materials, Radboud University of Nijmegen, Heijendaalseweg
20 135, 6525AJ Nijmegen, the Netherlands
21
22 4
23 Department for Complex Matter, Jozef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia
24 5
25 Jozef Stefan International Postgraduate School, Jamova 39, 1000 Ljubljana, Slovenia
26
6
27 Department of Applied Physics, Zernike Institute for Advanced Materials, University of
28
29 Groningen, Nijenborgh 4, NL-9747AG Groningen, The Netherlands
30
7
31 School of Mathematics and Physics and School of Chemistry and Chemical Engineering,
32
33 Queen's University Belfast, Belfast, BT95AL, United Kingdom
34
35 *colemaj@tcd.ie
36
37 ABSTRACT: While liquid exfoliation is a powerful technique to produce defect-free
38
39 nanosheets in large quantities, its usefulness is limited by broad nanosheet thickness
40
distributions and low monolayer contents. Here we demonstrate liquid processing techniques,
41
42 based on iterative centrifugation cascades, which can be designed to achieve either highly
43
44 efficient nanosheet size-selection and/or monolayer enrichment. The resultant size-selected
45
46
dispersions were used to establish quantitative metrics to determine monolayer volume
47 fraction, as well as mean nanosheet size and thickness, from standard spectroscopic
48
49 measurements. Such metrics allowed us to design and optimize centrifugation cascades to
50
51
enrich liquid exfoliated WS2 dispersions up to monolayer contents of 75%. Monolayer-rich
52 dispersions show relatively bright photoluminescence with narrow linewidths (<35 meV)
53
54 indicating the high quality of the nanosheets. The enriched dispersions display extinction
55
56
spectra with distinct features, which also allow the direct estimation of monolayer contents.
57
58 Keywords: exfoliation, monolayer, size-selection, luminescence
59
60 1
ACS Paragon Plus Environment
ACS Nano Page 2 of 31
1
2
3 Liquid phase exfoliation is a versatile technique for producing liquid suspensions
4
5 containing large quantities of 2-dimensional (2D) nanosheets.1-7 This method involves the
6 sonication2, 4
or shearing8, 9
of layered materials in liquids, resulting in the production of
7
8 generally defect-free nanosheets. If the liquid is appropriately chosen (e.g. suitable solvents5,
9 6, 10, 11
10 or surfactant3, 12-15 or polymeric solutions16-19) the nanosheets will be stabilized against
11 reaggregation.20 This method has been used to exfoliate a broad range of 2D materials
12
13 including graphene,2, 6, 9, 12, 17, 21
BN,5 transition metal dichalcogenides such as MoS2 and
14
15 WSe2,4, 8, 22, 23 as well as MoO3,24, 25 GaS,26 black phosphorous27-31 and MXenes.32 The liquid
16 suspensions produced in this way are extremely useful, as they can be easily processed into a
17
18 range of structures including films, coatings and composites.5, 23, 33-35
Such materials have
19
20 demonstrated superlative performance in a number of applications including reinforced
21 composites,5, 18, 36, 37 battery electrodes38, 39 and fiber lasers.40-42
22
23
24 However, liquid phase exfoliation tends to give nanosheets with very broad lateral size
25 (length, L) and thickness (expressed as number of monolayers per nanosheet, N)
26
27 distributions,15, 43 with ranges of 50nm≤L≤500nm and 1≤N≤10 not unusual for MoS2.15 Such
28
29 distributions are problematic as many applications require controlled nanosheet sizes: small
30
nanosheets are ideal for catalysis44 while large ones are needed for mechanical
31
32 reinforcement.37 These broad distributions also mean the monolayer content is always low,
33
34 generally <10%. This is a serious problem for a number of applications. For example, the use
35
of 2D materials in printed optoelectronic devices cannot be considered unless inks containing
36
37 very high monolayer contents are available.45 In the longer term, nanosheet printed
38
39 electronics will require monodisperse suspensions containing only N-layers where N is
40
defined by the application.
41
42
43 Such capabilities are a long way away. While some progress has been made toward
44
selection of nanosheets by size,15, 43 the available processes are generally inefficient, yielding
45
46 very small quantities of nanosheets. In addition, monolayer enrichment is much more
47
48 challenging, not least because the measurement of monolayer content generally involves
49
statistical microscopy which is very time consuming. While density gradient
50
51 ultracentrifugation has been used to separate TMD nanosheets by thickness,46 this procedure
52
53 is complex, gives low yields (concentrations <0.005 g/L) and is limited to polymer-stabilized
54
dispersions. What is needed is a scalable, universally applicable, high-yield technique to
55
56 either size-select nanosheets or produce highly monolayer-enriched dispersions. Any such
57
58 process will involve optimization, which is limited by the tedious microscopic
59
60 2
1
2
3 characterization required. Hence, to enable this, it will be necessary to develop fast
4
5 techniques to measure nanosheet size, thickness and monolayer content. Within this
6 manuscript we address these points.
7
8
9 Inspired by gas-separation centrifugation cascades, we have developed an efficient
10 centrifugation-based method, which allows nanosheet dispersions to be both separated by size
11
12 and enriched in monolayers in a controlled way. By studying the optical properties of WS2
13
14 dispersions as a function of nanosheet size, thickness and monolayer content, we identified
15 spectral properties which scale with monolayer content as well as mean nanosheet length and
16
17 thickness yielding quantitative metrics for these properties. The resultant ability to measure
18
19 L , N and monolayer content, allows us to design secondary cascades to further enhance
20
21 the monolayer content, reaching values as high as 75%. Such monolayer-rich dispersions
22
23
display easily observable photoluminescence (PL) and optical properties that distinct from
24 normally observed for WS2 ensembles.
25
26
27 Results
28
29 Size selection of WS2 nanosheets
30
31 As liquid-exfoliated nanosheets tend to be polydisperse in both size and thickness,
32
33 size selection is almost always required. A number of size-selection techniques exist, mostly
34
based on centrifugation.3, 15, 43, 46, 47 However, these techniques tend to be inefficient, yielding
35
36 small masses of size-selected nanosheets. To address this, we have developed a new
37
38 centrifugation-based technique which we denote liquid cascade centrifugation (LCC). We
39
start with a dispersion of liquid-exfoliated nanosheets, obtained by sonication of WS2 powder
40
41 in aqueous surfactant solution (see methods) where any unexfoliated crystallites had been
42
43 removed by low speed centrifugation (here 1.5 krpm in our centrifuge). This “stock”
44
dispersion contains nanosheets with a broad distribution of sizes and thickness and a small
45
46 but non-trivial population of monolayers with varying lateral sizes. The stock is then
47
48 centrifuged at a higher speed (here 2 krpm) and the sediment collected. This sediment
49
contains nanosheets from the larger end of the size distribution and, as they were collected
50
51 between centrifugation rates of 1.5 and 2 krpm, we refer to this sample as “1.5-2 krpm”.
52
53 Critical to LCC, this sediment can be redispersed48 completely by mild sonication in H2O-SC
54
(at SC concentrations as low as 0.1 g/L) to give dispersions with virtually any chosen
55
56 concentration. The supernatant produced during the 2 krpm centrifugation contains all but the
57
58 largest nanosheets. It can be centrifuged at a higher rate (here 3 krpm) to give a sediment with
59
60 3
1
2
3 slightly smaller nanosheets, which we label “2-3 krpm”. The associated supernatant can be
4
5 centrifuged again and the cascade continued for as many steps as are required with each step
6 using a continually increasing centrifugation rate. Critically, because the heavier, few-layer
7
8 nanosheets are removed in each step of the cascade, the resultant supernatants become more
9
10 and more monolayer enriched. After each step, the sediment contains smaller and smaller
11 nanosheets, resulting in effective size selection. Within this cascade, we terminate the
12
13 cascade at 10 krpm and only partially analyze the final supernatant as it mostly contained free
14
15 surfactant and extremely small WS2 nanosheets. This process is illustrated in Figure 1. One
16 very important feature of LCC is that virtually no material is wasted, resulting in the
17
18 collection of relatively large masses of nanosheets in each fraction (see below).
19
20 Importantly, the cascade can be designed according to the desired outcome. Here, we
21
22 wanted to produce a range of dispersions with varying nanosheet sizes and therefore we
23
24 performed a set of centrifugations with subsequently increasing rpm. However, if only a
25 specific size distribution is desired, the procedure can be simplified by trapping the desired
26
27 nanosheets between two fixed rpm. For example, if the stock is directly centrifuged at 7.5
28
29 krpm and the resultant supernatant then centrifuged at 10 krpm, the collected sediment will
30 contain nanosheets with a size distribution virtually identical to those obtained from the “7.5-
31
32 10 krpm” step in a cascade. Thus a cascade can be used when a set of different nanosheet
33
34 sizes is required but only two centrifugations are required to obtain a single size-selected
35 sample.
36
37
38
We have characterized the nanosheets collected in each fraction microscopically using
39 both transmission electron microscopy (TEM) and atomic force microscopy (AFM) with
40
41 typical images displayed in Figures 2 A-D (see supplementary Figures S1-S4). In order to
42
43
characterize the size-selection of LCC, the nanosheet length (i.e. the longest dimension) was
44 measured over all fractions. Example histograms are shown in Figures 2E and 2F (see
45
46 supplementary Figures S1, 4) for the 1.5-2 krpm and 7.5-10 krpm samples. These histograms
47
48
show a reduction in nanosheet length as the centrifugation rates are increased (i.e. as the
49 dispersion progresses through the cascade).
50
51
Along with nanosheet length and width (defined as the dimension perpendicular to
52
53 length), the nanosheet thickness, expressed as number of monolayers per nanosheet, N, was
54
55 measured by AFM with the effect of residual surfactant corrected for using step height
56
analysis (SI figure S5-6).9, 15, 26, 31 Typical layer number histograms are shown in Figures 2G-
57
58 H (see supplementary Figure S4). Again, we see a reduction in nanosheet thickness as the
59
60 4
1
2
3 dispersion progresses through the cascade. Interestingly, the N histograms are always log-
4
5 normal in shape (see SI section 1 1.1-1 1.2) as previously observed for LPE nanosheets.49
6
7 We can quantify these effects by plotting the mean nanosheet length (measured from
8
9 TEM and AFM) as a function of the centrifugal acceleration (g-force, see methods)
10 associated with the midpoint of the pair of rpms used in each step of the cascade (see Figure
11
12 2I). The mean nanosheet length falls off as (g-force)-0.5 as expected due to the close
13
14 relationship between <L> and the “cut size” (i.e. the size of the largest particle remaining
15 dispersed after centrifugation, see supplementary Figure S7).21 The lateral dimensions
16
17 measured by AFM are over-estimated due to tip broadening and pixilation effects. The data
18
19 in Figure 2I has been used to correct the lateral dimensions measured by AFM for the rest of
20 the analysis (see supplementary Figure S8). We note that we have also analyzed the very
21
22 small nanosheets that are discarded in the final supernatant of the cascade. TEM (figure S2)
23
24 confirms that these are indeed 2D nanosheets with <L> ~ 25 nm. Due to the small lateral
25 dimensions, a reliable AFM analysis was not possible.
