rights / license: research collection in copyright - non ... · 2 filtration is an effective method...

Research Collection

Doctoral Thesis

Physical characterizations of carbon nanotubes for the emissioncontrol and exposure modeling

Author(s): Bahk, Yeon Kyoung

Publication Date: 2015

Permanent Link: https://doi.org/10.3929/ethz-a-010603324

Rights / License: In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.

ETH Library

https://doi.org/10.3929/ethz-a-010603324

http://rightsstatements.org/page/InC-NC/1.0/

https://www.research-collection.ethz.ch

https://www.research-collection.ethz.ch/terms-of-use

DISS. ETH NO. 23055

Physical characterizations of carbon nanotubes for the emission control and exposure modeling

A thesis submitted to attain the degree of

DOCTOR OF SCIENCES of ETH ZURICH

(Dr. sc. ETH Zurich)

presented by

Yeon Kyoung Bahk

M.Sc., Pusan National University

born on 26.05.1980

citizen of Republic of Korea

accepted on the recommendation of

Prof. Jing Wang (ETH Zurich)

Prof. Sotiris E. Pratsinis (ETH Zurich)

Prof. Alfred Weber (TU Clausthal)

Prof. Jasmin Aghassi (KIT)

2015

1

Abstract

Carbon nanotubes (CNTs) are one of the most studied nanomaterials in the last two

decades, because of their special mechanical, electrical, thermal and optical properties. On the

other hand, their potential impacts on human beings and the environment have been heavily

considered. A number of literatures showed that different CNT’s properties, such as a shape,

size, length, chemistry, surface functionalization and agglomerate status, can influence on its

toxicity. Therefore, precise characterizations of CNTs are important to study their toxicities,

which have not been fully understood. In the present study, the efficacy of current

characterization methods for CNTs was investigated in order to measure the geometrical

length and effective density of CNTs.

To characterize CNTs, a good understanding of CNT dispersion in the liquid

suspension is crucial, because agglomeration states of CNTs in the suspension affect their

properties. Therefore, the relationship between concentration and agglomeration status of

CNTs in liquid suspensions was investigated and a hydrogen bond effect on the enhancement

of CNT dispersion in the liquid suspension was discovered. The demonstration using

polystyrene latex (PSL) particles showed that the discovered hydrogen bond effects by the

acetic environment can also be applied to other nanoparticles with the carboxyl functional

groups.

The effective densities for airborne CNTs and a fractal model to establish the

relationship between CNT properties were obtained through the differential mobility analyzer

–centrifugal particle mass analyzer (DMA-CPMA) tandem measurement method. The model,

which developed to compute the geometrical outer diameters and fractal dimensions of the

agglomerates, agreed well with the results by the scanning electron microscopy (SEM) image

analysis and literature data. The model can be used for monitoring CNT emissions and

transport at workplaces and also provide useful data for toxicity studies of airborne CNTs.

2

Filtration is an effective method to control nanoparticle emissions and more efficient

for elongated particles, because the deposition due to interception is more effective with

cylindrical particles which possess longer effective interception lengths than compact sphere-

like particles. Different filters such as metal screens, nanofibers and Nuclepore filters were

tested against either airborne or liquid borne CNTs.

Obtained physical characteristics were applied to the lung deposition models including

different deposition mechanisms and different particle models. The models for a spherical

particle and fibers were revised and showed reasonable results with given parameters. The

model calculations showed that CNT agglomerate status plays an important role on CNT

deposition in airways into the lung. The modified model can be applied to assess exposure of

CNTs or elongated particles at workplaces.

3

Zusammenfassung

Carbon Nanotubes (CNTs) sind eine der am besten untersuchten Nanomaterialien in den

letzten zwei Jahrzehnten, vor allem aufgrund ihrer besonderen mechanischen, elektrischen,

thermischen und optischen Eigenschaften. Andererseits stehen sie immer häufiger in Verdacht

ihre möglichen Auswirkungen auf Mensch und Umwelt zu haben. Eine große Auswahl an

Literatur zeigt, dass verschiedene CNT Eigenschaften, wie Form, Größe, Länge, Chemie,

Funktionalisierung und Agglomerat-Status, ihre Toxizität beeinflussen. Daher sind präzise

Charakterisierungen von CNTs wichtig um ihre Toxizitäten, die nicht vollständig verstanden

ist, zu studieren. In der vorliegenden Studie wurde die Wirksamkeit von aktuellen

Charakterisierungsmethoden für CNTs, genauer gesagt die Methoden zur Bestimmung der

geometrische Länge und effektive Dichte von CNTs, untersucht.

Zur Charakterisierung CNTs, ist ein gutes Verständnis der CNT-Dispersion von entscheiden

der Bedeutung, weil Eigenschaften der CNTs davon abhängen, ob sie agglomerieren oder

einzeln stehen. Daher wurde die Beziehung zwischen Konzentration und dem agglomerieren

Status von CNT Suspensionen untersucht und eine Wirkung von Wasserstoffbindungen auf

die Verbesserung der CNT Dispersion in flüssigen Suspensionen entdeckt. Die entdeckte

Wasser-stoff-Bindung in einer Essigsäure-Umgebung sind auch anwendbar auf andere

Nanopartikel mit Carboxyl-funktionellen Gruppen, die mit Polystyrollatex (PSL) Teilchen als

Beispiel gezeigt wurde.

Die effektiven Dichten für luftschwebende CNTs und einem Fraktal-Modell, wurde in dieser

Arbeit etabliert. Das entwickelte Modell basierend auf geometrischen Außendurchmesser und

fraktalen Dimensionen stimmten gut mit den Ergebnissen der Rasterelectronenmikroskop

(REM)-Bild-Analyse und Literaturdaten überein. Das Modell kann für die Überwachung der

CNT-Emissionen und ihrem Transport an Arbeitsplätzen eingesetzt werden und kann auch

nützliche Daten für Toxizitäts-studien von CNTs bieten.

4

Filtration ist ein effektives Verfahren zur Verminderung von Nanopartikel Emissionen und

effizienter für die länglichen Teilchen, weil die Abscheidung durch Abfangen wirksamer ist

mit zylindrischen Teilchen, die lange effektive Längen aufweisen. Verschiedene Filter, wie

beispielsweise Metallsiebe, Nanofasern und Nuclepore Filter wurden in dieser Arbeit

entweder mit Luft oder Flüssigkeits- getragen CNTs getestet.

Die erarbeiteten CNT Eigenschaften wurden auf ein Lungen Modell angewendet welches

verschiedener Abscheidungsmechanismen und unterschiedlicher Partikelmodelle einschloss.

Die Modelle für sphärischen Partikeln und Fasern wurden überarbeitet und zeigten

verständige Ergebnisse mit den verwendeten Parametern. Die Modellrechnungen zeigen, dass

CNT Agglomerat Zustand eine wichtige Rolle auf die CNT Ablagerung in dem Atemwegen

in der Lunge spielt. Das modifizierte Modell kann angewendet werden, um die Exposition der

CNTs oder länglichen Partikeln am Arbeitsplatz zu beurteilen.

