The authors report the characterization of individual free-standing 5 nm gold nanoparticles deposited on ultrathin graphene film by cluster secondary ion mass spectrometry (SIMS) in the transmission direction. For primary ions, the authors used C601,2+ and Au4004+ at impact energies of approximately ∼0.42, 0.83, and 1.3 keV/atom, respectively. The experiments were run as a sequence of single projectile impacts with each time separate recording of the secondary ions identified via time-of-flight-mass spectrometer. Graphene generates small mass secondary ions (SIs). It contributes little beyond m/z 120, facilitating the detection of moieties attached to the nanoparticles. From the SI yield of the molecular ion, it can be determined on which side of the graphene the analyte is deposited. Compared to the conventional reflection SIMS, the transmission SIMS shows a ∼4 times higher effective yield of molecular ions from dodecanethiol layer on nanoparticles. The SI yields from Au4004+ bombardment are ∼3 times higher than those from C602+ bombardment for the analysis of nanoparticles on graphene in the transmission direction.

The analysis of nanoparticles (NPs) with SIMS can be handled in one of two ways: analyzing an ensemble of NPs or testing them one by one. The latter enables to track changes in chemical reactivity with composition, a key issue when the surface to volume ratios are large.1 However, extracting chemical information from a single vanishingly small object is very difficult to impossible. We side-step the limitation by probing a large number of NPs one-by-one and record the emissions from each NP separately. A large collection of NP will likely contain subsets of like-nanoparticles. Their data can be summed for statistics.2 In this case, NPs must be dispersed to eliminate interaction among neighbors. Another concern is the contribution from the substrate. A solution is to deposit the NPs on graphene to reduce substrate contribution to the overall mass spectrum. Moreover, it then becomes feasible to run experiments in transmission, i.e., collecting the secondary ions (SIs) in the forward direction, where emission is enhanced in comparison to the conventional backward emission.3 We present here a study of transmission SIMS for the analysis of NPs, specifically 5 nm dodecanethiol-coated gold NPs deposited on graphene. We discuss below the characterization of graphene and of the NPs using C601,2+ and Au4004+ as projectiles at impact energies of ∼0.42, 0.83, and 1.3 keV/atom, respectively. The latter were chosen to maximize detection sensitivity, as they generate secondary ion yields, which are two to three orders of magnitude larger than those from equal velocity atomic ions.4 

The graphene films used in this study were 3–5 layers (3–5L) free-standing graphene films on a lacey carbon net supported by a 300 mesh 3.05 mm standard copper TEM grid (Ted Pella, Inc., Redding, CA 96003). The coverage of graphene was typically 70%–90%, and this was verified by scanning electron microscopy (Fig. 1). The grid was fixed on a 2 mm hole on a sample holder using silver print (MG Chemicals, Surrey, BC, Canada V4N 4E7). The dodecanethiol-coated 5 nm gold nanoparticles (Nanoprobes, Yaphank, NY 11980) were suspended in hexane and diluted to 0.2 mg/mL. One microliter of the solution was dropcast on the graphene film to form a submonolayer of Au NPs with a surface coverage of ∼50%. The Au NP has a 3–5 nm Au core (∼30 000 Au atoms) coated by a monolayer of ∼2 nm dodecanethiol. The TEM images of Au NPs (Fig. 2) verified that the deposited Au NPs are self-assembled to form a submonolayer without agglomeration. Au NPs with the same concentration and volume were also dropcast on bulk pyrolytic graphite.

Fig. 1.

(Color online) SEM images of the (a) graphene film on 3.05 mm Cu TEM grid fixed on a sample holder and (b) a square of TEM grid, showing the graphene film supported by a lacey carbon net.

Fig. 1.

(Color online) SEM images of the (a) graphene film on 3.05 mm Cu TEM grid fixed on a sample holder and (b) a square of TEM grid, showing the graphene film supported by a lacey carbon net.

Close modal
Fig. 2.

TEM image of the 5 nm dodecanethiol-coated Au NPs on 3–5L graphene.

Fig. 2.

TEM image of the 5 nm dodecanethiol-coated Au NPs on 3–5L graphene.