26
27
28
Similarly, the mean nanosheet thickness, <N>, as measured by AFM is plotted versus
29 g-force in Figure 2J. The mean number of layers falls with central rotational speed via (g-
30
31 force)-0.4. For the set of size-selected samples studied here, it is clear from this data that
32
33 L ∝ N . This is certainly a limitation of this method as, ideally, one would like to vary
34
35 L and N independently. Below and in the SI (figure S9) we discuss the relationship
36
37 between L and N in more detail. We find the distribution widths (i.e. the standard deviation:
38
39 ∆L or ∆N) to scale linearly with the mean for both L and N (see Figures 2K-L) with slopes
40
41 that are quite high (~50%). This is also not ideal as much smaller values of ∆N / N and
42
43 ∆L / L will be required for most applications. In this work, we have not optimised the
44
45 procedure to give reduced distribution widths. However, we believe that LCC is versatile
46
enough for such optimisation to be achievable.
47
48
49 AFM analysis can also be used to assess the population of monolayers expressed as
50 monolayer number fraction, Nmono/NT. and plotted in Figure 3A as a function of the central g-
51
52 force. Interestingly, we find Nmono/NT to scale linearly with g-force as the monolayer content
53
54 is enriched. After 7 steps in the cascade (i.e. at central acceleration of ~7000g), values of
55 Nmono/NT as high as 40 % have been achieved. In Figure 3B we plot Nmono/NT vs. <N>, finding
56
57
58
59
60 5
1
2
3 a very sharp fall off, emphasizing the importance in minimizing the mean nanosheet
4
5 thickness for ML enrichment.
6
7 Another way to describe the monolayer content is via the monolayer volume fraction,
8
9 Vf. This has been calculated from the AFM data using the knowledge of nanosheet length, L,
10 width, W, and thickness, t (which we define as N×0.6 nm) according to equation 1.
11
12
13 ∑ LWt ∑ LW
14 Vf =mono
= mono
(1)
15 ∑ LWt ∑ LWN
All All
16
17
18 We find Vf to scale as ( g − force)1.9 , reaching ~20% after 7 steps in the cascade (see Figure
19
20 3C). Because few-layer nanosheets are typically of higher volume than monolayers, Vf
21
22
<Nmono/NT.
23
24 Knowledge of Vf for each fraction is extremely useful as it allows us to assess the
25
mass of ML nanosheets produced in each step once the total WS2 mass is known. For each
26
27 step in the cascade, we measured the dispersed WS2 mass using a combination of weighing
28
29 and spectroscopy (see supplementary Figure S10). This data is plotted versus central g-force
30
in Figure 3D. The stock dispersion (80 mL in volume, ~150 mg of exfoliated WS2) was
31
32 separated into 7 fractions with the first four fractions containing 25-40 mg of WS2 nanosheets
33
34 each, considerably more than achieved using comparable processes. For example, a multistep
35
centrifugation procedure has reported graphene nanosheet quantities <1 mg43 while density
36
37 gradient ultracentrifugation has yielded WS2 monolayer-rich dispersions with concentrations
38
39 of <0.005 g/L.46
40
41 However, after the fourth step in the cascade, (i.e. for central g-forces above ~2000g)
42
43 the mass produced falls off sharply, reaching ~1 mg after step 7 (~7000g). The mass of
44
45
monolayers produced is related to the total WS2 mass by M ML = V f M T allowing us to
46
47
calculate the mass of monolayers as also plotted versus central g-force in Figure 3D. We find
48 the monolayer mass to increase over the first 4 steps of the cascade as the dispersions become
49
50 more and more enriched. However, even though enrichment continues in subsequent steps, it
51
52
falls off after step 4 as the total dispersed mass declines. This data clearly shows that highly
53 enriched samples can be achieved by collecting nanosheets using high centrifugation rates but
54
55 at the cost of low yield. Alternatively, the mass of monolayers collected can be optimized by
56
57
using midrange centrifugation rates.
58
59
60 6
1
2
3 Relationship between nanosheet length and thickness
4
5 The data in figure 1 I-J shows the mean nanosheet length and thickness of the
6
7 different samples to be in proportion to each other. However, this hides important trends
8
9 within each sample. Shown in supplementary figure S9A is a plot of L versus N measured for
10 all individual nanosheets over the seven samples prepared in the cascade described above.
11
12 While this graph appears to be consistent with the proportionality of L and N, closer
13
14 inspection shows that, for a given sample (e.g. 4-6 krpm as shown in inset), the nanosheet
15 length does not vary systematically with thickness. To investigate this more carefully, we
16
17 extracted the mean nanosheet length for each value of N, which we plotted versus N or each
18
19 sample in figure S9B. This shows that for a given step in the cascade, the mean nanosheet
20 length is roughly independent of N: e.g. for the 7.5-10 krpm sample, the mean monolayer
21
22 length is 41 nm while the mean 4-layer length is 45 nm. This implies that the separation
23
24 mechanism predominately involves nanosheet length rather than mass as might be expected.
25 However, lower central g-forces result in larger nanosheets in general and so larger
26
27 monolayers. For the relatively low central g-forces (2-3 krpm sample) the mean monomer
28
29 length was as high as 130 nm with individual monolayers as long as 200 nm observed.
30 Unfortunately, because these long monolayers are found in the low g-force samples, they
31
32 tend to be relatively rare. To date we have not found a way to highly enrich a sample with
33
34 long monolayers rather than the shorter (~50 nm) monolayers found in the 7.5-10 krpm
35 sample. We note that monolayer length distribution in the stock is probably set by the
36
37 exfoliation conditions (e.g. details of sonication regime and stabilizer). In the future, it would
38
39 be useful to identify conditions which would give a much larger initial population of large
40 monolayers.
41
42
43
Dependence of the optical extinction spectra on the nanosheet dimensions
44
45 It has previously been shown in the case of liquid-exfoliated MoS2, that the spectral
46
profile of optical extinction (or absorbance) spectra strongly depends on nanosheets
47
48 dimensions due to edge and confinement effects.15 Here we use the fractions produced by
49
50 LCC to investigate the effect of nanosheet size and thickness on the optical properties of LPE
51
WS2.
52
53
54 To do this, we first measured the optical extinction spectra for the fractions described
55
56 above (N.B. the extinction, Ext, is related to the transmittance, T, by T = 10 − Ext and to the
57 extinction coefficient, ε, by Ext = ε Cl where C is the nanosheet concentration and l is the
58
59
60 7
1
2
3 cell length). The extinction coefficients were calculated using the measured mass of dispersed
4
5 nanosheets (see supplementary Figure S10). As shown in Figure 4A, optical extinction
6 spectra of liquid-exfoliated WS2, display the characteristic excitonic transitions,27 but vary
7
8 systematically with nanosheet size and thickness due to edge and confinement effects.15
9
10 Similar behavior is observed for the absorbance and scattering coefficient spectra (see
11 supplementary Figure S11).15 Importantly, the scattering coefficients are relatively small for
12
13 these nanosheets, meaning that the absorbance and extinction are very similar. The edge
14
15 effects result in a dependence of the spectral profile on nanosheet length. As a result the
16 extinction coefficient, for example at the A-exciton (~620 nm), depends strongly on
17
18 nanosheet length as shown in Figure 4B. However, the extinction coefficient at 235 nm is
19
20 invariant with nanosheet length (ε235nm=47.7 Lg-1cm-1), allowing its universal use to measure
21
the concentration of WS2.
22
23
24 The effect of edges on the spectral shape can be quantified via the ratio of extinction
25 intensities at two different wavelengths e.g. at 235 nm to that at 290 nm, Ext235/Ext290, which
26
27 is plotted in Figure 4C. Similar to MoS2, the data in Figure 4C can be fitted to the following
28
29 equation15
30
31 Ext235 α C (235nm) L + 2 x(k + 1)∆α (235nm)
32 = (2)
Ext290 α C (290nm) L + 2 x(k + 1)∆α (290nm)
33
34
35 where α C is the absorption coefficient associated with the nanosheet basal plane,
36
37 ∆α = α E − α C where α E is the edge region absorption coefficient, and L, x and k are the
38
39 nanosheet length, edge thickness and aspect ratio respectively.15 We find this equation fits the
40
41 data very well allowing us to generate a function which relates the mean nanosheet length, L
42 to the extinction peak intensity ratio:
43
44
2.3 − Ext235 / Ext290
45 L(nm) = (3)
46 0.02 Ext235 / Ext290 − 0.0185
47
48 This relationship is extremely useful as it allows the mean nanosheet length in any dispersion
49
50 to be extracted from an extinction spectrum. Other peak intensity ratios yield similar
51
52 relationships as discussed in the SI (see supplementary Figure S12).
53
54 In addition, the extinction spectra also contain information on mean nanosheet
55
56 thickness due to confinement effects. These result in shifts of the A-exciton position (see
57 Figure 4A inset) towards lower wavelengths as the nanosheet thickness is reduced.
58
59
60 8
1
2
3 Interestingly, for the 7.5-10 krpm sample, previously unseen structure begins to appear with a
4
5 peak at ~612 nm and a shoulder at ~622 nm. The origin of this will be discussed in more
6 detail below.
7
8
9 The relationship between the center of mass position of the A-exciton, λA, (determined
10 from the second derivative of the A-exciton, supplementary Figure S13) and the mean WS2
11
12 nanosheet thickness is displayed in Figure 4D. As with MoS2, λA increases logarithmically
13
14 with nanosheet thickness.15 We find data for WS2 exfoliated in sodium dodecylbenzene
15 sulfonate (SDBS) and poly(vinyl alcohol) (PVA) to sit very close to the same line suggesting
16
17 solvatochromic effects to be small (but not nonexistent, see SI).50
18
19 In any case, fitting the data in Figure 4D to an empirical relation gives an equation,
20
21 which allows us to extract the mean nanosheet thickness from the wavelength associated with
22
23
the A-exciton:
24
25 N = 6.35 ×10−32 eλA ( nm )/8.51 (4)
26
27
28
This equation is at least applicable to aqueous SC-, SDBS- and PVA-stabilized WS2
29 dispersions and almost certainly gives approximate nanosheet thicknesses in a wide range of
30
31 liquid environments (see below and SI). Because these shifts in A-exciton position will
32
33
become important later in the paper, we need to verify this behavior as much as possible. We
34 found similar trends had been observed by Zhu et al. and Zhao et al. for
35
36 absorbance/reflectance spectra of micromechanically cleaved WS2.51, 52
We note that the
37
38
literature data from micromechanically cleaved WS2 is offset to lower energies compared to
39 our data presumably to the different dielectric environment.50, 53
In addition, we have
40
41 attempted to confirm these effects by calculating the electronic band structure of WS2
42
43
nanosheets as a function of N. As shown in Figure 4E, the computed direct transition at the
44 K-point (see section: methods) of the Brillouin zone follows a similar logarithmic
45
46 dependence on layer number as observed in the experimental data.
