Article BULLETIN OF THE

DOI: 10.1002/bkcs.11875 N. Joo and H. B. Lim KOREAN CHEMICAL SOCIETY

Density Determination of Au Nanoparticles Using Single Particle

ICP-MS
Nayoung Joo and H. B. Lim*

Department of Chemistry, Dankook University, Chungnam 31116, South Korea.

*E-mail: [email protected]
Received April 29, 2019, Accepted August 20, 2019

In this work, the densities of Au nanoparticles (NPs) were determined by single particle (sp) inductively
coupled plasma-mass spectrometry (ICP-MS). Single particle measurement at low concentration with a
dwell time of 100 μs showed good agreement with the size measured by scanning and transmission elec-
tron microscopy (SEM and TEM). The size of Au nanoparticles up to 100 nm showed linear function with
the measured counts, while 250 and 400 nm particles required signal suppression due to signal saturation.
The density of Au NPs was estimated, based on the particle size measured using TEM. From four
repeated measurement sets, the averaged densities of 50 and 100 nm Au NPs were determined to be 21.9
(
18.3%) and 20.8 (
11.0%) g/cm3, respectively, which were slightly larger than the bulk density of Au,
19.3 g/cm3.

Introduction In this work, Au NPs in the size range of 20 to 400 nm

were characterized by determining their particle size and
The analysis of nanoparticles has become an important density using low concentration sp-ICP-MS. Prior to the
issue in various industrial fields, such as environment, food, characterization, the following experimental parameters
biology, and electronics. In particular, the particle size dis- were estimated: the nebulization and transportation effi-
tribution (PSD) has been one of the major concerns for the ciency of NPs to the plasma, the dissociation and ionization
analysis, which has mostly been conducted by dynamic efficiency including the extraction and transmission of the
light scattering or electron imaging tools like scanning dissociated ion, and its conversion factor to produce a sig-
electron microscopy (SEM) or transmission electron nal in mass spectrometer. Although these parameters can be
microscopy (TEM).1 Single particle inductively coupled determined experimentally, factors such as dissociation effi-
plasma-mass spectrometry (sp-ICP-MS) has often been ciency still needed to be assumed for the calculation. Since
used for the measurement because it can provide additional no specific method to determine NP density is yet available,
information on chemical components.2–5 A PSD histogram a size convergence method using low concentration sp-
is developed from a time-resolved spectrum by correlating ICP-MS was demonstrated for Au NPs, of which the den-
the signal of NPs with that of a standard solution.6 sity was derived by converging the obtained particle size
Although sp-ICP-MS enables flexibility in application, it with that of reference size, i.e., SEM or TEM.16 So far, the
density of Au NPs was regarded as 19.3 g/mL of bulk Au
retains problems in attaining high accuracy because of the
with a crystal structure of face centered cubic (fcc).17
uncertainty of the parameters involved in the calculation.
Since the particle size was estimated based on the concept
Experimental
that the signal counts represent the number of atoms of
each particle, the particle density, shape, and instrumental For the PSD measurement, low concentration sp-ICP-MS
parameters need to be characterized prior to the analysis. was used with a quadrupole ICP-MS, produced by Agilent
Those parameters are important for measurement reliability; (model 7900, Santa Clara, CA, USA), a PFA nebulizer
for example, the transportation and dissociation efficiency (PFA-200, Elemental Scientific Inc., Omaha, NE, USA)
might be known for the calculation, which is critically and a PFA spray chamber (Elemental Scientific Inc). The
dependent on the instrumentation and NP properties. The detailed operating condition of the ICP-MS system is listed
density of NPs is another potentially erroneous parameter in Table S1 (Supporting Information). Standard Au solution
in sp-ICP-MS, because it has generally been substituted by of 10 000 μg/mL was purchased from AccuStandard, Inc.
the reference value of bulk solid materials with similar or (New Haven, CT, USA). Au NPs in the size range of
the same chemical composition. Currently, almost no infor- 20–400 nm were purchased from Sigma-Aldrich (Yongin,
mation on NP density is available in the literature, despite Korea). The solutions diluted with de-ionized water
its importance for the characterization of NPs in the fields (Millipore-Q, Burlington, MA, USA) were introduced into
of environmental,7–9 food, and clinical analysis.10–15 the system by a natural flow mode and the isotope signal of