Close modal

The experiments were run on a custom-built SIMS instrument with an effusive C60 source coupled to a linear time-of-flight mass spectrometer and a custom-built SIMS instrument with an Au liquid metal ion source coupled to a linear time-of-flight (TOF) mass spectrometer. Figure 3 shows the schematic of the C60 instrument. A detailed description can be found elsewhere.5 The samples were bombarded with individual 25 keV C60+, 50 keV C602+, and 520 keV Au4004+ projectiles. The bombardment rate was adjusted to ∼1000 projectiles per second; thus, it was virtually impossible that multiple impacts hit the same site for 106 total impacts (<0.1% of the surface is analyzed). The projectile-graphene bombardment angle was set at normal. The impact angle is critical for recovering a maximum of SIs. The SIs and secondary electrons from each individual impact were detected separately in the transmission direction by using the “event-by-event bombardment/detection mode.”6 The data were recorded and processed using custom-designed software.7 By selecting a specific ion of interest in the total mass spectrum, the coemitted and therefore colocalized ions were extracted, resulting in a coincidence ion mass spectrum.2 From the coincidence ion mass spectrum, one can calculate the effective yield (Ye), which is determined as follows:5 

Ye,A=IA,BIB,
(1)

where Ye,A is the effective yield of ion A, IA,B is the intensity of ion A in the coincidence ion mass spectrum with ion B, and IB is the intensity of ion B in the total mass spectrum. Ye is the number of a specified SI emitted per projectile impact on the NP, excluding impacts on the substrate. Ye also accounts for differences in NP coverage among samples.

Fig. 3.

(Color online) Schematic of the C60 SIMS instrument.

Fig. 3.

(Color online) Schematic of the C60 SIMS instrument.

Close modal

The 3–5 layer graphene was bombarded with 25 keV C60+ and 50 keV C602+ projectiles in the transmission mode (spectra shown in Fig. 4). The Cn carbon clusters ranging from C to C10 followed by CnH and CnH2 are the main features of this spectrum. Beyond m/z 120, the contribution from graphene becomes negligible, an advantageous feature for characterizing functionalized NPs. The first carbon peak C has a distinct tail shape in the spectra obtained with C60 projectiles at 25 and 50 keV, respectively. This feature does not appear on large carbon cluster peaks. The initial kinetic energy distribution of C extends up to 1/60 of the kinetic energy of the incident projectiles: 0.42 keV for 25 keV C60+ and 0.83 keV for 50 keV C602+, which is contributed by the knocked-on carbon atoms from the graphene and the shattered carbon atoms from the projectiles.8 

Fig. 4.

(Color online) Negative mass spectra of the 3–5L graphene in transmission bombarded with (a) 25 keV C60+, (b) 50 keV C602+, and (c) details of the high-energy tails of C peaks (peak height is normalized to total events).

Fig. 4.

(Color online) Negative mass spectra of the 3–5L graphene in transmission bombarded with (a) 25 keV C60+, (b) 50 keV C602+, and (c) details of the high-energy tails of C peaks (peak height is normalized to total events).

Close modal

The negative spectra of Au NPs on graphene bombarded with 50 keV C602+ are presented in Fig. 5. In spectrum (a), C602+ bombarded the graphene film first, then the Au NPs, while in spectrum (b), C602+ bombard Au NPs first, and then the graphene film. It must be noted that the graphene film is supported by a lacey carbon net, which has a thickness of ∼100 nm.9 Thus, in the transmission direction, no start signal can be obtained from the lacey carbon net or Cu grid. Thus, virtually all signals are from impacts on graphene. In both spectra, the peaks of Au, Au adduct ions, and the oxidized molecular ion (C12H25SO3, m/z 249) from the layer of dodecanethiol were observed. The mass resolution of the Au peak is about 350. The effective yields of Au are similar in both cases (Ye = 1.7% and 1.8%, respectively). All effective yields are measured in the coincidence mass spectra with SH, which is a characteristic peak of the Au NPs. However, the effective yield of the molecular ion peak in (b) is lower than that in (a) (Ye = 0.15% and 0.62%, respectively). We attribute the lower yield in case (b) to the emission of the molecular moiety being blocked by the graphene film, while single Au atomic ions and small Au adducts are able to penetrate through the graphene film. However, when the projectiles impact on graphene first, there is no hindrance to the SI emission. Therefore, the effective yield of the oxidized molecular ion of dodecanethiol enables to determine on which side of the graphene the Au NPs are deposited.

Fig. 5.

(Color online) Negative mass spectra of the 5 nm Au NPs deposited on 3–5L graphene in transmission bombarded with 50 keV C602+. (a) Graphene was bombarded first; (b) Au NPs were bombarded first (peak height is normalized by the number of projectile impacts).

Fig. 5.