47
48 It is worth noting that equations 3 and 4 might only be expected to hold for surfactant
49
50 exfoliated WS2 nanosheets. To test this we performed a cascade for WS2 nanosheets
51
exfoliated in the solvent N-methyl-pyrrolidone (NMP).4 We found the length metric
52
53 described by equation 3 to apply very well to NMP-exfoliated WS2. However, we observed a
54
55 small degree of solvatochromism which resulted in shifts in λA due to environmental effects
56
57 (~1.5 nm shift between NMP- and SC-stabilized nanosheets). This means that although
58
59
60 9
1
2
3 equation 4 can be used to approximately find nanosheet thicknesses from extinction spectra
4
5 of NMP-based dispersions, slight modifications are required to accurately extract nanosheet
6 thicknesses in NMP. This modified metric is given in Figure S23.
7
8
9 Dependence of the photoluminescence on the monolayer content in the dispersions
10
11 While quantitative spectroscopic metrics for length and thickness are instructive, a
12
13
metric for ML content would be even more useful. The most obvious candidate is the
14 nanosheet photoluminescence (PL) as this is only appreciable in monolayers for the common
15
16 group VI-TMDs.51, 54-56 However, previous experiments on MoS2 dispersions showed that PL
17
18
can only be detected using standard luminescence spectrometers for dispersions with high
19 monolayer content,15 making routine PL characterization impossible.
20
21
Here we use the superior sensitivity of a Raman spectrometer to detect PL even in non-
22
23 centrifuged stock dispersions with low ML content. We found that, when acquiring a Raman
24
25 spectrum in liquid drops (λexc = 532 nm), the photoluminescence of the monolayer is typically
26
detected at high wavenumbers (~2460 cm-1). N.B. the measurement has to be carried out with
27
28 extreme care (see methods). The Raman/PL spectra normalized to the WS2 2LA(M) Raman
29
30 mode (355 cm-1, 543 nm) of size-selected WS2 nanosheet dispersions are plotted in Figure 5A
31
as a function of wavelength (also compare figure S14). The WS2 Raman modes plotted
32
33 versus wavenumber are shown in the inset. In addition to the WS2 Raman modes, in all cases,
34
35 the typical photoluminescence peak of the WS2 is detected at ~612 nm (the feature at ~650
36
nm is the Raman response of water).
37
38
39 A typical PL spectrum measured in this way is plotted versus photon energy in Figure
40 5B. All spectra fit reasonably well to single Lorentzians, representing excitonic emission.
41
42 While we initially expected inhomogeneous broadening to render Gaussian fitting more
43
44 appropriate, we consistently found more reliable fits using Lorentzians (see supplementary
45 Figure S15). The Lorentzian is centered at 2.023 eV invariant of centrifugation conditions
46
47 (see Figure 5C inset). However, as shown in Figure 5C, the linewidth appears to fall off with
48
49 increasing centrifugation rate suggesting larger nanosheets to be slightly more defective than
50 smaller ones. It must be noted that these PL peaks are very narrow, displaying widths as low
51
52 as 30 meV. WS2 usually displays PL line widths between 22 and 75 meV,51, 57 with the lower
53
54 values found only on non-perturbing surfaces like BN. This suggests that, not only does the
55 water/SC environment not significantly dope the nanosheets, but also that they are largely
56
57 defect-free (see supplementary Figure S16).
58
59
60 10
1
2
3 As shown in Figure 5D, we found the ratio of PL to Raman intensity to increase
4
5 strongly with central g-force as the monolayer population increases. This suggests that this
6 intensity ratio, IPL/IRaman, can be quantitatively linked to the monolayer content. We propose
7
8 that the intensity of the Raman spectrum is proportional to the total number of WS2 formula
9
10 units probed by the laser. However, because only the monolayers are luminescent, the PL
11 intensity should be proportional to only the WS2 formula units associated with monolayers.
12
13 This implies that IPL/IRaman should scale linearly with the monolayer volume fraction. To test
14
15 this, in Figure 5E we plot IPL/IRaman versus the monolayer volume fraction, Vf, (as measured
16 by AFM, Figure 3C). We find a linear relationship which holds of >2 decades of both
17
18 IPL/IRaman and Vf. Interestingly, equivalent measurements for WS2 nanosheets exfoliated in
19
20 SDBS and PVA give data which falls on the same curve strongly suggesting that this metric
21 is quite robust towards changes in the stabilizer, even though it may depend on the quality of
22
23 the WS2 starting material. To further test the robustness of this metric, as well as the metrics
24
25 based on UV-Vis spectroscopy, we have redispersed WS2 size-selected by LCC in SC of
26 varying concentrations from 0-5 g/L. As shown in figures S17-18, spectra and hence the
27
28 results of the metric analysis do not change as a function of stabilizer concentration. In
29
30 addition, this shows that the LCC size-selected WS2 can even be redispersed in water without
31 aggregation occurring immediately.
32
33
34 From the fit line in Figure 5E, we find
35
36 1 I PL
37
Vf = (5)
17 I Raman
38
39
40 This relationship holds over a broad range of sizes and monolayer contents, but we note
41
that it eventually breaks down for very small nanosheets, as edges may activate non-radiative
42
43 decay. A first indication that this is indeed the case is obtained from an analysis of the very
44
45 small nanosheets that are discarded in the supernatant after centrifuging at 10 krpm (see SI
46
47 figure S14, L ~ 25 nm). While TEM and optical extinction show no obvious deviation from
48
49 larger nanosheets, the PL/Raman ratio is significantly lower than in the 7.5-10 krpm sample
50 suggesting edge effects to play a role. This issue is addressed further in the supporting
51
52 information.
53
54 Nonetheless, this is an important result, as it means that the monolayer volume fraction
55
56 in a dispersion of WS2 nanosheets can be quantified very simply and easily from a Raman/PL
57
58 measurement. This is hugely advantageous over traditional techniques such as AFM or TEM,
59
60 11
1
2
3 which are very time consuming and can require skill and experience from the experimenter.
4
5 Critically, it allows us to track the monolayer content as we perform further monolayer
6 enrichment as described below.
7
8
9 Optical properties of dispersions highly enriched in monolayers
10
11 In the preceding sections we described a very simple primary centrifugation cascade
12
13
which resulted in the production of a range of size-selected fractions at much higher yields
14 than homogeneous centrifugation (see supplementary Figure S19), the last of which was
15
16 monolayer-enriched up to Nmono/NT~40% and Vf~20%. However, the strength of LCC is its
17
18
versatility. For example, it is also applicable to nanosheets exfoliated in solvents (see SI
19 Figures S20-23). In addition, much more complex cascades can be designed, resulting in
20
21 higher degrees of monolayer enrichment. For example, we find the monolayer content to rise
22
23
rapidly with iteration cycle when repeating the centrifugation at fixed rpm (see
24 supplementary Figure S24).
25
26
In the following, we used the ability to spectroscopically measure the volume fraction
27
28 of monolayers to design secondary cascades with the goal of maximizing IPL/IRaman. Details
29
30 are presented in the Supporting Information (see supplementary Figures S25-38). In brief, we
31
designed centrifugation protocols for monolayer enrichment consisting of multiple iterations
32
33 starting from size-selected dispersions. These typically involved centrifugation at relatively
34
35 low speeds and longer times (overnight) to remove few-layered WS2 and centrifugation at
36
high speeds and short times (1-4h) to remove very small nanosheets (see supplementary
37
38 Figure S25). We note that when designing such a cascade, if highly luminescent dispersions
39
40 are required, the small-nanosheet-removal step is crucial as we found PL/Raman ratios to be
41
significantly lower for very small nanosheets even at high ML contents, probably due to edge
42
43 effects (see supplementary Figures S33-S34).
44
45 Below we will describe secondary cascades (S.C.), which used strongly (6-8 krpm,
46
47 S.C.1, Figure S26-28) and weakly (1.5-10 krpm, S.C.2, Figure S29-31) size-selected
48
49 dispersions as starting points for monolayer enrichment. Shown in Figure 6A is an AFM
50 layer number histogram for the S.C.2 dispersion with a typical image in the inset. It is clear
51
52 that this sample is dominated by monolayers with AFM analysis giving Nmono/NT~74% and
53
54 Vf~70% (see table 1 for all monolayer content data). The Raman/PL spectra of the 6-8 krpm
55 dispersion and both ML enriched dispersions are shown in Figure 6B. While the 6-8 krpm
56
57 dispersion displayed IPL/IRaman=2.8, equivalent to Vf=0.16, it is clear that the enriched samples
58
59
60 12
1
2
3 display considerably more intense PL. We found IPL/IRaman=10.9 and 12.7, for the S.C.1 and
4
5 S.C.2 dispersions, giving Vf of 65% and 75%, respectively, in good agreement with AFM
6 statistics (Table 1).
7
8
9 Shown in Figure 6C are extinction spectra for both the 6-8 krpm sample and the S.C.1
10 and S.C.2 dispersions. The shapes of these spectra are almost indistinguishable because high
11
12 speed centrifugation steps have been included in the cascade to remove very small nanosheets
13
14 giving similar <L> values for these dispersions (confirmed by AFM statistics, see
15 supplementary Figure S28). We note that in this nanosheet size range, scattering15 is virtually
16
17 non-existent so the extinction spectrum is identical to the absorbance spectrum (see
18
19 supplementary Figure S33).
20
21 However, much more interesting is the shape of the A-excitonic components of the
22
23
extinction spectra as shown in the inset (see Figure 4A inset). These show changes not only
24 in center of mass peak position but also in shape. Unlike the spectra associated with thicker
25
26 nanosheets, the 6-8 krpm spectrum clearly has a peak at ~612 nm and a shoulder at ~622 nm
27
28
as discussed above (see Figure 4A). However, the ML-rich samples clearly have A-excitonic
29 responses, which are dominated by a peak close to 610 nm (2.032 eV). This is extremely
30
31 close to the position of the A-excitonic PL peak (2.023 eV, implying a Stokes shift of 10
32
33
meV). Thus we associate the feature at ~610 nm in the extinction/absorbance spectrum of the
34 ML-rich sample with the absorbance of monolayer WS2. This suggests the component at
35
36 ~622 nm represents the combined contribution of few-layered WS2.
37
38 To test this in more detail, we carefully smoothed the extinction spectra of all samples
39
40 (see SI section 7) before finding the second derivative with respect to energy, d2(Ext)/dE2.