5

Acknowledgements

I would first like to thank my PhD advisor, Prof. Jing Wang, for invaluable support, guidance and encouragement throughout a long process. Discussions with him were always so fruitful and helpful. I would also like to thank my co-advisor, Prof. Sotiris E. Pratsinis, who gave me great impressions and showed how an active scientist looks like, for kind advice regarding my thesis and carrier. I would like to give special thanks to the committee members, Prof. Alfred Weber and Prof. Jasmin Aghassi, and Chairman, Prof. Irena Hajnsek, for their valuable advice and feedback on my work.

To all members of the APR group: it was a great and unforgettable experience to spend time and do research with you. Thank you to Kiny for a lot of scientific and unscientific chats and kind helps, Panagiota and Xu for a lot of valuable discussions and giving chances for involvements in your research, Beni for hundred times of linguistic supports without complaining, Ari for kind greeting and smiling every morning, Jelena for a great time sharing from the beginning of the lab, Lukas S for keeping the next seat for four years and sharing a lot of useful information, Lukas D for helpful supports regarding instruments, Manos for being the first master student with me and one of my best friends, also thank you to former bachelor and master students, who spent time together with me in APR group.

Special thanks to Korean colleagues in EMPA and EAWAG, Songhak, Wookjin, Jaebong, Junho, Hansang, Pyoungjik, Minju, Soyoun and Nayoung, for interesting scientific discussions and advice, wonderful gathering and trips. I also appreciate my colleagues in Analytical Chemistry Lab. in EMPA helping and supporting during my stay in EMPA. Thank you to Noemi for being my best coffee-break mate, Nobert for kind helps and concerns regarding my normal life, Heinz and Andy for a lot of support from the beginning.

There are many people who trust and encourage me to finish my PhD successively; I appreciate Prof. Go, who was my advisor of master’s study, Prof. Pui and Dr. Chen in University of Minnesota. I would like to thank Donghyun, who is my best friend and reliable collaborator, for putting a lot of effort to fabricate my strange micro devices without complaining. There are many thanks to all my friends and colleagues in ETH Zruich.

Lastly and most importantly, I must thank my lovely family, my wife Soo Yeoun and little girl, La On, and my parents and my brother, who support me mentally, especially for many times of understanding the abnormal situations during my PhD.

6

Contents

1. Introduction

1.1. Author’s contributions to publications included in dissertation

2. Physical characterization I: CNT Dispersion

2.1. Concentration effect on agglomeration of CNTs in aqueous solution

2.2. Acetic effect on the enhancement of dispersion stability of CNTs inaqueous solution

3. Physical characterization II: Geometrical length measurement

3.1. Electron microscopy analysis

3.2. Filtration method

4. Physical characterization III: Effective density measurement and fractalmodel for CNT agglomerate

4.1. Densities of MWCNTs

4.2. Fractal model for CNT agglomerate

5. Emission control: Filtration of air-borne CNTs

6. Exposure modeling: Lung deposition model for air-borne CNTs

7. Conclusions

8. Outlook

Bibliography

Appendix

Curriculum Vitae

············· 7

············· 13

············· 16

············· 18

············· 21

············· 43

············· 44

············· 49

············· 64

············· 65

············· 70

············· 84

············· 95

··········· 117

··········· 121

··········· 123

··········· 148

··········· 159

7

1. Introduction

Carbon nanotubes (CNTs) have been heavily studied for the last few decades, because

of their attractive mechanical, electrical, thermal and optoelectronic properties. Together with

increasing of their fascinating applications, toxicological concerns have also been increased

due to their special geometrical structures. CNTs possess extremely high aspect ratios than

other materials and geometrical similarity with asbestos fibers increased concerns of

inhalation toxic of airborne CNTs. Previous studies have shown not only lung disease

associated with asbestos fibers (1,2), but also long industrial fibers possess potential

pathogenicity (3). Recently several studies also showed asbestos-type pathology associated

with exposure to long and straight CNTs (4-6). Davis et al. (2) have studied the effect of the

geometrical length to pathogenicity of asbestos fiber and showed that long fibers were more

toxic than short fibers with their experimental set-up. A number of studies have found that

inhaled CNTs possess cytotoxicity, inflammatory cell influx and potential to promote lung

tumors (7-13). Furthermore, CNTs can cause immunosuppression and inflammation, and

change molecular signaling in extra pulmonary tissues (14).

The lengths of CNTs have significant effects on agglomeration status and toxicity. Wick et al.

(15) showed CNT agglomerates are more cytotoxic than well dispersed CNTs. Pauluhn (7)

concluded that the toxic effects of multiwalled CNTs (MWCNTs) are determined by density

of agglomerate structures, not fibrillar structures. Murphy et al. (16) examined three CNT

samples of differing lengths and found only the long CNT sample caused acute neutrophilic

inflammation in bronchoalveolar lavage at 1 week. In addition, the length and density of

CNTs also affect their mobility and transport properties. These parameters play critical roles

for prediction of lung deposition due to interception, impaction and gravitational settling,

therefore they are critical in inhalation dosimetry studies (8, 17). Characterization of CNTs is

significant not only for toxicology studies, but also for their functionality. For instance,

8

improved mechanical property of CNT composite is achieved with a higher aspect ratio of

single-walled CNTs (SWCNTs) (18); gas adsorption properties are closely related to fractal

dimension of CNT assemblies (19).

A number of studies have been dedicated to determination of the CNT characteristic

parameters and the length and diameter of CNTs are usually determined by an electron

microscopy. On the other hand, airborne particle measurement systems were used to

determine the length of individual airborne CNTs. For example, Kim and Zachariah (20)

combined an electric mobility measurement and model for cylinder shaped particles in

electrical fields to compute the length of SWCNTs. In present study, the geometrical lengths

of CNTs were determined by the scanning electron microscopy (SEM) analysis and results are

presented in Chapter 3. Additionally, a filtration method is introduced to measure the length

of MWCNTs. The filtration method based on the single fiber filtration theory including

particle capture mechanisms, such as diffusion, interception and impaction (21), and is

applicable to cylinder-like individual CNTs’ real-time characterization.