Bull. Korean Chem. Soc. 2019 © 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Wiley Online Library 1
Article BULLETIN OF THE
ISSN (Print) 0253-2964 | (Online) 1229-5949 KOREAN CHEMICAL SOCIETY

Au was measured at the mass-to-charge ratio (m/z) of 197. efficiency of the particles remains unknown, the obtained
The time-resolved spectrum was obtained at the dwell time nebulization and transportation efficiency would be accu-
of 100 μs using a low particle concentration. A lab-made rate because the introduced NPs were rarely overlapped
program using the excel software was used for the calcula- since the particles were well separated at the low concentra-
tions. The particle sizes of Au NPs were also measured tions. In addition, the particles had a greater chance to stay
using SEM (Hitachi S4300, Tokyo, Japan) and TEM in the central channel of the plasma, which made the peaks
(LIBRA 120, Carl Zeiss, Germany) at a 120 kV source. identified clearly. From the result, a significant proportion
of the NPs, i.e., about 98.2% to 98.3%, was lost during the
nebulization and transportation to the plasma.
Results and Discussion Particle Size Measurement by SEM. For reference parti-
cles, the measured sizes of about 100 NPs were averaged to
Determination of System Parameters: Nebulization, be 23.32
3.05, 52.38
4.50, 104.08
8.72, 222.9

Transportation, and Conversion Efficiencies of ICP-MS. 15.06, and 409.18
36.31 nm, for 20, 50, 100, 250, and
For the characterization of Au NPs, the size range of 400 nm (Figure 1). The averaged sizes with standard devia-
20–400 nm was selected. First, instrumental parameters of tions were slightly larger than expected because the parti-
sp-ICP-MS were determined prior to the characterization, cles were coated before taking the SEM images.
such as transmission efficiency of ions in ICP-MS, sample Noticeably, the NPs had sharp edges of a (1,1,1) bi-
uptake rate, and the nebulization, transport, and dissociation pentagonal structure, not a round shape, which implied that
efficiencies of NPs, etc. For this, a time-resolved spectrum the PSD calculations inevitably have a chance to produce
was obtained using a low particle concentration adjusted to erroneous results in size measurement. TEM images were
about 1.2 × 108 particles/L after sonication for 30 s. Since also taken for high accuracy17 and the results are shown in
the inner diameter of the capillary of the PFA-200 nebulizer Figure 2. From the images, the averaged sizes were 19.3,
was about 0.3 mm, the diluted Au NPs were introduced 48.6, 94.8, 233.5, and 387.2 nm, for the sizes of 20, 50,
without clogging. The sample uptake rate of 282.2 μL/min 100, 250, and 400 nm, respectively, with the standard devi-
was experimentally determined at 19  C. From the obtained ation of 10.0%–11.3%. Those were used as reference diam-
time-resolved spectrum, the nebulization and transport effi- eters (dr) throughout this work. Generally, the sizes of
ciency was estimated by counting the number of peaks with TEM were slightly smaller than those of SEM.
respect to the number of particles introduced into the nebu- Particle Size Measurement by Sp-ICP-MS. Next, a time-
lizer.18 The number concentration of the NPs given by the resolved spectrum and PSD histogram at 3 ms dwell time
manufacturer was confirmed from the SEM images, for were obtained by sp-ICP-MS for comparison. The averaged
which the number of particles introduced during the acqui- nominal particle sizes of four replicate measurements at the
sition time of 3 ms was estimated by considering sample optimized conditions were estimated to be 18.9
2.8,
uptake rate and water density of 0.99843 g/mL. The aver- 46.8
7.9, 78.9
19.2, 128.5
48.0, and 134.6