41
Such a procedure is known to narrow contributing peaks roughly threefold, allowing
42
43 resolution of closely spaced peaks.58 Samples of the resultant second derivatives are shown in
44
45 Figure 6D. Monolayer-poor samples such as 1.5-2 krpm are dominated by one peak at ~1.98
46
eV while the monolayer-rich sample (S.C.1) is dominated by one peak at ~2.03 eV, with
47
48 intermediate samples showing both components.
49
50 We can extract more information about the components by fitting the second
51
52 derivatives. In the simplest form, a Lorentzian line can be described by
53
54 h
55 L( E ) = (6)
56   ( E − E0 )  2 
57 1 +   
58   w / 2  
59
60 13
1
2
3 where h represents the height, E0 the centre and w the FWHM. As described in the SI (see SI
4
5 section 5), we have fit the second derivative curves to the sum of two doubly differentiated
6 Lorentzians, allowing us to extract height, width and position for both (see supplementary eq.
7
8 S10 and Figure S39). As before, Lorentzians gave better fits than Gaussians (see Figures S40-
9
10 41). We find extremely good fits in all cases (see solid lines in Figure 6D). The higher energy
11 component always had a position of E0=2.035 eV consistent with it representing absorbance
12
13 of the monolayer A-exciton. The lower energy component was found between 1.98 and 1.997
14
15 eV. We associate this component with the individually unresolvable sum of few-layer A-
16 exciton absorptions. In fact, the monolayer component can be differentiated from the few-
17
18 layers only because of the logarithmic dependence embodied in equation 4. This equation
19
20 implies that the energy difference between A-excitonic transitions for 1- and 2- layer
21 nanosheets is 19 meV while being only 11 meV between 2- and 3-layer nanosheets and 8
22
23 meV between 3- and 4-layer nanosheets.
24
25 If the mono- and few-layer assignment described above is correct, then we expect the
26
27 width of the monolayer A-exciton absorbance peak to be invariant with the width of the
28
29 nanosheet thickness distribution. Conversely, the width of the few-layer A-exciton
30 absorbance peak should increase as the thickness distribution broadens. This is exactly what
31
32 we observed as plotted in Figure 6E (the lines are guides to the eye). The FL peak width
33
34 increases from ~65 to ~110 meV as the AFM thickness histogram width increases from 1-4.5.
35 Conversely the monolayer peak is always ~ 55 meV wide. Interestingly this is almost twice
36
37 as wide as the PL peak. This likely partly stems from the fact that all monolayered nanosheets
38
39 are probed in absorbance, while the smallest nanosheets are not captured in the PL
40 measurement (see supplementary Figures S23-S24).
41
42
43
With this peak assignment in mind, we propose that the area under the ML A-exciton
44 extinction peak should scale with the monolayer content in the dispersion. Because the area
45
46 under any Lorentzian is proportional to h × w , we have calculated the metric SA:
47
48 hML wML
49 SA = (7)
50 hML wML + hFL wFL
51
52 which we suggest, should scale with the monolayer volume fraction. We have plotted SA
53
54 versus Vf in Figure 6F, finding good linearity as described by S A = (0.8 ± 0.05)V f . The fact
55
56 that the proportionality constant is so close to 1 is a strong indicator that our assignment is
57
58 correct. This allows us to use SA as an alternative metric for the monolayer volume fraction:
59
60 14
1
2
3 V f = (1.25 ± 0.08) S A (8)
4
5
6 The availability of highly monolayer-enriched dispersions allows us to measure the
7 photoluminescence using a standard PL spectrometer exciting with Xe lamp to obtain
8
9 excitation emission contour plots as shown in Figure 6G. Emission from the A-exciton can be
10
11 well resolved with fitting of the emission spectrum (see supplementary Figure S32) giving
12 position and widths virtually identical to those measured in the Raman spectrometer.
13
14
15 The real advantage of measurement with the PL spectrometer is the ability to measure
16 excitation spectra. Such spectra generally allow the measurement of the absorption spectrum
17
18 of the luminescent species at higher resolution than would usually be possible. As shown in
19
20 Figure 6H, the measured excitation spectrum (λem=617 nm) shows the same spectral features
21
as the extinction spectrum. However, these features, which can be attributed to B, C and D
22
23 excitonic transitions, are considerably sharper in the excitation spectrum. As an inset in
24
25 Figure 6I, we also compare the measured PL spectrum with the deconvoluted A-exciton
26
absorbance contributions from mono-and few layer WS2 (calculated from the second
27
28 derivative fit parameters). The monolayer A-exciton absorbance is very close to the PL with
29
30 only a slight Stokes shift (~10 meV).
31
32 To test our understanding of this behavior, we have used ab initio GW plus Bethe-
33
34 Salpeter method to calculate the absorption spectrum of both mono- and bi-layer WS2 as
35 shown in Figure 6I (see section methods). We find a strong dependence on layer number with
36
37 the A-exciton shifting from 2.06 eV for the ML to 1.99 eV in the bilayer in reasonable
38
39 agreement with the data in Figure 4D. We can model the absorption spectrum of a monolayer
40 rich dispersion (S.C.2) by calculating the weighted average of the theoretical mono- and
41
42 bilayer spectra where the weighting factors are the measured mono- and bilayer volume
43
44 fractions (see Figure 6A, we ignore the small tri-layer population for simplicity). We find
45 excellent agreement between measured and calculated spectra in the A-exciton regime once
46
47 the calculated spectra have been normalized and downshifted by 25 meV. We attribute the
48
49 shift to a combination of environmental and temperature effects.
50
51 Conclusion
52
53 Liquid cascade centrifugation is a simple, powerful and broadly applicable technique
54
55 to separate liquid exfoliated nanosheets by size. It has a number of advantages over other
56
57 techniques; notably its high yield and lack of wastage; the ability to control the concentration
58
59
60 15
1
2
3 of size-selected suspensions, even up to very high concentrations. Probably most important is
4
5 its versatility: cascades can be designed to produce the desired size and thickness
6 distributions and the required degree of monolayer enrichment. Ultimately, we believe
7
8 cascades will be designed to produce dispersions containing only a given nanosheet thickness
9
10 at a predetermined lateral size. In addition, this technique can be applied to virtually any 2D
11 material stabilized by solvents, surfactants or polymers using only benchtop centrifuges.
12
13
14 The ability to easily size-select nanosheets has enabled us to study their optical
15 properties as a function of size and thickness. Similar to MoS2,15 we found the extinction
16
17 spectra of WS2 suspensions to contain quantitative information describing both mean length
18
19 and thickness of the nanosheets while the PL spectra quantify the monolayer content.
20 Spectroscopic measurement of size, thickness and monolayer content allows cascade design
21
22 to achieve further monolayer enrichment, enabling the study of the fundamental optical
23
24 properties of nanosheet ensembles. For example, once the monolayer content was increased
25 beyond Nmono/NT ~25%, fine structure began to appear in the extinction spectra in the vicinity
26
27 of the A-exciton. We found it possible to differentiate the contributions to A-exciton
28
29 extinction of mono- and few-layer nanosheets allowing us to use a simple extinction
30 spectrum to measure the monolayer content of a suspension.
31
32
33
We believe that the ability to easily size select and monolayer enrich dispersions,
34 coupled with the availability of quantitative spectroscopic metrics to assess mean nanosheet
35
36 length, thickness and monolayer volume fraction will have impact for both applications and
37
38
fundamental studies. Nanosheet sizes and thicknesses will be precisely tailored according to
39 needs in samples, which will be available in large quantities. This will be especially critical
40
41 for applications such as printed electronics where dispersions of electronically identical, and
42
43
so uniformly sized, nanosheets are required. Optical analysis will become a tool for quality
44 control allowing the concentration, size, thickness and monolayer content to be assessed for
45
46 any dispersion. In addition, PL will be used as a fingerprint, which is sensitive to doping and
47
48
environmental effects allowing an in situ probe for intermolecular interactions.
49
50
51
52 Methods
53
54 Sample preparation: WS2 dispersions were prepared by probe sonicating the powder. WS2
55
56 (20 g/L) was immersed in 80 mL of aqueous surfactant solution (Csurf= 6g/L). The mixture
57
58 was sonicated under ice-cooling in a metal beaker by probe sonication using a solid flathead
59
60 16
1
2
3 tip (Sonics VX-750) for 1 h at 60 % amplitude. The dispersion was centrifuged in a Hettich
4
5 Mikro 220R centrifuge equipped with a fixed-angle rotor 1016 at 5 krpm (2660 g) for 1.5 h.
6 The supernatant was discarded and the sediment collected in 80 mL of fresh surfactant and
7
8 subjected to a second sonication 5 h at 60 % amplitude with a pulse of 6 s on and 2 s off.
9
10 To select nanosheets by size, we used controlled centrifugation with sequentially
11
12 increasing rotation speeds. Two different rotors were used (see SI). In the standard primary
13
14 cascade, unexfoliated WS2 was removed by centrifugation at 1.5 krpm (240 g, 2 h). The
15 supernatant was subjected to further centrifugation at 2 krpm (426 g, 2 h). The sediment was
16
17 collected in fresh surfactant at reduced volume (3-8 mL), while the supernatant was
18
19 centrifuged at 3 krpm (958 g, 2 h). Again, the sediment was collected and the supernatant
20 subjected to centrifugation at higher speeds. This procedure was repeated with the following
21
22 speeds: 4 krpm (1700 g, 2 h), 5 krpm (2660 g, 2 h), 6 krpm (3506 g, 2 h), 7.5 krpm (5480 g, 2
23
24 h), 10 krpm (9740 g, 2 h). The data presented in Figure 1 uses the central rpm/g-force to
25 express the consecutive centrifugation. For example, the sediment collected from the
26
27 centrifugation between 2-3 krpm has a central rpm of 2.5 krpm (665 g).
28
29 To perform the monolayer enrichment, a sample size-selected by the standard
30
31 procedure was subjected to further iterative centrifugation steps. Details are described in
32
33
Section 4 of the Supplementary Information. The sample of S.C.1 shown in Figure 6 was
34 produced from a standard size selection between 6-8 krpm. The dispersion was then
35
36 centrifuged at 4 krpm (1560 g) for 6 h. the sediment was discarded and the supernatant
37
38
centrifuged at 5 krpm (2435 g) for 14 h. The sediment was discarded and the supernatant
39 centrifuged at 9 krpm (7890 g, 4 h). Very small nanosheets were removed in the supernatant
40
41 after further centrifugation at 15 krpm (21915 g, 1 h). Alternatively, for S.C.2 a dispersion
42
43
containing nanosheets sedimenting between 1.5-10 krpm was centrifuged as follows: 2.5
44 krpm (609 g, 16 h), supernatant subjected to 4 krpm (1560 g, 14 h), supernatant subjected to
45
46 10 krpm (9740 g, 1 h), sediment collected in 1.5 mL and subjected to 5 krpm (2436 g, 5 h),
47
48
supernatant subjected to 8 krpm (6235 g, 2 h), sediment collected and subjected to 3 krpm
49 (877 g, 12 h). The supernatant after this last centrifugation step was collected and had a
50
51 monolayer volume fraction of ~70-75%.