The literature on CNT density measurements showed substantial discrepancies. Kim et al. (22)

used an aerosol particle mass analyzer (APM) to measure airborne MWCNTs with

consideration of their cylindrical shape and reported the condensed phase density as 1.74

g/cm3. Chen et al. (23) found the effective density of airborne MWCNT agglomerates with

the geometric mean outer diameter and length of 100 nm and 3 µm, respectively, to be 0.71 –

0.88 g/cm3. They determined the effective density by taking the ratio between the

aerodynamic diameter and projected area diameter obtained from electron micrographs. The

bulk density of Baytubes provided by the manufacturer was 0.14 – 0.16 g/cm3 (24) measured

by the EN ISO 60 method, which determines the apparent density of a powder by pouring it to

a cylinder with a known volume. Ma-hock et al. (8) used 0.15 g/cm3 as the apparent density

for their MWCNTs, but the stated value represented effects of the large pores between the

more densely packed agglomerates. They considered the pores less than 1 μm as determined

9

by the mercury porosimetry and reported the density of agglomerated MWCNTs as 0.39

g/cm3. Oberdӧrster (25) pointed out the CNTs usually occur as tangles of hollow tubes with

different shapes, and their effective densities were different from that of solid carbon. The

measurement results of the intrinsic density and effective density of airborne MWCNTs by a

combination of micro-balance and gas pycnometer and the differential mobility analyzer -

Couette centrifugal particle mass analyzer (DMA-CPMA) tandem, respectively, are shown in

Chapter 4.

In order to characterize CNTs, certain degree of control between well dispersed and

agglomerated CNTs is essential, since their physical characteristics including geometrical

sizes and densities vary a lot depending on their agglomeration status in air and liquid. In

addition, their physical characteristics are also important for technical applications (26,27).

Therefore, the dispersion stability of CNTs in suspensions has been widely studied.

Hydrophilic oxygen-containing functionalization by strong oxidizing agents is one of popular

surface modifications to facilitate uniform dispersion of CNTs in the suspensions. For

example, HNO3, HNO3/H2SO4, O3, KMnO4 or H2O2 are commonly used in order to

functionalize CNTs (28-30). Another method to enhance dispersion of CNTs involves the use

of surfactants (31). Several studies investigated enhanced stabilization with natural organic

matter as a surfactant for dispersion of not only CNTs, but also colloids in the aqueous phase

(32-34). Smith et al. (29) investigated the relationship between the colloidal stability and pH

of electrolyte solutions and showed that the stability of acid-treated MWCNTs increases with

increasing pH. For metal-based electrolyte, they also determined the critical coagulation

concentration (CCC) of the electrolyte for acid-treated MWCNTs and the results were

consistent with the Derjaguin Landau Verwey Overbeek (DLVO) theory prediction (35). In

Chapter 2, the investigated relationship between concentration and agglomerate status of

CNTs suspensions and the discovery of a hydrogen bond effect on the enhancement of CNT

dispersion in a liquid suspension are presented.

10

In addition to the density determination in Chapter 3, a fractal model for MWCNT

agglomerates has been developed to establish the relationship between the MWCNT

properties including tube diameters, length, intrinsic density and the agglomerate properties

such as the mass, porosity, effective density, surface area and characteristic diameters. Many

studies demonstrated that a variety of distinctively different forms of CNTs can coexist,

including individual CNTs that are cylinder-like or bended or coiled, CNTs bundles, and

quasi-isometric shaped agglomerates (8,23,36-39). Therefore characterization can hardly be

achieved by using uniform parameters for all the CNT particles even in the same sample.

Models have been developed for cylinder-like CNTs but are lacking for CNT agglomerates.

Thus we performed tandem measurement of the mobility size and mass of airborne MWCNT

agglomerates and demonstrated that the mass increase with size follows a power law,

therefore the MWCNT agglomerates can be described as fractal-like particles. The fractal

geometry has been investigated for bulk CNT assemblies (19,40,41) however, has not been

applied to airborne dispersed CNT agglomerates.

Due to increasing concerns over control the CNT emissions in workplaces or environment,

NIOSH proposed a recommended exposure limit (REL) for CNTs (also for carbon nanofibers,

CNFs) of 1 μg/m3 of respirable elemental carbon as an 8-hour time-weighted average (TWA)

concentration (42). NIOSH suggested that efforts should be made to reduce the CNT exposure

as much as possible because the effects of their shape, size, length, chemistry,

functionalization, etc. on their toxicity have not been fully understood (4-7,43-48). Because of

the potential high toxicity of CNTs (49), Environmental Protection Agency (EPA, USA) has

promulgated Significant New Use Rules (SNURs) for premanufacture notice (PMN)

substances for the multi-walled carbon nanotubes (MWCNTs, PMN P-08-177), single-walled

carbon nanotubes (SWCNTs, PMN P-08-328) and certain multi-walled carbon nanotubes

(certain MWCNTs, PMN P-08-199). In the notices, the EPA proposed a series tests

attempting to reduce their potential serious adverse health effects and environmental impacts

11

(50,51). Now, EPA further promulgated SNURs for functionalized MWCNTs (PMN P-12-44,

52), which could offer an improved dispersion capability for some solution-based

nanotechnology applications such as to be the additive for rubber and batteries (53-55).

Usually, the functionalization of CNTs (including MWCNTs and SWCNTs) is conducted by

attaching certain functional groups (e.g. -COOH and -OH) to their smooth sidewalls and open

ends of tubes via harsh oxidative process. One of the mostly used functionalization methods is

the nitric acid refluxing in which not only functionalization takes place but also purification is

carried out by dissolving and removing the metal catalysts impurities. Such purified CNTs are

also widely used in many nanotechnology chemical and biological sensors (56,57). However,

the functionalized MWCNTs were found to be more toxic than the pristine (raw) MWCNTs

in human T cells toxicity tests (58). Besides, Zhao and Liu (59) indicated that the functional

groups can have a negative impact on cell viability. Therefore, in the new SNUR of the

functionalized MWCNTs, in addition to the relevant toxicity tests (in vivo and in vitro), the

measurements of their size distribution, shape and length of the functionalized MWCNT were

recommended to be the future research focus. Therefore, filtration through different filters

was employed to control CNT emission in air and liquid, because filtration is one of common

methods controlling the nanoparticle emissions and more efficient to the elongated particles

as CNTs. Previous filtration studies on CNTs (60-63) or asbestos fiber (64) indicated that the

major difference between fiber and sphere-like particles was caused by the interception

mechanism. Different filters such as metal screens, nanofiber filters and Nuclepore filters

were tested against airborne and liquid-borne CNTs and the results are shown in Chapter 5.

There are increasing concerns on the precise prediction of CNT deposition in the lung due to

the toxicity reports regarding CNTs. Several studies have investigated CNT deposition in the

human lung by existing models, such as an aerodynamic diameter concept model (65) and

multiple path particle dosimetry (MPPD) model. Sturm (65) employed the theoretical

approximation by the concept of aerodynamic diameters in order to assess CNT deposition in

12

human alveoli and showed CNT depositions under different conditions such as different

exercise levels, aspect ratios of CNTs, etc. Recently, Su and Cheng (66) have estimated

carbon fiber deposition in a human respiratory tract replica and the results showed the local

deposition fraction of CNTs in the airways in the lung. They adopted carbon fibers with

uniform diameter, 3.66 µm, to investigate fiber deposition mechanisms in the human nasal

airway and they presented that impaction was a dominant mechanism for the tested fibers.