aged peaks appearing in a spectrum were in the range of 63.4 nm for the corresponding Au NPs. Among them, the
45–900 during 3 ms when 4900 to 53 200 NPs were intro- particle sizes of 20 to 100 nm were close to the averaged
duced. Since each Au particle contains enough atoms to size (dr) of TEM, but they became smaller than dr as the
produce a single peak, the number of observed peaks repre- size increased over 100 nm. For confirmation, additional
sents the number of particles reached at the plasma. As study was performed, i.e., the sizes were measured at the
shown in Table 1, the efficiency of the nebulization and dwell time of 100 μs after dilution, and the particle sizes
transportation to the plasma was measured in the range of were estimated from the integrated count of each peak rep-
1.69% to 1.84%, with a relative standard deviation of resenting the number of atoms. Since this type of single
1.76% for five replicate measurements, regardless of the peak conversion may produce a large error, the measure-
particle size. Noticeably, even though the dissociation ment was repeated and averaged to reduce the random error
by ensemble averaging. About 200 peaks were converted
Table 1. Transport efficiency of au nanoparticles (NPs) in low for the calculation throughout this work. The obtained pro-
concentration sp-ICP-MS with PFA-200 nebulizer. file of each peak was composed of about 9–11 data points,
as shown in Figure S1.
Transport
Additionally, the signal conversion efficiency of the MS
Size of Number of Number of efficiency
Au particles particles (%) = (B/A) system, i.e., the rate of signal counts produced by ions pres-
NPs introduced (A) detected (B) × 100 ented in the plasma, can be estimated, assuming the com-
plete dissociation and ionization of NPs. For example, each
20 nm 8400 150 1.79%
100 nm Au particle contains about 3.09 × 107 atoms to be
50 nm 4900 90 1.84%
dissociated into ions. Since the integrated count of each
100 nm 53 200 900 1.69%
peak was in the range of 3.05–3.09 × 103, the extraction
250 nm 9940 180 1.81%
and conversion efficiency of the ICP-MS system was esti-
400 nm 2660 45 1.69%
mated to be about 0.0099% (
0.0001), as shown in

Bull. Korean Chem. Soc. 2019 © 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.bkcs.wiley-vch.de 2
 BULLETIN OF THE
Article Density Determination Using Single Particle ICP-MS KOREAN CHEMICAL SOCIETY

Figure 1. SEM images of au nanoparticles (NPs): Particles sizes measured with standard deviations were (a) 23.32
3.05 nm,
(b) 52.38
4.50 nm, (c) 104.08
8.72 nm, (d) 222.9
15.06 nm, and (e) 409.18
36.31 nm for 20, 50, 100, 250, and 400 nm,
respectively.