52
53 Characterization: Optical extinction was measured on a Varian Cary 500 in quartz cuvettes
54
55 1 nm increments. Bright field transmission electron microscopy imaging on Holey carbon
56
grids (400 mesh) was performed using a JEOL 2100, operated at 200 kV. Statistical analysis
57
58 was performed of the flake dimensions by measuring the longest axis of the nanosheet and
59
60 17
1
2
3 assigning it “length” and the dimension perpendicular to the longest axis which we defined as
4
5 “width”. Atomic force microscopy (AFM) was carried out on a Veeco Nanoscope-IIIa
6 (Digital Instruments) system in tapping mode after depositing a drop of the dispersion (10
7
8 µL) on a pre-heated (150 °C) Si/SiO2.The apparent thickness was converted to number of
9
10 layers using previously elaborated step-height analysis of liquid-exfoliated nanosheets.15
11 Raman and photoluminescence spectroscopy was performed on the liquid dispersions using a
12
13 Horiba Jobin Yvon LabRAM HR800 with 532 nm excitation laser in air under ambient
14
15 conditions. Great care must be taken during these measurements, as changes in the focal
16 plane during the acquisition will introduce an error in the PL/Raman ratio (see SI methods).
17
18 A drop (~ 40 µL) of a high concentration dispersion was placed on a glass slide and the drop
19
20 edge was optically focused using a 10× objective. The focus for the measurement with the
21 100× objective was readjusted in such a way that the laser was focused slightly above the
22
23 drop. The average of ~5 measurements are displayed. Photoluminescence to obtain the
24
25 contour plot and excitation spectra was measured in quartz cuvettes using an Edinburgh
26 Instruments FS920 PL spectrometer equipped with a Xe lamp (450 W) and a S900
27
28 photomultiplier tube detector at room temperature with single monochromators in excitation
29
30 and emission.
31
32 Theoretical Optical Gap Calculations:
33
34 The optical gap is extracted from the optical conductivity calculated by using the tight-
35
36 binding propagation method (TBPM).59, 60 We adopt an 11-band TB model of few-layered
37
38
WS2 proposed by R. Roldán et al in Ref.61, consisting with five d orbitals of W atom and six
39 p orbitals of S atom as the follows:
40
41
W atoms: , , , , ,
42
43
44 S atoms: , , , , , , , , , , , ,
45
46 where the t and b indexes indicate the top and bottom planes of S atoms within the same
47
48 layer, respectively. The Slater-Koster parameters used to construct the intralayer W-S, W-W
49
50 and S-S hopping matrixes are (in unit of eV) ∆ = −0.872, ∆ = 0.42, ∆ = −2.065, ∆ =
51
−3.468, ∆ = −3.913, !" = 3.603, !# = −0.942, !!" = −1.216, !!# =
52
53 0.177, = 0.243, = 0.749, = 0.236, and for the interlayer S-S hoppings
!!$ " #
54
55 are %" = −0.55, %# = −0.6. The spin-orbital couplings originating from the W and S
56
57 atoms are &' = 0.215, &( = 0.057.
58
59
60 18
1
2
3 The real part of the optical conductivity at finite frequency is calculated via the Kubo’s
4
5 formula as59-61
6
∞
7 1 − e − β hω
σ (ω ) = lim ε → +0 dte i (ω + iε )τ 2i Im φ | J [1 − f (H )]J (τ ) f (H ) | φ ,
hωΩ ∫0
8
9
10
11 where Ω is the sample area, β = 1 / k B T is the inverse temperature, H is the tight-binding
12
13 Hamiltonian, f (H ) = 1 /{exp [β (H − µ )] + 1} is the Fermi-Dirac distribution operator, and
14
15 J (τ ) = exp( iH τ / h ) J exp( −iH τ / h ) is the current operator in the Heisenberg picture. The state
16
17 φ is a normalized random state which covers all the eigenstates in the whole spectrum.59
18
19 The time evolution operator and Fermi-Dirac distribution operator are represented as the
20
21 Chebyshev polynomial expansions.
22
23 Ab initio GW-BSE simulations:
24
25 We start the optical calculations using the Kohn-Sham eigenvectors and energy
26
27 eigenvalues previously calculated within the density-functional theory at generalized-gradient
28
approximation,62 for monolayer and bilayer WS2, using plane waves basis set and periodic
29
30 boundary conditions as implemented in the PWscf63 and Vasp64, 65 codes. Norm-conserving66
31
32 and PAW67, 68’ pseudopotentials are used with a plane wave energy cutoff of 900 eV, with
33
partial-core states included. Atomic coordinates were allowed to relax until all forces were
34
35 smaller in magnitude than 0.01 eV/Å. Relevant lattice constants (in-plane and out-of-plane)
36
37 were optimized for each system. To avoid interactions between supercell images, the distance
38
between periodic images of the WS2 layers along the direction perpendicular to the plane was
39
40 always larger than 20 Å. Spin-orbit interactions are included in the calculations perturbatively
41
42 through the calculations of the spinor wave functions, which are used as an input for the
43
44 calculation of the dielectric functions ε2(ω,q) afterwards. The GW-BSE calculations are done
45 using the Yambo code69 using 300 unoccupied bands in the integration of the self-energy
46
47 term. The number of k-points was chosen according to the Monkhorst–Pack scheme70 and
48
49 was set to the equivalent of a 21 × 21 × 1 grid in the primitive unit cell of WS2, which was
50 previously converged for all structures. The BSE Hamiltonian was created using the ten
51
52 highest valence bands and the six lowest conduction bands using the Tamm-Dancoff
53
54 approach. The response functions were obtained in a fine grid of 1000 energy points using a
55 broadening of 0.04 eV in all calculations.
56
57
58
59
60 19
1
2
3 Supporting Information Available: Detailed methods, materials characterisation and
4
5 description of a range of cascades. This material is available free of charge via the Internet at
6 http://pubs.acs.org.
7
8
9
10
11 Acknowledgement
12
13 The research leading to these results has received funding from the European Union Seventh
14
15 Framework Program under grant agreement n°604391 Graphene Flagship. We have also
16
received support from the Science Foundation Ireland (SFI) funded centre AMBER
17
18 (SFI/12/RC/2278). In addition, JNC acknowledges the European Research Council
19
20 (SEMANTICS) and SFI (11/PI/1087) for financial support. CB acknowledges the German
21
research foundation DFG (BA 4856/1-1). VV-M and CG acknowledge Marie Curie ITN
22
23 network “MoWSeS” (grant no. 317451). EJGS acknowledges the use of computational
24
25 resources provided by the Extreme Science and Engineering Discovery Environment
26
(XSEDE), supported by NSF grants number TG-DMR120049, TG-DMR150017; as well as
27
28 the Queen’s Fellow Award through the startup grant number M8407MPH. We thank David
29
30 McAteer for help preparing figure 1.
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 20
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19 Figure 1: Schematic describing the basic centrifugation cascade employed in this study. The
20
21 sediment discarded after the first centrifugation contains exfoliated layered crystallites while
22
23 the supernatant discarded after the last centrifugation step contains extremely small
24 nanosheets. Size-selected dispersions are prepared by re-dispersing the collected sediments in
25
26 1 g/L aqueous sodium cholate after subsequently increasing centrifugation speeds.
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 21
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21 Figure 2: Microscopic characterization of size-selected WS2 nanosheets. A-D) TEM (A-B)
22 and AFM (C-D) images of nanosheets, size-selected with upper and lower centrifugation
23
24 speeds of (A&C) 1.5 krpm and 2 krpm and (B&D) 6 krpm and 7.5 krpm respectively. E-J)
25
26 Statistical analysis of dimensions of nanosheets extracted from images such as those in A-D.
27 This is presented as histograms representing data for lateral nanosheet size measured by TEM
28
29 (E-F) and nanosheet thickness (number of monolayers per nanosheet, N) determined by AFM
30
31 (G-H). Data is shown for two combinations of centrifugation speed, 1.5-2 krpm and 6-7.5
32 krpm. I, J) Data extracted from histograms for a range of centrifugation conditions. Mean
33
34 values of I) nanosheet length, L , and J) nanosheet thickness, N , as a function of the
35
36 central centrifugation acceleration, presented as central g-force. K-L) Standard deviation of
37
38 nanosheet length (K) and thickness (L) plotted versus mean length and thickness respectively.
39
40 The dashed lines in K and L represent ∆L / L =0.46 and ∆N / N =0.59 respectively.
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 22
1
2
3 A 0.6 B 0.6
4
5
6
Nmono/NT
Nmono/NT
7
0.1 0.1
8 -2.2
9 1
10
11 0.02 0.02
12 200 1000 10000 2 3 4 5 6 78
13 Central g-force (g) Mean N
14 C 0.5 D 100
15
Mass produced (mg)
16 0.1
10
Monolayer Vf
17 All WS2
18 ML-WS2
0.01 1
19 1.9
20
21 0.1
VInitial=80 mL
1E-3
22
23 200 1000 10000 200 1000 10000
24 Central g-force (g) Central g-force (g)
25
26
27 Figure 3: Monolayer population data extracted from histograms for a range of centrifugation
28
conditions. A) Number fraction of monolayers, Nmono/NT, as a function of the central
29
30 centrifugation acceleration, presented as g-force. B) Plot of Nmono/NT versus mean nanosheet
31
32 thickness, N. C) Monolayer volume fraction as a function of g-force. D) Total mass of all
33
WS2 as well as only monolayer-WS2 produced in each fraction as a function of g-force.
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 23
1
2
3 A
50 7
4 6
5
40
5 ε (Lg-1cm-1)
4
3
6 30 600 620 640
1.5-2 krpm
7 20 2-3 krpm
A-exciton
8 3-4 krpm
4-5 krpm Size
10
9 5-6 krpm
6-7.5 krpm
10 0 7.5-10 krpm
11 200 300 400 500 600 700 800
12 Wavelength (nm)
13 B 60 C 2.2
2
14
ε (Lg-1cm-1)
Ext235/Ext290 1.8
40
15 ε235nm = 1.6
47.7 Lg-1cm-1 1.4
16
20 1.2
17 εA
1
18 0
19 10 100 700 10 100 600
20 TEM Length, <L> (nm) TEM Length, <L> (nm)

D E 1.10
21 630
λA from theory (au)
SC
22 625 SDBS
λA (nm)
23 620
PVA
1.05
24
615
25 From PL
Zhu et al.