Existing studies regarding calculations of CNT deposition commonly assumed a CNT as a

cylindrical particle, which possesses the high aspect ratio and alignment effect on the air flow

direction. However, results in the present study showed the coexistence of CNT agglomerates

and single standing CNTs in the same airborne samples depending on the dispersion status in

the suspension and mobility size of airborne CNTs. Therefore, the relationship between the

fraction of CNT agglomerates and mobility diameter of CNTs was obtained by the SEM

analysis in order to apply the fraction in the lung deposition model. Developed theoretical

models by Yeh and Schum (67), Anjilvel and Asgarian (68), which are used in MPPD model,

and Ding et al. (69) have been revised and applied in the present study to predict CNT

deposition in entire airways to the lung. The results are presented in Chapter 6.

13

1.1 Author’s contributions to publications included in the dissertation

Paper I: Determination of geometrical length of airborne carbon nanotubes by electron

microscopy, model, calculation, and filtration method (Published in Aerosol Science and

Technology, 2013)

In this study, determination of geometrical lengths of airborne CNTs was investigated. A new

approach (called a filtration method) to calculate the CNT length using the combination of

mobility measurement and filtration was introduced. The filtration method provides the

possibility of fast measurement of the geometrical length of elongated particles. In addition,

the generation methods of airborne CNTs from a liquid suspension and method to control the

degree of dispersion by changing the concentration of CNTs in the suspension were

investigated. The author of the dissertation designed the study and performed the entire

experiments and analysis. The author also developed the filtration model and wrote the paper.

The results from this paper are introduced in Chapter 2, 3 and 5.

Paper II: Filtration and length determination of airborne carbon nanotubes in the

submicrometer range using nanofiber filters (Published in Aerosol and Air Quality Research,

2014)

This study investigated the filtration method developed in the previous study using stainless

steel screens. Here nanofiber filters were used instead of the screens. Furthermore, the

theoretical model for nanofiber filter media including the single fiber theory was investigated

against airborne MWCNTs. Nanofiber samples with different solidities and fiber sizes were

tested and the results were compared with the model. All experiments and modelling were

done by the author of the dissertation and paper was also written by this author. The results of

this study are presented in Chapter 3 and 5.

14

Paper III: Carbon nanotube penetration through fiberglass and electret respirator filter and

Nuclepore filter media: experiments and models (Published in Aerosol Science and

Technology, 2014)

Different respirator filters and Nuclepore filter were challenged with airborne MWCNTs. The

model based on the single fiber theory was used to predict the CNT and NaCl penetration

through the filters and study the deposition mechanisms of nanoparticle through the filters.

The filtration model using the Nuclepore filter was investigated and test to confirm the

alignment effect of CNT in the air flow with a high velocity was performed. The author of the

dissertation has contributed to develop the filtration model and produced the CNT samples for

the experiments. Also this author has contributions in discussions for the study, writing the

paper and modelling for the particle filtration. The contents are presented in Chapter 3 and 5.

Paper IV: Characteristics of airborne fractal-like agglomerates of carbon nanotubes

(Published in Carbon, 2015)

We characterized MWCNT agglomerates in order to understand their physical properties. The

effective density and tube length in the agglomerate were determined by mass measurements

and developed fractal model. The model provides geometrical outer diameters and fractal

dimensions of the agglomerates using mobility-mass measurement. It can be used for real-

time monitoring of CNT emissions at workplaces. The author of the dissertation has

contributed in discussions to develop the fractal model for CNT agglomerates and writing the

manuscript, and designed and performed all experiments except the experiments using APM,

which performed by a co-author in UMN. The results are mainly included in Chapter 4.

15

Paper V: Enhanced dispersion stability and mobility in porous media of carboxyl

functionalized carbon nanotubes in aqueous solutions through strong hydrogen bonds

(Published in Journal of Nanoparticle Research, 2015)

The enhancement of the dispersion of MWCNTs by acetic electrolytes was discovered in this

study. The experiments confirming the acetic effects on carboxyl functional groups on

nanoparticles were demonstrated using UV-Vis spectrophotometry, dispersion observations

and aerosolization-quantification method. MWCNTs and PSL particles with and without the

functional group were tested. The author of the dissertation discussed and discovered the

effect, and designed the confirmation study. The experiments were mainly demonstrated and

repeated by the author of dissertation, and results were summarized and the paper was written

by this author. This study is presented in Chapter 2.

Paper VI: Lung deposition model of airborne multi-walled carbon nanotubes for inhalation

exposure assessment (Preparing to submission for publication)

In this study, lung deposition models were revised and CNT agglomerate data was applied to

assess CNT deposition in the lung. Obtained parameters from the previous characterization

study for airborne MWCNTs were applied in these model calculations. The model

calculations showed that CNT agglomerate status plays an important role on CNT deposition

in airways into the lung. Agglomerate CNTs in the entire size range and small single standing

CNTs can reach the deep inside the lung under the low flow rate condition. The author of the

dissertation has done all necessary modelling and modification of existing lung deposition

models. This author prepared the manuscript to publish in a peer reviewed journal. These

results are included in Chapter 6.

16 Bahk et al. (2014) Aerosol Sci. Tech. 47(7):776-734; Bahk et al. (2015) J. Nanopart. Res. 17:396

2. Physical characterization I: CNT Dispersion

Control of dispersion degree of MWCNTs is significant for aerosol research. For

instance, to define the geometrical characteristic of airborne MWCNTs, well dispersed

MWCNTs samples with distinguishable individual MWCNTs are required. For density

measurement of single standing or agglomerated CNTs, control of certain dispersion degree

of CNTs is necessary. Although the CNT functionalization method adopted to lead good

dispersion in the study, certain dispersion degree of CNTs in the suspensions was difficult to

achieve due to the strong adhesion force between CNTs.

The generation methods for airborne CNTs using an atomizer and electrospray

introduced previous studies (60,70-72), and concentration effect on the degree of CNT

dispersion were investigated in the present study (73). Furthermore, enhancement of

dispersion stability of CNT suspension by acetic electrolytes was discovered in the study and

this study showed that acetic solutions, for instance, CH3CO2NH4 and CH3CO2Na, in certain

concentration range had opposite effects to the common electrolytes such as NaCl, KCl, etc.:

they enhanced the dispersion stability of aqueous suspension of COOH functionalized

MWCNTs (COOH-MWCNTs) (74). The reason lies in that the electrolytic acetic CH3COO-

forms a stronger hydrogen bond with carboxyl than the bond between carboxyl and water

molecule (75).