difference increased with increasing size. This large devia-

tion for 250 and 400 nm NPs was attributed to incomplete
dissociation and/or signal saturation of the detector. For
closer examination, the averaged profiles of 25 peaks are
shown in Figure 4. Noticeably, Gaussian profiles are clearly
apparent for the NPs of ≤100 nm. But, the top of the peak
became erratic and relatively flat as the particle size
increased over 100 nm. This flat-top peak appearance was
further investigated by changing the sensitivity of ICP-
MS.19 Since the ICP-MS system generally retains a cou-
nting limitation, normally approximately 2 ns, due to a
pulse dead time, the integrated counts of each peak tended
Figure 2. TEM images and determined sizes of au nanoparticles
to be smaller than the counting limit. Since the data point
(NPs): (a) 18.4 nm for 20 nm and (b) 48.3 nm for 50 nm. The par-
of each peak obtained in this work was about 10–11 for
ticles generally had sharp edges rather than a round shape.
100 μs dwell time, each NP contained about
10 datapoints/s. Therefore, the integrated counts of the par-
Table S2. The efficiency was slightly increased to 0.0109% ticles larger than 100 nm may have been in the range of
for 50 nm and even higher than that at 0.0112% for 20 nm. 5–7 × 108 cps, which was close to the counting limit of the
This slight efficiency increased with decreasing particle size detector, causing the flat-top peaks. For further confirmation,
maybe resulted from the difference in the distribution and the sensitivity was reduced by adjusting the extraction voltage
extraction of ions, i.e., ions produced from a large particle of the ion lens from 0 to −89 V (Table S1). At this condition,
distributed in a relatively large area, which reduced the the time-resolved profile was changed to a Gaussian shape, as
extraction through the interface, sampler, and skimmer shown in Figure 4(f). At this reduced sensitivity, the measured
cones, compared to a small one. Since the difference was particle sizes were changed to 229.8 nm (
15.0) and
not significant, the extraction and conversion efficiency of the 316.2 nm (
29.0) for the particles of 250 nm and 400 nm,
system was regarded as about 0.01% in this work. Although respectively, as shown in Figure 3(c). In fact, the size gap was
those efficiencies were indispensable for the estimation of narrowed, but still remained smaller than the dr of TEM at this
PSD in sp-ICP-MS, it was difficult to measure them accu- condition. This result implied that the appearance of flat-top
rately using aqueous standard solutions. Fortunately, for NPs, peaks could be definitively attributed to the signal saturation
they can be determined more accurately because particles of the detector, although the possibility of incomplete dissoci-
appeared as peaks once reached at the plasma. ation cannot be completely ruled out.
From the time-resolved spectrum of sp-ICP-MS
(Figure S3), the particle sizes were estimated and plotted Estimation of Nanoparticle (NP) Density. For further
against the corresponding dr obtained from TEM, as shown characterization, the density, which directly influences the
in Figure 3. From the figures, the obtained sizes for parti- particle size, was considered. Au NPs are generally produced
cles of >100 nm were smaller than expected and the in a liquid chemical method by the reduction of chloroauric

Bull. Korean Chem. Soc. 2019 © 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.bkcs.wiley-vch.de 3
 BULLETIN OF THE
Article Density Determination Using Single Particle ICP-MS KOREAN CHEMICAL SOCIETY

Figure 3. Comparison of the particle size of au nanoparticles (NPs) measured by sp-ICP-MS at the dwell time of 3 MS (a), 100 μs (b),
and 100 μs at reduced sensitivity (c).

acid (H[AuCl4]) to neutral gold atoms. As the solution  3

4 ds
becomes supersaturated, gold gradually starts to precipitate in C m = πρ N np ð1Þ
3 2
the form of sub-nanometer particles. The remaining gold
atoms stick to the existing particles, producing an fcc structure where Cm is the elemental mass concentration and ρ the
with fairly uniform size. Noticeably, the fcc structure did not density of NPs. From the equation, the parameters influenc-
exhibit a round shape in this case, as shown in Figure 1, ing ds are Cm, the density, ρ, and Nnp. The particle number
which can induce a measurement error in sp-ICP-MS. Gener- concentration, Nnp, can be determined experimentally, con-
ally, the density of Au nanoparticle has been regarded as sidering the dilution factor, the sample uptake rate, and the
19.3 g/mL, i.e., that of bulk Au solid material. measurement dwell time. Then, Cm of the introduced NPs
In this work, the density of the NPs was determined by a can be calculated from the experimentally obtained signal
size convergence technique designed to be suitable for intensity, R, which is a product function of Cm, the nebuli-
unknown NPs. It approximately matched the density as the zation (Kneb) and transportation efficiency (Ktrp), and the
particle size measured by this low concentration sp-ICP- conversion efficiency of ICP-MS (Kconv), i.e.,
MS, ds, with the reference size, dr, measured by TEM. For
the measurement of ds, a peak conversion method was used
R
from the spectrum, of which the integrated counts of a Cm = ð2Þ
peak, R, in a low particle number concentration (Nnp) solu- K neb K trp K conv
tion were converted to the particle sizes associated with the
calibration of a standard solution. For the calculation, the Therefore, the density can be calculated from the follow-
following equation was used:2 ing equation;