(Color online) Negative mass spectra of the 5 nm Au NPs deposited on 3–5L graphene in transmission bombarded with 50 keV C602+. (a) Graphene was bombarded first; (b) Au NPs were bombarded first (peak height is normalized by the number of projectile impacts).

Close modal

The difference in the data obtained from the NPs in transmission versus from the same NPs deposited on a thick substrate is illustrated in Fig. 6. Spectrum (a) is from a submonolayer of Au NPs deposited on bulk pyrolytic graphite substrate. The SIs were obtained in the conventional reflection direction. Spectrum (b) is from Au NPs on 3–5L graphene bombarded with C602+. The surface coverage of Au NPs on substrates was similar in both cases (∼30%). It should be noted that the effective yield of C12H25SO3 is ∼4 times higher in the transmission mode than that in the reflection mode (Ye = 0.62% vs 0.16%). In the reflection mode, most of the SIs are from direct impacts of C602+ on Au NPs, which leads to a higher yield of Au atomic ions (Ye = 4.2% in the reflection mode and 1.7% in the transmission mode) and Au adducts in the reflection direction. Given the thick substrate, the ejecta result from a high density collision cascade.11 In the transmission mode, the SIs are from grazing impacts, favoring the emission of fragments and molecular ions from the dodecanethiol layer. However, overlapping collision cascades cannot develop in graphene, yet the effective yield of the molecular ion is higher in the transmission mode. The possible mechanism(s) are discussed below. The comparison shows that transmission SIMS is more suitable for the characterization of molecular ions attached to NPs than conventional reflection SIMS.

Fig. 6.

(Color online) Negative mass spectra of the 5 nm Au NPs deposited on (a) bulk pyrolytic graphite measured in the reflection mode and (b) 3–5L graphene measured in the transmission mode, bombarded with 50 keV C602+ (peak height is normalized by the number of projectile impacts). See the supplementary material for full spectra (Ref. 10).

Fig. 6.

(Color online) Negative mass spectra of the 5 nm Au NPs deposited on (a) bulk pyrolytic graphite measured in the reflection mode and (b) 3–5L graphene measured in the transmission mode, bombarded with 50 keV C602+ (peak height is normalized by the number of projectile impacts). See the supplementary material for full spectra (Ref. 10).

Close modal

A comparison of the spectra obtained with different projectiles is shown in Fig. 7. Spectrum (a) is from 520 keV Au4004+ bombardment and spectrum (b) is from 50 keV C602+ bombardment. The two spectra contain similar peaks: Au, Au2 (not shown), Au adduct ions, and the ions from dodecanethiol (C12H25S, C12H25SO3, C12H25SO4, etc.). It should be noted that the y-axis scales on the two spectra are different. The yields of SIs from Au4004+ bombardment are ∼3 times higher than the yields of the same SIs from C602+ bombardment. For instance, the dodecanethiol molecular ion peak at m/z 249 has a Ye of 2.0% from Au4004+ bombardment and a Ye of 0.62% from C602+ bombardment. The high Ye of Au from Au4004+ bombardment compared to that from C602+ bombardment (10.0% vs 1.7%) is because part of the Au is from the Au4004+ projectiles. The comparison of the effective yields for all cases discussed above is listed in Table I.

Fig. 7.

(Color online) Negative mass spectra of the 5 nm Au NPs deposited on 3–5L graphene in transmission bombarded with (a) 50 keV C602+ and (b) 520 keV Au4004+ (peak height is normalized to total events).

Fig. 7.

(Color online) Negative mass spectra of the 5 nm Au NPs deposited on 3–5L graphene in transmission bombarded with (a) 50 keV C602+ and (b) 520 keV Au4004+ (peak height is normalized to total events).

Close modal
Table I.

Effective yields of Au and C12H25SO3 of Au NPs coincidental with SH.

C602+ bombardmentAu4004+ bombardment
GFa, TbNPFc, TGraphite, RdGF, T
Au 1.7% 1.8% 4.2% 10.0% 
C12H25SO3 0.62% 0.15% 0.16% 2.0% 
C602+ bombardmentAu4004+ bombardment
GFa, TbNPFc, TGraphite, RdGF, T
Au 1.7% 1.8% 4.2% 10.0% 
C12H25SO3 0.62% 0.15% 0.16% 2.0% 
a

The projectiles impact the graphene first (GF), and then the Au NPs.

b

In the transmission mode.

c

The projectiles impact nanoparticles first (NPF), and then the graphene.

d

In the reflection mode.