In the DLVO theory, the stability of the suspension depends on the dispersion attractive force

(van der Waals force) and electrostatic repulsion due to the presence of electric double layer

of the charged objects (76). The total free energy between two interacting objects (e.g.

between two nanoparticles or a nanoparticle and plane surface) is the sum of potential energy

due to the electrostatic repulsive force and van der Waals force (77). It should be noted that

most of colloidal interactions take place in aqueous environment and unique characteristics of


water bring non-DLVO interactions, such as hydrogen bond, hydrophobic interactions,

chemical (e.g. steric repulsion, bridging attraction, depletion force) and physical interactions

(e.g. hydrodynamic attraction), etc. The dispersion stability can be strongly affected by the

non-DLVO interactions under specific conditions (78-83). When common electrolytes, such

as NaCl, KCl, etc., are added in the solutions, the electrolytic cations and anions form

hydrated ions due to the strong interaction among water molecules, cations and anions (84-87).

Based on the Lewis acid-base interaction theory, this leads to decreasing hydrophilicity of the

MWCNTs caused by the favorability of interaction between water molecules and electrolyte

ions. Besides, the electrolytes also compress the electrical double layer, which favor the

attraction of CNTs. It is already known that increasing electrolyte concentration in the

suspension leads to decreasing solubility of oxidized CNTs (29,88). On the contrary,

discovered acetic effects lead to the enhancement of dispersion stability.

The influence of acetate on COOH-MWCNTs in the aqueous phase was investigated using

CH3CO2NH4 and CH3CO2Na solutions, and enhanced dispersion stability was observed. The

enhanced dispersion stability also corresponded to higher mobility in porous media, which

was demonstrated by the penetration of COOH-MWCNTs through a model membrane filter.

The mobility and dispersion status are closely interlinked, since penetration of singly

dispersed CNTs is easier than that of aggregated CNTs when sieving, interception, impaction

and gravity settling are the major capturing mechanisms. A series of experiments were

performed to confirm the proposed effects of the electrolytic acetate. The stability of

MWCNTs in KCl solutions was tested and compared with those of acetic samples.

Additionally, polystyrene latex (PSL) particles and carboxyl functionalized PSL (COOH-PSL)

particles were examined in order to compare the acetate effects on the particles, with and

without COOH functional group. The results are expected to be applicable not only for

dispersion of COOH-MWCNTs, but also for other nanoparticles with such a functional group.


2.1. Concentration effect on agglomeration of CNTs in aqueous solution

Materials and methods

The generation method for airborne MWCNTs using a Collison type atomizer and

electrospray (TSI 3480) were investigated. Obtained typical mobility size distributions by two

particle generation methods were presented in the Appendix A1. The multi walled carbon

nanotubes (MWCNTs) from the generators went through a diffusion dryer to ensure that

liquid vapors were removed. In order to classify the generated airborne CNTs, a DMA (TSI

3081) was used. The mobility size distribution was obtained by using the scanning mobility

particle sizer (SMPS). The distributions from different generation methods were compared to

investigate the characteristics of each method. In order to control the degree of CNT

dispersion, the concentration of CNTs in the suspensions was varied and the resultant size

distributions of aerosolized CNTs were measured. MWCNTs with diameters about 15-20 nm

(Baytubes, BMS, Germany), which were functionalized by the nitric acid refluxing method

(57), and about 20-30 nm (Cheaptubes, USA), which possessed a functional group COOH,

were used in the experiments and dispersed in deionized (DI) water.

Dispersion degree of MWCNTs

Figure 1 shows that modes of the size distributions of airborne MWCNTs were changed by

varying the concentration of MWCNTs in the suspension. Smaller mode sizes were obtained

from the suspensions with lower concentrations. The concentration of airborne MWCNTs was

also reduced when the concentration of MWCNTs decreased in the suspension. By varying

the concentration of MWCNTs in the suspension, it was able to generate well dispersed

MWCNTs or agglomerated MWCNTs with certain degree of control. Well dispersed

MWCNTs were obtained with the low concentration suspension and agglomerated MWCNTs

were generated from the high concentration suspension. For Baytubes, well dispersed


MWCNTs and agglomerated MWCNTs were obtained from 0.001 wt% to 0.005 wt% and

from 0.01 wt% to 0.05 wt%, respectively. The dispersion degree was determined from SEM

images of the airborne MWCNTs collected on silicon substrates, as shown in the inset in

Figure 1. Short sampling time was used to ensure that the collected samples had significant

distances among themselves and no apparent overlapping. The agglomerates showed tightly

intertwined structure, which was highly unlikely to be resulted from sample overlapping on

the substrate. Thus we believe the agglomerates observed on the substrates reflected the

agglomeration status of airborne MWCNTs.

Figure 1. Size distributions from suspensions of different concentrations and SEM images

from each mode for MWCNTs from Baytubes.


Results using MWCNTs from Cheaptubes were shown in Figure 2. The smaller size peaks

were from residual particles and larger size peaks were from MWCNTs. When the MWCNT

concentration in the suspension increased, the mode of the MWCNTs distribution were

increased slightly and the concentration for the mode size increased significantly. All the error

bars in Figures 1 and 2 indicated the maximum and minimum values of experimental data.

The residual peak was comparable to the main aerosol peak in the case of MWCNTs from

Cheaptubes, whereas it was much lower than the main peak in the case of MWCNTs from

Baytubes, so that it was not prominent in Figure 1. The difference between number

concentrations of airborne CNTs from the two types of CNTs even with similar mass

concentration in suspensions might be due to the functionalization process. During the

process, geometrical properties of CNTs were changed due to processing conditions. For

example, dependent on processing time of nitric acid oxidation, which is a common method to

oxidize CNTs, CNTs can be damaged (57) and it can be a reason why CNTs from different

providers possess different number concentrations in the specific size range. The airborne

CNT size distribution and agglomeration level were dependent on the dispersion status in the

suspension, which was in turn dependent on the CNT diameter, length, surface properties and

solvent properties. Several kinds of MWCNTs from different manufacturers were checked

and the controlling method for the degree of CNTs dispersion by varying the concentration in

the suspension could not be used when the CNTs were not dispersed in the suspension

properly.


Figure 2. Size distributions from suspensions of different concentrations for MWCNTs from

Cheaptubes.

2.2. Acetic effect on the enhancement of dispersion stability of CNTs in aqueous

solution


In order to investigate the dispersion stability of MWCNT suspensions, UV-Vis

spectrophotometry, dispersity observations documented by a digital camera and

measurements of the particle penetration through the porous media were employed in the

study. Pristine MWCNTs (Baytubes, BMS, Germany) possessing 15-20 nm diameters and

carboxyl functionalized MWCNTs (which are referred to as C-MWCNTs in the following

part) by the nitric acid refluxing method (57) were used to study the electrolyte effect.


Suspensions of C-MWCNTs were prepared with DI water, CH3CO2NH4, CH3CO2Na or KCl

solutions. Different concentrations of ammonium acetate in the solutions were prepared for

dispersity observation and pristine MWCNT samples in DI water and 0.02 moles of

CH3CO2NH4 per liter of solution were also prepared for comparison. The samples were kept

for 14 days to observe aggregation of MWCNTs in the solution and measured by a UV-Vis

spectrophotometer (Cary 50, Varian, USA) every day.