610 1.00
26 Zhao et al.
27 1 10 1 10
28 AFM Thickness, <N> Theory N
29
30 Figure 4: Dependence of the optical properties of nanosheet dispersions on the nanosheet
31
32 dimensions. A) Optical extinction coefficient spectra measured for WS2 dispersions
33
34 (water/SC) prepared using different centrifugation conditions, and so with different mean
35 nanosheet lengths and thicknesses. Inset: magnified A-exciton region. B) Extinction
36
37 coefficient, measured at 235 nm and at the A-exciton position (~615nm) plotted versus mean
38
39 nanosheet length, as measured by TEM. The mean value of ε235 is 47.7 Lg-1cm-1. C) Ratio of
40
extinction at 235 nm to that at 290 nm plotted versus mean nanosheet length, as measured by
41
42 TEM. The dashed line is a fit to eq. 2. D) A-exciton center of mass position (determined from
43
44 second derivatives) plotted versus mean nanosheet thickness. Also included is data for WS2
45 nanosheets dispersed in SDBS and PVA and from the literature.51, 52
The open symbol
46
47 represents the A-exciton position from photoluminescence measurements. The dashed line
48
49 shows an empirical relationship between λA and N according to eq. 4. E) Calculated relative
50
51
wavelength associated with the optical gap of WS2 (i.e. the direct band gap at the K-point).
52
53
54
55
56
57
58
59
60 24
1
2
3 A 7.5-10 krpm
1.5-2k 2LA(M) WS2 PL
4 4 1
+ E2g
6-7.5 krpm
5-6 krpm
Intensity (au)
5 A1g 4-5 krpm
6 3-4 krpm
2-3 krpm
7 2
200 300 400 500 1.5-2 krpm
-1
8 Wavenumber (cm ) Water Raman
9
WS2 Raman
10
0
11
550 575 600 625 650 675
12
Wavelength (nm)
13
B C 38
14 1.0
WS2 4-5k 36
Intensity (au)
15 0.8
PL w (meV) 34
16 0.6
32
0.4 Water
17
EPL (eV)
2.03
Raman 30 2.02
18 0.2 2.01
L-fit 28
200 1000 10000
0.0
19 26
1.9 2.0 2.1 200 1000 10000
20
21 E (eV) Central g-force (g)
D 10 E
22 10
23
IPL/IRaman
1
IPL/IRaman
1
24
1
25 0.1 0.1 WS2/SC
WS2/PVA
26 0.01 WS2/SDBS
0.01
27 200 1000 10000 1E-3 0.01 0.1 1
28 Central g-force (g) Monolayer Vf
29
30
31 Figure 5: Photoluminescence data for dispersions with different monolayer content. A)
32
33 Photoluminescence (~610 nm) and Raman (~540 nm) spectra of surfactant-stabilized WS2
34 dispersions prepared with different centrifugation conditions and measured in liquid using a
35
36 Raman spectrometer (λexc=532nm). The spectra were normalized to the 2LA(M) Raman
37
38 mode of WS2. The feature at ~650 nm is the water Raman peak. Inset: Raman spectrum of
39 dispersed WS2 nanosheets plotted versus wavenumber. B) Photoluminescence spectrum of
40
41 WS2, plotted on an energy scale and fitted to a Lorentzian. C) PL linewidth, from Lorentzian
42
43 fit, plotted versus central centrifugation acceleration (expressed in units of g). Inset: PL
44 position vs. g-force. D) Ratio of PL intensity to Raman intensity, IPL/IRaman, plotted vs. g-
45
46 force. E) IPL/IRaman plotted versus monolayer volume fraction. The dashed line represents
47
48 I PL / I Raman = 17 × V f .
49
50
51
52
53
54
55
56
57
58
59
60 25
1
2
3 A B
180 12 S.C.2
4
150
Intensity (au)
5 9 S.C.1
Counts
120
6 90 6
100
7 60 nm 6-8
3 300 400
8 30 krpm
9 0 0
1 2 3 4 5 6 0 1000 2000 3000 4000
10
Number of Layers, N Raman Shift (cm-1)
11
12 C 2 D
13 0.22
1.5-2k
Ext/Ext290nm
14
0.18
15 1
16 600 620
5-6k
17 d2(Ext)/dE2 (au)
18 0
200 400 600 800
19 Wavelength (nm)
6-8k
20
E 120
21
AbsA width, w (meV)
22 FL
90
23
24 60
ML ML rich (S.C.1)
25
26 Mean PL width
30
27 1 2 3 4 5 1.90 1.95 2.00 2.05 2.10
28 AFM histo width, ∆N Energy, E (eV)
29 F G
550
hMLwML/[hMLwML+hFLwFL]
30 1 Vf from PL
500 Raman
31 Vf from AFM
λexc (nm)
32 450
33 0.1
400
34
35 350
0.01
36 0.01 0.1 1 580 600 620 640
37 Monolayer Vf λem (nm)
38 H I
20
39 6 PL
Extinction
Intensity (au)
Intensity (au)
40 15
AbsA,ML
4
AbsA,FL
41 10 1.8 1.9 2.0 2.1

42
2 5
43 Excitation 2L
1L
44 0 0
45 1.8 2.2 2.6 3.0 1.8 2.0 2.2 2.4
46 Energy (eV) Energy (eV)
47
48
49 Figure 6: Characterization of second stage monolayer enriched dispersions. A) AFM
50 nanosheet thickness (expressed as layer number) histogram and representative image of WS2
51
52 nanosheets from a dispersion enriched in monolayers by a refined LCC. B) Raman/PL spectra
53
54 of the ML-rich dispersion (red) compared to the first stage size-selected dispersion used for
55 the ML enrichment centrifugation (blue). An alternative secondary cascade yielded a similar
56
57 dispersion (yellow). C) Optical extinction spectra of the three dispersion. Inset: Zoom in of
58
59
60 26
1
2
3 the A-exciton showing clear changes in spectral shape. D) Second derivative spectra for a
4
5 number of WS2-SC dispersions fitted to the sum of the second derivatives of two Lorentzians.
6 With increasing ML volume fraction, a well separated component at 2.033(3) eV (~ 610 nm)
7
8 can clearly be identified which is attributed to ML-WS2. E) Width of absorbance Lorentzian
9
10 curves, representing mono- and few-layer nanosheets, found by fitting second derivatives
11 such as those shown in D plotted versus the full width at half maximum of the corresponding
12
13 AFM thickness histogram. Shown for comparison is the mean width of the PL spectra in
14
15 Figure 5A. F) A-exciton shape monolayer metric, obtained from fitting the second derivative
16 of the extinction spectra to two Lorentzians, as a function of ML volume fraction. The violet
17
18 squares represent data where Vf was measured from AFM, whereas Vf was determined from
19
20 the PL/Raman ML metric in the case of the red data points. The dashed line shows a linear
21 relation that can be used to determine the Vf from the shape of the A exciton according to
22
23 equation 8. G) Excitation-emission contour plot of the ML-rich dispersion measured in a PL
24
25 spectrometer. H) Extinction spectrum of the ML-rich dispersion compared to the excitation
26 spectrum at the ML emission (after subtraction of the water background). The same excitonic
27
28 features are evident in both spectra. Lower inset: A-exciton absorbance deconvoluted into the
29
30 individual components of ML and few-layer WS2 as well as their sum. Upper inset: Measured
31 PL spectrum (450 nm excitation). I) Theoretical absorption curves for monolayer (1L) and bi-
32
33 layer (2L) WS2. Inset: Extinction curve measured for S.C.2 monolayer enriched sample (open
34
35 symbols). Also shown is the weighted sum ( Abs = 0.71 Abs1L + 0.24 Abs2L , black line) of the
36
37 theoretical monolayer and bilayer absorption spectrum. These weightings were chosen to
38 reflect the volume fraction of monolayers and bilayers in the S.C.2 sample as measured by
39
40 AFM (see Figure 6A) where we have neglected the small population of trilayers for
41
42 simplicity. N.B the theory curve has been downshifted to 25 meV to match the experimental
43 data.
44
45
46
47
Sample <N> Nmono/NT Vf IPL/IRaman
48
49 6-8 krpm 2.4 0.30 0.15 2.7
50 S.C.1 1.6 0.73 0.64 10.8
51 1.5-10 krpm 9.0 0.04 0.01 0.18
52 S.C.2 1.6 0.74 0.70 12.7
53
54
55 Table 1: Summary of data relating to monolayer content for size-selected starting dispersions
56 and dispersions enriched by secondary cascades.
57
58
59
60 27
1
2
3 ToC fig
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21 References
22
23 1. Nicolosi, V.; Chhowalla, M.; Kanatzidis, M. G.; Strano, M. S.; Coleman, J. N., Liquid
24 Exfoliation of Layered Materials. Science 2013, 340, 1420-+.
25 2. Hernandez, Y.; Nicolosi, V.; Lotya, M.; Blighe, F. M.; Sun, Z.; De, S.; McGovern, I. T.;
26 Holland, B.; Byrne, M.; Gun'Ko, Y. K.; Boland, J. J.; Niraj, P.; Duesberg, G.; Krishnamurthy, S.;
27 Goodhue, R.; Hutchison, J.; Scardaci, V.; Ferrari, A. C.; Coleman, J. N., High-Yield Production of
28 Graphene by Liquid-Phase Exfoliation of Graphite. Nat. Nanotechnol. 2008, 3, 563-568.
29 3. Green, A. A.; Hersam, M. C., Solution Phase Production of Graphene with Controlled
30 Thickness Via Density Differentiation. Nano Lett. 2009, 9, 4031-4036.
31 4. Coleman, J. N.; Lotya, M.; O'Neill, A.; Bergin, S. D.; King, P. J.; Khan, U.; Young, K.;
32 Gaucher, A.; De, S.; Smith, R. J.; Shvets, I. V.; Arora, S. K.; Stanton, G.; Kim, H.-Y.; Lee, K.; Kim,
33 G. T.; Duesberg, G. S.; Hallam, T.; Boland, J. J.; Wang, J. J., et al., Two-Dimensional Nanosheets
34 Produced by Liquid Exfoliation of Layered Materials. Science 2011, 331, 568-571.
35 5. Zhi, C. Y.; Bando, Y.; Tang, C. C.; Kuwahara, H.; Golberg, D., Large-Scale Fabrication of
36 Boron Nitride Nanosheets and Their Utilization in Polymeric Composites with Improved Thermal and
37 Mechanical Properties. Adv. Mater. 2009, 21, 2889-+.
38 6. Bourlinos, A. B.; Georgakilas, V.; Zboril, R.; Steriotis, T. A.; Stubos, A. K., Liquid-Phase
39 Exfoliation of Graphite Towards Solubilized Graphenes. Small 2009, 5, 1841-1845.
40 7. Ciesielski, A.; Samori, P., Graphene Via Sonication Assisted Liquid-Phase Exfoliation.