A question that arises is that of the mechanism(s) of ejecta emission and ionization, given that the dimensions of the NPs are not sufficient for complete projectile energy deposition.12 The SIs originate either from the support, the Au NP, or its self-assembled layer of dodecanethiol. During the impact of C602+ on graphene, the projectile is shattered and atomized via atom–atom collisions. The ejected carbon atoms from the projectile and knocked-on carbon atoms from graphene then interact with the Au NPs in the transmission direction.8 In contrast, when Au4004+ impacts on graphene the projectile is not shattered but penetrates through graphene and interacts with the AuNPs.13 The Au–Au collision is more efficient for kinetic energy transfer than a C-Au collision. The projectile impact parameter plays a role:8 in the case of C602+ bombardment, the SI signals are from grazing impacts, while in the case of Au4004+ bombardment, the SI signals can be obtained from both grazing and direct impacts on the Au NPs.

We present a characterization of the graphene film alone and the Au NPs deposited on the graphene substrate. Graphene is a promising ultrathin substrate for the analysis of NPs in transmission SIMS since it does not interfere with peaks from the NPs above m/z 120. Transmission SIMS readily indicates on which side of the graphene film the analyte is deposited. Compared to the conventional reflection SIMS on bulk support, the transmission SIMS provides ∼4 times higher yield for the molecular ion attached to 5 nm NPs. Comparing projectiles in transmission, the yields for the molecular moiety are ∼3× higher in Au4004+ bombardment at 1.3 keV/atom than in C602+ bombardment at 0.83 keV/atom. It is important to recall that the respective mechanisms of projectile-graphene/NP interactions are fundamentally different.8,13 A final caveat for reproducible transmissions experiments is the requirement of a well-defined projectile-target geometry.

This work was supported by the National Science Foundation Grant No. CHE-1308312.

1.
S. C. S.
Lai
,
P. V.
Dudin
,
J. V.
Macpherson
, and
P. R.
Unwin
,
J. Am. Chem. Soc.
133
,
10744
(
2011
).
2.
M. A.
Park
,
K. A.
Gibson
,
L.
Quinones
, and
E. A.
Schweikert
,
Science
248
,
988
(
1990
).
3.
Y.
Huang
,
J.
Wu
, and
K.
Hwang
,
Phys. Rev. B
74
,
245413
(
2006
).
4.
J. D.
Debord
,
S.
Della-Negra
,
F. A.
Fernandez-Lima
,
S. V.
Verkhoturov
, and
E. A.
Schweikert
,
J. Phys. Chem. C
116
,
8138
(
2012
).
5.
M. J.
Eller
,
S. V.
Verkhoturov
,
S.
Della-Negra
, and
E. A.
Schweikert
,
Rev. Sci. Instrum.
84
,
103706
(
2013
).
6.
S. V.
Verkhoturov
,
M. J.
Eller
,
R. D.
Rickman
,
S.
Della-Negra
, and
E. A.
Schweikert
,
J. Phys. Chem. C
114
,
5637
(
2010
).
7.
L.-J.
Chen
,
J. H.
Seo
,
M. J.
Eller
,
S. V.
Verkhoturov
,
S. S.
Shah
,
A.
Revzin
, and
E. A.
Schweikert
,
Anal. Chem.
83
,
7173
(
2011
).
8.
S. V.
Verkhoturov
,
S.
Geng
,
B.
Czerwinski
,
A. E.
Young
,
A.
Delcorte
, and
E. A.
Schweikert
,
J. Chem. Phys.
143
,
164302
(
2015
).
10.
See supplementary material at http://dx.doi.org/10.1116/1.4943027 for negative spectra of a sub-monolayer of 5 nm AuNPs deposited on (a) bulk pyrolytic graphite substrate and (b) 3‐5L graphene bombarded with 50 keV C602+.
11.
P.
Sigmund
,
Nucl. Instrum. Methods Phys. Res., Sect. B
27
,
1
(
1987
).
12.
C.-K.
Liang
,
S. V.
Verkhoturov
,
L.-J.
Chen
, and
E. A.
Schweikert
,
Int. J. Mass. Spectrom.
334
,
43
(
2013
).
13.
M. J.
Eller
,
C.-K.
Liang
,
S.
Della-Negra
,
A. B.
Clubb
,
H.
Kim
,
A. E.
Young
, and
E. A.
Schweikert
,
J. Chem. Phys.
142
,
044308
(
2015
).

Supplementary Material