C-MWCNTs dispersed in aqueous CH3CO2NH4 and KCl solutions at different concentrations

were used for the penetration measurements. The Nuclepore filters used in the study were

47mm Whatman®-Track-Etched Polycarbonate Membrane Filters with 3 µm pore diameters,

which were treated as a model membrane filter due to their uniform pore sizes and a flat front

surface. The porosity and thickness of the filter were 0.14 (2×106 pores/cm2) and 0.9 µm,

respectively.

Figure 3 shows the system set-up for the penetration measurement including UV-Vis

spectrometry as well as the aerosolization-quantification method. Suspensions of C-

MWCNTs were driven through the Nuclepore filter by a peristaltic pump, which provided a

stable flow rate of 44 ml/min. The average face velocity was calculated with the effective

surface area of the Nuclepore filter (8.0×10-4 m2) and the value was 9.1×10-4 m/s. C-MWCNT

concentrations upstream and downstream were measured by either the UV-Vis

spectrophotometry or aerosolization-quantification method (89). Ling et al. (89) investigated

the method to evaluate the collection efficiency of liquid-borne nanoparticles by the filter.

They found satisfactory correlations between the concentrations of liquid-borne and air-borne

particles when the liquid-borne particles were aerosolized and measured by the SMPS. The

SMPS delivers particle number vs. size distributions, where the size is the equivalent mobility

size representing the particle mobility in an electrical field.


Figure 3. Experimental system for mobility tests of C-MWCNTs in the aqueous solutions

through the porous media.

The UV-Vis spectrophotometer was calibrated with different concentrations of C-

MWCNTs in the DI water and the results are shown in Figure 4. Stock suspensions of 0.02,

0.015, 0.01, 0.005 wt% C-MWCNTs were prepared and diluted by 5 or 10 times to obtained

lower concentration samples. The samples a, b, c, d, e, f, and g correspond to 0.005, 0.004,

0.003, 0.002, 0.0015, 0.001, 0.0005 wt%, respectively; the sample h is DI water. The samples

were measured by the UV-Vis spectrophotometer and the ranking of the resultant absorbance

was consistent with the expected C-MWCNT concentrations. The curves of the scanning

results for the wave length range from 350 to 900 nm are clearly distinguished for different

samples as shown in Figure 4. The 700 nm wavelength was chosen as the index for the C-

MWCNT concentration in the suspension and Figure 5 demonstrates the proportional

relationship between the concentration and UV absorbance.


Figure 4. Calibration of UV-Vis spectrometer with different concentrations of C-MWCNTs in

the suspensions.

Figure 5 Linear fitting for UV absorbance (arbitrary unit) of different concentrations of C-

MWCNTs in the suspensions (wavelength = 700 nm).


The amount of dispersed C-MWCNTs was measured by the aerosolization-quantification

method, in which the C-MWCNT suspensions were aerosolized by a Collison type atomizer

and measured by the SMPS. Several typical size distributions of airborne C-MWCNTs are

shown in Figure 6. The size of the C-MWCNT in such a measurement represents the diameter

of an equivalent sphere which has the same electrical mobility as the C-MWCNT. The ratio of

the upstream and downstream concentrations gave the penetration through the porous media,

which was computed for different mobility sizes based on the SMPS results. The penetrations

are expressed as:

𝑃𝑃 = 𝐶𝐶𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝐶𝐶𝑢𝑢𝑢𝑢𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑

, (1)

where Cupstream and Cdownstream are particle concentrations measured by SMPS, at the upstream

and downstream of the porous media, respectively. The size range from 100 to 200 nm, where

enough particle concentrations could be measured to calculate the penetration for each

mobility size and differentiate C-MWCNTs from residue particles or electrolytes in water,

was chosen for aerosolized C-MWCNT measurements. On the other hand, the penetration of

each sample was also calculated with the upstream and downstream concentrations of C-

MWCNT suspensions determined by the UV-Vis spectrophotometry, which was an integral

measurement and does not give size specific information.


Figure 6. Number vs. mobility size distributions of C-MWCNTs in the CH3CO2NH4 solution

measured by the SMPS after aerosolization. The size distributions of the C-MWCNT

suspension upstream and downstream of the filter had mode values around 50-55 nm and the

background distribution of the CH3CO2NH4 solution had the mode at 18 nm.

The aerosolization-quantification method may not work for certain electrolytes and particle

suspensions, because the electrolytes, surfactants or impurities in the aerosolized droplets

could condense and form particles in similar size range as the target particles, thus obscuring

the intended particle size distribution. For instance, distributions of C-MWCNTs dispersed in

KCl samples could not be distinguished from the background distributions obtained with the

KCl solutions without any C-MWCNTs as shown in Figure 7. Therefore for KCl-based C-

MWCNT suspensions, we obtained the total penetration of C-MWCNTs only by the UV-Vis

spectrophotometer.


Figure 7. Number vs. mobility size distributions of C-MWCNTs in KCl solution measured by

the SMPS. The intended C-MWCNT distributions cannot be distinguished from the

background distribution of the KCl solution.

PSL particles of different sizes, such as 100, 200 and 300 nm (Magshpere Inc., USA), as well

as COOH-PSL particles, such as 109, 217 and 300 nm (Agilent Technologies, USA), were

used to test the acetate effect on dispersions of nanoparticles with and without carboxylic

functional groups. Proprietary surfactants from the manufacturer existed in the purchased PSL

and COOH-PSL suspensions which facilitated stable dispersion in aqueous environment.

Both UV-Vis spectrophotometry and aerosolization-quantification method were used to

investigate the effect. Whatman®-Track-Etched 47 mm Polycarbonate Membrane Filters with

1 µm pore diameters were used for the penetration experiments.

Dispersion stability of MWCNT suspensions

Dispersion stability observations by a digital camera were shown in Figure 8 in which the

pictures were taken immediately, 24 hours and one week after the solutions had been


sonicated. 0.005 wt% of pristine MWCNTs and C-MWCNTs were dispersed in DI water,

CH3CO2NH4 and CH3CO2Na solutions with different electrolyte concentrations, such as

0.01, 0.02, 0.03, 0.05 and 0.1 M. The lower transparency of the suspension and homogeneous

black solution can be considered as a qualitative indication of better dispersion (90). As

shown in Figure 8, all pristine MWCNT samples aggregated considerably after the sonication

process and quickly settled down to the bottom. On the other hand, C-MWCNT samples in

CH3CO2NH4 and CH3CO2Na solutions showed stable dispersion when they were just

prepared. 24 hours after preparation, the C-MWCNT samples in both 0.1 M CH3CO2NH4 and

CH3CO2Na solutions showed settlements on the bottom, leaving a transparent supernatant

above. Other samples with concentrations from 0.01 M to 0.05 M of CH3CO2NH4 and

CH3CO2Na, and with DI water exhibited stable dispersion. One week after the sample

preparation, the C-MWCNT samples with 0.05 M CH3CO2NH4 and CH3CO2Na showed

aggregation and the others with lower acetic electrolyte concentrations or DI water displayed

no visually discernible differences of dispersion stability. The observation results show that

C-MWCNTs were stable for at least a week in the suspensions with CH3CO2NH4 or

CH3CO2Na concentrations lower than 0.05 M, however, when the electrolyte concentrations

were higher than 0.05 M the stability decreased dramatically. The non-monotonic relation

between the electrolyte concentration and the suspension stability was hypothetically caused

by the competition between the effects of the hydrogen bond and the compression of the

electrical double layer. Figure 9 shows the results of dispersion stability observation of C-

MWCNTs in CH3CO2Na solutions by the digital camera. CH3CO2Na samples showed

similar results with the results of CH3CO2NH4 samples presented in Figure 8. Pristine

MWCNTs aggregated quickly after the sonication process. The samples with 0.01 M, 0.02 M

and 0.03 M CH3CO2Na kept stable dispersion for more than a week; 0.1 M and 0.05 M

samples settled down after a day and a week, respectively, from when the samples were

prepared.