41 Chem. Soc. Rev. 2014, 43, 381-398.
42 8. Varrla, E.; Backes, C.; Paton, K. R.; Harvey, A.; Gholamvand, Z.; McCauley, J.; Coleman, J.
43 N., Large-Scale Production of Size-Controlled Mos2 Nanosheets by Shear Exfoliation. Chem. Mater.
44 2015, 27, 1129-1139.
45 9. Paton, K. R.; Varrla, E.; Backes, C.; Smith, R. J.; Khan, U.; O’Neill, A.; Boland, C.; Lotya,
46 M.; Istrate, O. M.; King, P.; Higgins, T.; Barwich, S.; May, P.; Puczkarski, P.; Ahmed, I.; Moebius,
47 M.; Pettersson, H.; Long, E.; Coelho, J.; O’Brien, S. E., et al., Scalable Production of Large
48 Quantities of Defect-Free Few-Layer Graphene by Shear Exfoliation in Liquids. Nat. Mater. 2014, 13,
49 624-630.
50
10. Hernandez, Y.; Lotya, M.; Rickard, D.; Bergin, S. D.; Coleman, J. N., Measurement of
51
Multicomponent Solubility Parameters for Graphene Facilitates Solvent Discovery. Langmuir 2010,
52
26, 3208-3213.
53
11. Cunningham, G.; Lotya, M.; Cucinotta, C. S.; Sanvito, S.; Bergin, S. D.; Menzel, R.; Shaffer,
54
M. S. P.; Coleman, J. N., Solvent Exfoliation of Transition Metal Dichalcogenides: Dispersibility of
55
Exfoliated Nanosheets Varies Only Weakly between Compounds. ACS Nano 2012, 6, 3468-3480.
56
57
12. Lotya, M.; Hernandez, Y.; King, P. J.; Smith, R. J.; Nicolosi, V.; Karlsson, L. S.; Blighe, F.
58 M.; De, S.; Wang, Z. M.; McGovern, I. T.; Duesberg, G. S.; Coleman, J. N., Liquid Phase Production
59
60 28
1
2
3 of Graphene by Exfoliation of Graphite in Surfactant/Water Solutions. J. Am. Chem. Soc. 2009, 131,
4 3611-3620.
5 13. Smith, R. J.; King, P. J.; Lotya, M.; Wirtz, C.; Khan, U.; De, S.; O'Neill, A.; Duesberg, G. S.;
6 Grunlan, J. C.; Moriarty, G.; Chen, J.; Wang, J.; Minett, A. I.; Nicolosi, V.; Coleman, J. N., Large-
7 Scale Exfoliation of Inorganic Layered Compounds in Aqueous Surfactant Solutions. Adv. Mater.
8 2011, 23, 3944-3948.
9 14. Wu, M.; Wang, Y.; Lin, X.; Yu, N.; Wang, L.; Wang, L.; Hagfeldt, A.; Ma, T., Economical
10 and Effective Sulfide Catalysts for Dye-Sensitized Solar Cells as Counter Electrodes. Phys. Chem.
11 Chem. Phys. 2011, 13, 19298-301.
12 15. Backes, C.; Smith, R. J.; McEvoy, N.; Berner, N. C.; McCloskey, D.; Nerl, H. C.; O’Neill, A.;
13 King, P. J.; Higgins, T.; Hanlon, D.; Scheuschner, N.; Maultzsch, J.; Houben, L.; Duesberg, G. S.;
14 Donegan, J. F.; Nicolosi, V.; Coleman, J. N., Edge and Confinement Effects Allow in Situ
15 Measurement of Size and Thickness of Liquid-Exfoliated Nanosheets. Nature Commun. 2014, 5,
16 4576.
17 16. May, P.; Khan, U.; Hughes, J. M.; Coleman, J. N., Role of Solubility Parameters in
18 Understanding the Steric Stabilization of Exfoliated Two-Dimensional Nanosheets by Adsorbed
19 Polymers. J. Phys. Chem. C 2012, 116, 11393-11400.
20 17. Liang, Y. T.; Hersam, M. C., Highly Concentrated Graphene Solutions Via Polymer
21 Enhanced Solvent Exfoliation and Iterative Solvent Exchange. J. Am. Chem. Soc. 2010, 132, 17661-
22 17663.
23 18. Khan, U.; May, P.; O'Neill, A.; Bell, A. P.; Boussac, E.; Martin, A.; Semple, J.; Coleman, J.
24 N., Polymer Reinforcement Using Liquid-Exfoliated Boron Nitride Nanosheets. Nanoscale 2013, 5,
25 581-587.
26 19. Bourlinos, A. B.; Georgakilas, V.; Zboril, R.; Steriotis, T. A.; Stubos, A. K.; Trapalis, C.,
27 Aqueous-Phase Exfoliation of Graphite in the Presence of Polyvinylpyrrolidone for the Production of
28
Water-Soluble Graphenes. Solid State Commun. 2009, 149, 2172-2176.
29
20. Coleman, J. N., Liquid Exfoliation of Defect-Free Graphene. Acc. Chem. Res. 2013, 46, 14-
30
22.
31
21. Nacken, T. J.; Damm, C.; Walter, J.; Ruger, A.; Peukert, W., Delamination of Graphite in a
32
High Pressure Homogenizer. RSC Adv. 2015, 5, 57328-57338.
33
22. Zhou, K. G.; Mao, N. N.; Wang, H. X.; Peng, Y.; Zhang, H. L., A Mixed-Solvent Strategy for
34
35
Efficient Exfoliation of Inorganic Graphene Analogues. Angew. Chem.,Int. Ed. 2011, 50, 10839-
36 10842.
37 23. Bang, G. S.; Nam, K. W.; Kim, J. Y.; Shin, J.; Choi, J. W.; Choi, S. Y., Effective Liquid-
38 Phase Exfoliation and Sodium Ion Battery Application of Mos2 Nanosheets. ACS Appl. Mater.
39 Interfaces 2014, 6, 7084-7089.
40 24. Alsaif, M.; Balendhran, S.; Field, M. R.; Latham, K.; Wlodarski, W.; Ou, J. Z.; Kalantar-
41 Zadeh, K., Two Dimensional Alpha-Moo3 Nanoflakes Obtained Using Solvent-Assisted Grinding and
42 Sonication Method: Application for H-2 Gas Sensing. Sens. Actuators, B 2014, 192, 196-204.
43 25. Hanlon, D.; Backes, C.; Higgins, T. M.; Hughes, M.; O’Neill, A.; King, P.; McEvoy, N.;
44 Duesberg, G. S.; Mendoza Sanchez, B.; Pettersson, H.; Nicolosi, V.; Coleman, J. N., Production of
45 Molybdenum Trioxide Nanosheets by Liquid Exfoliation and Their Application in High-Performance
46 Supercapacitors. Chem. Mater. 2014, 26, 1751-1763.
47 26. Harvey, A.; Backes, C.; Gholamvand, Z.; Hanlon, D.; McAteer, D.; Nerl, H. C.; McGuire, E.;
48 Seral-Ascaso, A.; Ramasse, Q. M.; McEvoy, N.; Winters, S.; Berner, N. C.; McCloskey, D.; Donegan,
49 J.; Duesberg, G.; Nicolosi, V.; Coleman, J. N., Preparation of Gallium Sulfide Nanosheets by Liquid
50 Exfoliation and Their Application as Hydrogen Evolution Catalysts. Chem. Mater. 2015, 27, 3483–
51 3493.
52 27. Yasaei, P.; Kumar, B.; Foroozan, T.; Wang, C.; Asadi, M.; Tuschel, D.; Indacochea, J. E.;
53 Klie, R. F.; Salehi-Khojin, A., High-Quality Black Phosphorus Atomic Layers by Liquid-Phase
54 Exfoliation. Adv. Mater. 2015, 27, 1887-1892.
55 28. Vishnu Sresht; Agilio A.H. Padua; Blankschtein, Liquid-Phase Exfoliation of Phosphorene:
56 Design Rules from Molecular Dynamics Simulations. ACS Nano 2015, DOI:
57 10.1021/acsnano.5b02683.
58
59
60 29
1
2
3 29. Kang, J.; Wood, J. D.; Wells, S. A.; Lee, J.-H.; Liu, X.; Chen, K.-S.; Hersam, M. C., Solvent
4 Exfoliation of Electronic-Grade, Two-Dimensional Black Phosphorus. ACS Nano 2015, 9, 3596-3604.
5 30. Brent, J. R.; Savjani, N.; Lewis, E. A.; Haigh, S. J.; Lewis, D. J.; O'Brien, P., Production of
6 Few-Layer Phosphorene by Liquid Exfoliation of Black Phosphorus. Chem. Commun. 2014, 50,
7 13338-13341.
8 31. Hanlon, D.; Backes, C.; Doherty, E.; Cucinotta, C. S.; Berner, N. C.; Boland, C.; Lee, K.;
9 Lynch, P.; Gholamvand, Z.; Harvey, A.; Zhang, S.; Wang, K.; Moynihan, G.; Pokle, A.; Ramasse, Q.
10 M.; McEvoy, N.; Blau, W. J.; Wang, J.; Abellan, G.; Hauke, F., et al., Liquid Exfoliation of Solvent-
11 Stabilised Few-Layer Black Phosphorus for Applications Beyond Electronics. Nature Commun. 2015,
12 6, 8563.
13 32. Naguib, M.; Mashtalir, O.; Carle, J.; Presser, V.; Lu, J.; Hultman, L.; Gogotsi, Y.; Barsoum,
14 M. W., Two-Dimensional Transition Metal Carbides. ACS Nano 2012, 6, 1322-1331.
15 33. Zhou, K.; Jiang, S.; Bao, C.; Song, L.; Wang, B.; Tang, G.; Hu, Y.; Gui, Z., Preparation of
16 Poly(Vinyl Alcohol) Nanocomposites with Molybdenum Disulfide (Mos2): Structural Characteristics
17 and Markedly Enhanced Properties. RSC Adv. 2012, 2, 11695-11703.
18 34. Paton, K. R.; Varrla, E.; Backes, C.; Smith, R. J.; Khan, U.; O'Neill, A.; Boland, C.; Lotya,
19 M.; Istrate, O. M.; King, P.; Higgins, T.; Barwich, S.; May, P.; Puczkarski, P.; Ahmed, I.; Moebius,
20 M.; Pettersson, H.; Long, E.; Coelho, J.; O'Brien, S. E., et al., Scalable Production of Large Quantities
21 of Defect-Free Few-Layer Graphene by Shear Exfoliation in Liquids. Nat. Mater. 2014, 13, 624-30.