Bull. Korean Chem. Soc. 2019 © 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.bkcs.wiley-vch.de 4
 BULLETIN OF THE
Article Density Determination Using Single Particle ICP-MS KOREAN CHEMICAL SOCIETY

Figure 4. Peak profiles of au nanoparticle (NPs) measured at 100 μs of LC-sp-ICP-MS: (a) 20 nm, (b) 50 nm, (c) 100 nm, (d) 250 nm,
and (e) 400 nm at normal sensitivity and (f) 400 nm at reduced sensitivity.

6R for the corresponding sizes of 50 and 100 nm Au NPs.

ρ= ð3Þ
πds K neb K trp K conv N np
3 Noticeably, the densities were larger than expected, indicat-
ing that more fine tuning of parameters and experimental
The efficiencies of Kneb × Ktrp and Kconv for Au NPs conditions were required. Secondly, the measured density
were 1.76% and 0.01%, respectively. Noticeably, once a was reduced with increasing particle size. As discussed pre-
particle reached the plasma, it should have produced a peak viously, the Au NPs were not perfectly round, but had the
if the size was large enough to be detected. Thus, this sin- edges of a (1,1,1) bi-pentagonal structure, which might
gle peak conversion by sp-ICP-MS can ignore the effi- induce errors in the calculation of size and density. The
ciency of Kneb × Ktrp. Since a known number of particles usefulness of the proposed method was its flexibility in
were introduced into the ICP-MS with the PFA nebulizer application to various unknown NPs so long as the refer-
system, the nebulization and transport efficiencies were use- ence particle size can be obtained. In conclusion, despite
ful to estimate the elemental concentration. Therefore, from some accuracy limitations remained for multi-structure
Equation (3), if ds was known for a certain particle, the NPs, it can produce the measurement of densities and sizes
unknown density can be determined, and vice versa. of various NPs using low concentration sp-ICP-MS.
The convergence method proposed herein for density
estimation commenced from the experimental measurement Conclusion
of R, using the reference density of 19.3 g/cm3, ρr, from
which the particle size of the sample, ds, was determined. If A characterization method using low concentration sp-ICP-
jds−drj = 0, ρ1 will be equal to ρr. If jds−drj > 0, ρ1 will be MS to determine the size and density of NPs was demon-
changed to an approximate density, ρ2, and the process strated for Au NPs of 20, 50 and 100 nm. The measured
should be repeated until satisfying jds−drj = 0. Eventually, densities were slightly larger than the expected values but
the density will approach the target density, thereby satisfy- this difference decreased with increasing particle size. The
ing the condition. Therefore, the target density, ρ, can be characterization revealed that the particle size of larger than
directly obtained from Equation (3) by substituting ds with 100 nm required instrumental adjustment for accurate mea-
dr. Based on the convergence, the densities of the Au NPs surement because of instrumental limitation. In addition,
of 50 and 100 nm were measured at the optimized condi- the instrumental parameters for NPs, such as nebulization,
tions. Since the particles of 250 nm and 400 nm showed a transport, extraction, and pulse conversion in the ICP-MS
large deviation from the TEM observation, those particles system, could be determined clearly and easily. Despite a
were ruled out for the demonstration. From four repeated few limitations, low concentration sp-ICP-MS showed a
sets of the measurement, the averaged densities of ρ were great potential to be used as a tool to determine particle
obtained to be 21.9 (
18.3%) and 20.8 (
11.0%) g/cm3 density for various unknown NPs.