Figure 8. Dispersion statuses of pristine MWCNTs and C-MWCNTs in different solutions

such as DI water and CH3CO2NH4.


Figure 9. Dispersion statuses of pristine MWCNTs and C-MWCNTs in CH3CO2Na solutions.

C-MWCNTs suspensions with DI water and 0.01, 0.02, 0.03 and 0.05 M of CH3CO2NH4 or

CH3CO2Na, were stored for two weeks and measured several times using UV-Vis

spectrophotometry to investigate the sample stability. Figure 10 shows the measurement

results of CH3CO2NH4 and the results of CH3CO2Na samples can be found in Figure 11. All

the freshly prepared C-MWCNT samples in CH3CO2NH4 and CH3CO2Na solutions showed

higher absorbance, which indicated better dispersity, than the sample in DI water. The UV

absorbance of the C-MWCNTs in the 0.05M samples started to drop significantly one day

after sample preparation, accompanied by visual observation of aggregation of C-MWCNTs

and separation of the sediment and supernatant. Absorbance for all other samples decreased

slightly with time, but stayed relatively stable for two weeks. The two samples 0.01 and 0.02

M for each electrolyte showed higher absorbance than the DI water sample for the entire

period. Both 0.03 M samples with CH3CO2NH4 and CH3CO2Na showed steeper decreasing

trend of absorbance than 0.01, 0.02 M samples and DI water sample.


Figure 10. Dispersion stability of C-MWCNT dispersed suspensions with different

concentrations of CH3CO2NH4 measured by UV-Vis spectrophotometer.

Figure 11. Dispersion stability of C-MWCNT dispersed suspensions with different

concentrations of CH3CO2Na measured by UV-Vis spectrophotometer.


Figure 12 shows the absorbance of C-MWCNTs in the suspensions with increasing electrolyte

concentrations. The results are for samples one week after preparation and the two different

acetic suspensions showed similar trends. The suspension stability first increased with the

increasing electrolyte concentration until it reached the peak, which was about 0.01 M for

CH3CO2NH4 samples and 0.02 M for CH3CO2Na samples, and then decreased when the

concentrations increased further. The discrepancy between the two electrolytes was caused by

different cations which affected the electrical double layer thickness. Since NH4+ possesses

bigger ionic radius (1.43 Å) than Na+ (0.95 Å), it leads to a weaker ionic hydration (91,92).

Thus the electrical double layer is compressed more due to weaker repulsion among NH4+

counterions and the particle surfaces (93-95). Therefore NH4+ is more efficient to promote

agglomeration of colloidal particles than Na+, on the other hand, the acetic effect for

dispersion of the samples with identical concentrations of the two electrolytes is the same.

Thus the concentration for the most stable condition with the CH3CO2NH4 samples was

lower than that of CH3CO2Na samples.

Figure 12. UV absorbance of C-MWCNT dispersed suspensions with different concentrations

of electrolytes in the solutions (a week after preparation of the samples).


In order to check the consistency of the aerosol method with the UV-Vis spectrophotometry,

samples prepared for penetration experiments were measured by both UV-Vis

spectrophotometer and the SMPS. Figure 13 shows comparison between the two methods,

which demonstrated similar trends depending on the ionic concentration. The ionic

concentration of CH3CO2NH4 was varied as 0, 0.007, 0.01, 0.014, 0.021 and 0.3 M, and the

measured absorbance and particle number concentration at the mode of mobility size

distribution were normalized by the values for DI water samples. The measured values

increased when the ionic concentration increased until it reached 0.014 M and decreased with

further increase of the ionic concentration. The inset figure in Figure 13 presents SMPS

results for the samples. The change of the peak size, which is closely related to the dispersion

status (73), was apparent with variation of the ionic concentration.

Figure 13. Normalized absorbance measured by UV-Vis spectrophotometer and number

concentrations measured by SMPS at the mode of mobility size distributions for C-MWCNT

suspensions with DI water and different concentrations of CH3CO2NH4.


Penetration tests using C-MWCNT suspensions

The penetration experiments were performed using C-MWCNT samples in DI water and

solutions with different concentrations of CH3CO2NH4 and KCl. The samples with

CH3CO2NH4 concentrations below 0.02 M showed higher penetrations than DI water and

KCl samples (Figure 14 and Table 1). The inset of Figure 14(a) shows pictures of downstream

samples after penetration experiments and the samples with lower transparency generally

corresponded to higher penetration. This result was in agreement with our observation of

better dispersity of C-MWCNTs in CH3CO2NH4 solutions shown in Figure 13. Because

singly dispersed C-MWCNTs have higher mobility and are more likely to penetrate the

porous media due to smaller geometrical sizes and possible alignment effect with the liquid

flow. In contrast, C-MWCNTs in KCl solutions penetrated less than in DI water as shown in

Table 1 which is consistent with results in the literature (96) and can be explained by the

classical DLVO theory. The calculated DLVO results for MWCNTs and PSL particles in

electrolytes solutions are presented in the Appendix A2. The electrical double layers of both

the C-MWCNTs and porous media are compressed when the ionic strength increases, hence

the electrostatic repulsive force and energy are reduced as shown in Figure A2-1. Furthermore

according to the Lewis acid-base interaction theory, decreased hydrophilicity of C-MWCNTs

caused by increasing interaction between electrolyte ions and water molecules, can also

support the lower penetration of C-MWCNTs in KCl solutions than in DI water. When

filtration efficiency was high and the dispersion stability was low, C-MWCNTs were easily

captured by the filter and then they formed the filter cakes and clogged the pores as shown in

Figure 15. Suspensions with different concentrations of C-MWCNTs, such as 0.003, 0.005,

0.007 and 0.01 wt%, were additionally tested against the Nuclepore filter with 3 µm pore size.

Better dispersion was achieved with lower C-MWCNT concentrations; on the contrary, higher

concentration samples comprised more agglomerated C-MWCNTs (73). As shown in the


figure, pores were more severely clogged by the C-MWCNTs in the same filtration duration

by higher concentration samples than lower concentration samples.