22 35. Eksik, O.; Gao, J.; Shojaee, S. A.; Thomas, A.; Chow, P.; Bartolucci, S. F.; Lucca, D. A.;
23 Koratkar, N., Epoxy Nanocomposites with Two-Dimensional Transition Metal Dichalcogenide
24 Additives. ACS Nano 2014, 8, 5282-5289.
25 36. Sorrentino, A.; Altavilla, C.; Merola, M.; Senatore, A.; Ciambelli, P.; Iannace, S., Nanosheets
26 of Mos2-Oleylamine as Hybrid Filler for Self-Lubricating Polymer Composites: Thermal,
27 Tribological, and Mechanical Properties. Polym. Compos. 2015, 36, 1124-1134.
28
37. May, P.; Khan, U.; O'Neill, A.; Coleman, J. N., Approaching the Theoretical Limit for
29
Reinforcing Polymers with Graphene. J. Mater. Chem. 2012, 22, 1278-1282.
30
38. Wang, J.-Z.; Lu, L.; Lotya, M.; Coleman, J. N.; Chou, S.-L.; Liu, H.-K.; Minett, A. I.; Chen,
31
J., Development of Mos2-Cnt Composite Thin Film from Layered Mos2 for Lithium Batteries.
32
Advanced Energy Materials 2013, 3, 798-805.
33
39. Pumera, M.; Sofer, Z.; Ambrosi, A., Layered Transition Metal Dichalcogenides for
34
35
Electrochemical Energy Generation and Storage. J. Mater. Chem. C. 2014, 2, 8981-8987.
36 40. Woodward, R. I.; Kelleher, E. J. R.; Howe, R. C. T.; Hu, G.; Torrisi, F.; Hasan, T.; Popov, S.
37 V.; Taylor, J. R., Tunable Q-Switched Fiber Laser Based on Saturable Edge-State Absorption in Few-
38 Layer Molybdenum Disulfide (Mos2). Optics Express 2014, 22, 31113-31122.
39 41. Zhang, M.; Howe, R. T.; Woodward, R.; Kelleher, E. R.; Torrisi, F.; Hu, G.; Popov, S.;
40 Taylor, J. R.; Hasan, T., Solution Processed Mos2-Pva Composite for Sub-Bandgap Mode-Locking of
41 a Wideband Tunable Ultrafast Er:Fiber Laser. Nano Research 2015, 8, 1522-1534.
42 42. Sun, Z.; Hasan, T.; Torrisi, F.; Popa, D.; Privitera, G.; Wang, F.; Bonaccorso, F.; Basko, D.
43 M.; Ferrari, A. C., Graphene Mode-Locked Ultrafast Laser. ACS Nano 2010, 4, 803-810.
44 43. Khan, U.; O'Neill, A.; Porwal, H.; May, P.; Nawaz, K.; Coleman, J. N., Size Selection of
45 Dispersed, Exfoliated Graphene Flakes by Controlled Centrifugation. Carbon 2012, 50, 470-475.
46 44. Jaramillo, T. F.; Jorgensen, K. P.; Bonde, J.; Nielsen, J. H.; Horch, S.; Chorkendorff, I.,
47 Identification of Active Edge Sites for Electrochemical H-2 Evolution from Mos2 Nanocatalysts.
48 Science 2007, 317, 100-102.
49 45. Torrisi, F.; Coleman, J. N., Electrifying Inks with 2d Materials. Nat. Nanotechnol. 2014, 9,
50 738-739.
51 46. Kang, J.; Seo, J.-W. T.; Alducin, D.; Ponce, A.; Yacaman, M. J.; Hersam, M. C., Thickness
52 Sorting of Two-Dimensional Transition Metal Dichalcogenides Via Copolymer-Assisted Density
53 Gradient Ultracentrifugation. Nat Commun 2014, 5, 5478.
54 47. O’Neill, A.; Khan, U.; Coleman, J. N., Preparation of High Concentration Dispersions of
55 Exfoliated Mos2 with Increased Flake Size. Chem. Mater. 2012, 24, 2414-2421.
56 48. Barwich, S.; Khan, U.; Coleman, J. N., A Technique to Pretreat Graphite Which Allows the
57 Rapid Dispersion of Defect-Free Graphene in Solvents at High Concentration. J. Phys. Chem. C 2013,
58 117, 19212-19218.
59
60 30
1
2
3 49. Kouroupis-Agalou, K.; Liscio, A.; Treossi, E.; Ortolani, L.; Morandi, V.; Pugno, N. M.;
4 Palermo, V., Fragmentation and Exfoliation of 2-Dimensional Materials: A Statistical Approach.
5 Nanoscale 2014, 6, 5926-5933.
6 50. Lin, Y.; Ling, X.; Yu, L.; Huang, S.; Hsu, A. L.; Lee, Y.-H.; Kong, J.; Dresselhaus, M. S.;
7 Palacios, T., Dielectric Screening of Excitons and Trions in Single-Layer Mos2. Nano Lett. 2014, 14,
8 5569-5576.
9 51. Zhao, W.; Ghorannevis, Z.; Chu, L.; Toh, M.; Kloc, C.; Tan, P.-H.; Eda, G., Evolution of
10 Electronic Structure in Atomically Thin Sheets of Ws2 and Wse2. ACS Nano 2012, 7, 791-797.
11 52. Zhu, B.; Chen, X.; Cui, X., Exciton Binding Energy of Monolayer Ws2. Sci. Rep. 2015, 5.
12 53. Scheuschner, N.; Ochedowski, O.; Kaulitz, A.-M.; Gillen, R.; Schleberger, M.; Maultzsch, J.,
13 Photoluminescence of Freestanding Single- and Few-Layer Mos2. Phys. Rev. B 2014, 89, 125406.
14 54. Mak, K. F.; Lee, C.; Hone, J.; Shan, J.; Heinz, T. F., Atomically Thin Mos2. A New Direct-
15 Gap Semiconductor. Phys. Rev. Lett. 2010, 105, 136805/1-136805/4.
16 55. Splendiani, A.; Sun, L.; Zhang, Y.; Li, T.; Kim, J.; Chim, C.-Y.; Galli, G.; Wang, F.,
17 Emerging Photoluminescence in Monolayer Mos2. Nano Lett. 2010, 10, 1271-1275.
18 56. Gutiérrez, H. R.; Perea-López, N.; Elías, A. L.; Berkdemir, A.; Wang, B.; Lv, R.; López-
19 Urías, F.; Crespi, V. H.; Terrones, H.; Terrones, M., Extraordinary Room-Temperature
20 Photoluminescence in Triangular Ws2 Monolayers. Nano Lett. 2012, 13, 3447-3454.
21 57. Kobayashi, Y.; Sasaki, S.; Mori, S.; Hibino, H.; Liu, Z.; Watanabe, K.; Taniguchi, T.;
22 Suenaga, K.; Maniwa, Y.; Miyata, Y., Growth and Optical Properties of High-Quality Monolayer
23 Ws2 on Graphite. ACS Nano 2015, 9, 4056-4063.
24 58. Saakov, V. S.; Drapkin, V. Z.; Krivchenko, A. I.; Rozengart, E. V.; Bogachev, Y. V.;
25 Knyazev, M. N., Derivative Spectrophotometry and Electron Spin Resonance (Esr) Spectroscopy for
26 Ecological and Biological Questions. Springer-Verlag Wien: 2013.
27 59. Yuan, S.; De Raedt, H.; Katsnelson, M. I., Modeling Electronic Structure and Transport
28
Properties of Graphene with Resonant Scattering Centers. Phys. Rev. B 2010, 82, 115448.
29
60. Yuan, S.; Roldán, R.; De Raedt, H.; Katsnelson, M. I., Optical Conductivity of Disordered
30
Graphene Beyond the Dirac Cone Approximation. Phys. Rev. B 2011, 84, 195418.
31
61. Roldán, R.; López-Sancho, M. P.; Guinea, F.; Cappelluti, E.; Silva-Guillén, J. A.; Ordejón, P.,
32
Momentum Dependence of Spin–Orbit Interaction Effects in Single-Layer and Multi-Layer Transition
33
Metal Dichalcogenides. 2D Materials 2014, 1, 034003.
34
35
62. Perdew, J. P.; Burke, K.; Ernzerhof, M., Generalized Gradient Approximation Made Simple.
36 Phys. Rev. Lett. 1996, 77, 3865-3868.
37 63. Giannozzi, P.; Baroni, S.; Bonini, N.; Calandra, M.; Car, R.; Cavazzoni, C.; Ceresoli, D.;
38 Chiarotti, G. L.; Cococcioni, M.; Dabo, I.; Dal Corso, A.; de Gironcoli, S.; Fabris, S.; Fratesi, G.;
39 Gebauer, R.; Gerstmann, U.; Gougoussis, C.; Kokalj, A.; Lazzeri, M.; Martin-Samos, L., et al.,
40 Quantum Espresso: A Modular and Open-Source Software Project for Quantum Simulations of
41 Materials. Journal of Physics-Condensed Matter 2009, 21, 395502.
42 64. Kresse, G.; Furthmuller, J., Efficient Iterative Schemes for Ab Initio Total-Energy
43 Calculations Using a Plane-Wave Basis Set. Phys. Rev. B 1996, 54, 11169-11186.
44 65. Kresse, G.; Hafner, J., Ab-Initio Molecular-Dynamics for Open-Shell Transition-Metals.
45 Phys. Rev. B 1993, 48, 13115-13118.
46 66. Troullier, N.; Martins, J. L., Efficient Pseudopotentials for Plane-Wave Calculations. Phys.
47 Rev. B 1991, 43, 1993-2006.
48 67. Blochl, P. E., Projector Augmented-Wave Method. Phys. Rev. B 1994, 50, 17953-17979.
49 68. Kresse, G.; Joubert, D., From Ultrasoft Pseudopotentials to the Projector Augmented-Wave
50 Method. Phys. Rev. B 1999, 59, 1758-1775.
51 69. Marini, A.; Hogan, C.; Gruening, M.; Varsano, D., Yambo: An Ab Initio Tool for Excited
52 State Calculations. Comput. Phys. Commun. 2009, 180, 1392-1403.
53 70. Monkhorst, H. J.; Pack, J. D., Special Points for Brillouin-Zone Integrations. Phys. Rev. B
54 1976, 13, 5188-5192.
55
56
57
58
59
60 31

Production of Highly Monolayer Enriched Dispersions of Liquid-Exfoliated Nanosheets by Liquid Cascade Centrifugation

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Production of Highly Monolayer Enriched Dispersions of Liquid-Exfoliated Nanosheets by Liquid Cascade Centrifugation

Hochgeladen von

Copyright:

Verfügbare Formate

Subscriber access provided by ORTA DOGU TEKNIK UNIVERSITESI KUTUPHANESI

11 200 300 400 500 600 700 800

20 TEM Length, <L> (nm) TEM Length, <L> (nm)

28 AFM Thickness, <N> Theory N

41 10 1.8 1.9 2.0 2.1

Das könnte Ihnen auch gefallen