Bull. Korean Chem. Soc. 2019 © 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.bkcs.wiley-vch.de 5
 BULLETIN OF THE
Article Density Determination Using Single Particle ICP-MS KOREAN CHEMICAL SOCIETY

Acknowledgments. The authors thank Prof. R. S. Houk of 7. F. Gottschalk, T. Sonderer, R. Scholz, B. Nowack, Environ.
Iowa State University-Ames Laboratory and Dr. Karen Sci. Technol. 2009, 43, 9216.
E. Murphy of NIST for discussion and comments on the sin- 8. B. Nowack, Nano Today 2009, 4, 11.
gle peak profiles. This work was supported by the National 9. N. Mueller, B. Nowack, Environ. Sci. Technol. 2008,
42, 4447.
Research Foundation of Korea (NRF) grant funded by the
10. R. J. B. Peters, Z. H. Rivera, G. van Bemmel,
Korea government (MSIT) (No.2018R1D1A1B07048110).
H. J. P. Marvin, S. Weigel, H. Bouwmeester, Anal. Bioanal.
Supporting Information. Additional supporting informa- Chem. 2014, 406, 3875.
tion is available in the online version of this article. 11. K. Chaloupka, Y. Malam, A. M. Seifalian, Trends Biotechnol.
2010, 28, 580.
12. M. R. Monroe, G. G. Daaboul, A. Tuysuzoglu, C. A. Lopez,
References F. F. Little, M. S. Unlu, Anal. Chem. 2013, 85, 3698.
13. S. Egger, R. Lehmann, M. Height, M. Loessner,
1. 1B. Rice, C. Chan, S. C. Brown, P. Eschbach, L. Han, M. Schuppler, Appl. Environ. Microbiol. 2009, 75, 2973.
D. S. Ensor, A. B. Stefaniak, J. Bonevich, A. E. Vlad’ar, 14. D. M. Mitrano, A. Barber, A. Bednar, P. Westerhoff, C. P. Higgins,
A. R. HightWalker, J. Zheng, C. Starnes, A. Stromberg, J. F. Ranville, J. Anal. At. Spectrom. 2012, 27, 1131.
J. Ye, E. A. Grulke, Metrologia 2013, 50, 663. 15. S. Hu, R. Liu, S. Zhang, Z. Huang, Z. Xing, X. Zhang,
2. F. Laborda, E. Bolea, J. Jimenez-Lamana, Anal. Chem. 2014, J. Am. Soc. Mass Spectrom. 2009, 20, 1096.
86, 2270. 16. J. –. H. Shim, S. –. C. Lee, B. –. J. Lee, J. –. Y. Suh,
3. C. Degueldre, P.-Y. Favarger, S. Wold, Anal. Chim. Acta Y. W. Cho, J. Cryst. Growth 2003, 250, 558.
2006, 555, 263. 17. K. Loeschner, J. Navratilova, C. Købler, K. Mølhave,
4. C. Degueldre, P.-Y. Favarger, R. Rosse, S. Wold, Talanta S. Wagner, F. von der Kammer, E. H. Larsen, Anal. Bioanal.
2006, 68, 623. Chem. 2013, 405, 8185.
5. B. Franze, I. Strenge, C. Engelhard, J. Anal. At. Spectrom. 18. H. E. Pace, N. J. Rogers, C. Jarolimek, V. A. Coleman,
2012, 27, 1074. C. P. Higgins, J. F. Ranville, Anal. Chem. 2011, 83, 9361.
6. J. Tuoriniemi, G. Cornelis, M. Hassellov, Anal. Chem. 2012, 19. J. Liu, K. E. Murphy, R. I. MacCuspie, M. R. Winchester,
84, 3965. Anal. Chem. 2014, 86, 3405.

Bull. Korean Chem. Soc. 2019 © 2019 Korean Chemical Society, Seoul & Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.bkcs.wiley-vch.de 6