Furthermore, the aerosolization-quantification method was used to obtain the upstream and

downstream concentrations for the penetration experiments using C-MWCNTs as shown in

Figure 14(b). Adding CH3CO2NH4 up to 0.014 M into the suspension led to increasing

penetrations of C-MWCNTs and good agreement with the better dispersity shown above. The

penetration decreased when the CH3CO2NH4 concentration further increased to 0.021 M and

the resultant curve crossed that for the DI water sample. The two penetration curves were in a

similar range, which agrees with the fact that the overall efficiencies for the two samples were

similar as shown in Figure 14(a). The details regarding filtration will be discussed in Chapter

5.


Figure 14. Penetration of C-MWCNTs in the CH3CO2NH4 suspensions through the porous

media measured by (a) UV-Vis spectrophotometer, (b) SMPS for 100-200 nm mobility size

range. The inset of (a) shows photos of the samples with different ionic concentrations after

penetration and the numbers above the data points indicate the samples.


Table 1. Penetration of C-MWCNTs through the porous media with different ionic

concentrations of KCl in the solutions.

Ionic concentration (moles/liter)

Average penetration Maximum penetration

Minimum penetration

0.01 0.063 0.078 0.038

0.02 0.104 0.145 0.065

0.03 0.12 0.173 0.069

Figure 15. SEM images of pores of Nuclepore filters with 3 µm pore size after penetration

experiments for C-MWCNTs.


Dispersion stability of PSL suspensions

The dispersion stability of PSL particles in DI water and CH3CO2NH4 solutions were also

measured by the UV-Vis spectrophotometer and the results were presented in Figure 16.

COOH-PSL particle suspensions with 0.01 M and 0.02 M CH3CO2NH4 as well as PSL

samples with and without carboxyl functional groups in DI water showed stable conditions

for more than two weeks. The absorbance of the COOH-PSL sample with 0.05 M

CH3CO2NH4 continuously decreased for the entire period; the UV absorbance for the PSL

samples without the carboxyl functional group with 0.01 M and 0.02 M CH3CO2NH4

decreased at similar rates. The above results showed that the COOH-PSL samples in the

acetic solutions with low ionic concentrations (0.01 M and 0.02 M) were stable, however, the

dispersion of PSL samples in DI water was already stable and no enhancement by the

carboxyl-acetate hydrogen bond was observed. In the PSL sample with a high ionic

concentration (0.05 M), the normal electrolyte effect became apparent and particle

agglomeration was promoted.


Figure 16. Changes of dispersion stabilities of PSL and COOH-PSL particle suspensions with

different concentrations of electrolytes in the solutions.

Penetration tests using PSL suspensions

The aerosolization-quantification method was used for the penetration experiments using PSL

and COOH-PSL particles in different solutions and the results were shown in Figure 17. The

penetrations for PSL particles in DI water were close to 100 %, which were attributed to the

good dispersion and low attachment efficiency of PSL particles on the filter surfaces. The

calculated energy barrier of the system by DLVO theory, which can be found in Figure A2-2

in the Appendix, was so high that PSL particles could hardly overcome it to achieve

permanent attachment to the contact surfaces. The penetrations of COOH-PSL particles in

CH3CO2NH4 solutions with different concentrations were essentially as high as the PSL

sample in DI water. In contrast, CH3CO2NH4 acted as a normal electrolyte and adding

CH3CO2NH4 resulted in decreasing penetrations for PSL particles without the functional

group. Our penetration tests did not show improved mobility due to the carboxyl-acetate


hydrogen bond compared to the PSL sample in DI water, however, the results confirmed that

the acetic electrolyte acted differently than normal electrolytes. The beneficial effects of the

carboxyl-acetate hydrogen bond on dispersity and mobility seemed to balance out the normal

electrolyte effects for COOH-PSL particles.

Figure 17. Penetrations of PSL particles, with and without the carboxyl functional group, in

CH3CO2NH4 suspensions through the Nuclepore filter.

Discussion

The enhancement of dispersion stability achieved in the study can be explained by a strong

hydrogen bond between the electrolytic CH3COO- and carboxyl. To enhance dispersion of

nanoparticles with the carboxyl functional group in aqueous solutions, the hydrogen bond

between CH3COO- and carboxyl should be stronger and more competitive than other

influencing bonds, such as hydrogen bonds between carboxyl and water molecules, acetate

and water molecules, and suppression of the double layer by electrolyte ions in the solutions

additionally. Figure 18 describes the strength of intermolecular O―H…O hydrogen bonds


between different molecules and it can be reflected by the mean distance of O…O. The values

are 0.2544 nm, 0.294nm (or 0.2591nm due to different hydrogen donor and acceptor) and

0.2807 nm for bonds between carboxyl and CH3COO-, carboxyl and water molecule,

CH3COO- and water molecule, respectively (75). So the hydrogen bond between carboxyl

and CH3COO- is the strongest in the suspension system. Quantitative comparison between the

effect of the carboxyl-CH3COO- hydrogen bond and the effect of electrolyte ions on the

electrical double layer is not yet theoretically available, however, according to our results, the

hydrogen bond was more influential on the dispersity than the effect of electrolyte ions on the

shrinkage of electrical double layer when concentrations of acetic ions were below certain

concentrations (approximately 0.03 M for both CH3CO2NH4 and CH3CO2Na in the study),

and vice versa when the acetic concentration was higher than the threshold.

Figure 18. The mean distances of O atoms in different hydrogen bonds, such as carboxyl and

CH3COO-, carboxyl and water molecule, CH3COO- and water molecule.

Previous studies (97,98) show that the ionization of the carboxyl groups on CNT surface is

varied by concentration of electrolyte, due to the correlated change in the pH value. The

carboxyl group tends to be dissociated and negatively charged in a high pH solution so that


CNT dispersity is improved with increase of the pH value. However, this effect cannot

explain the results in the present study, since both of the acetic electrolytes employed in our

experiments are commonly used to prepare buffer solutions. The pH values of ammonium

acetate solutions stayed in a narrow range (6.98-7.11) for the entire concentration range (0.01-

0.2M), meanwhile for sodium acetate, that value increased mildly from 6.99 to 7.95 with the

concentration increase. In the measured pH range, the zeta potential of the CNTs, which

corresponds to the electrostatic repulsion between particles, only has a minor change (98) and

cannot account for the observed change of the CNT dispersity in our experiments.

43 Bahk et al. (2013) Aerosol Sci. Tech. 47(7):776-734; Bahk et al. (2014) Aerosol Air Qual. Res. 14(5):1352-1359

3. Physical characterization II: Geometrical length measurement

The geometrical length is an important intrinsic property of CNTs and the extremely

high length-to-diameter ratio is one of the hallmarks of CNTs. The geometrical length plays

an important role on the mechanical, electrical, and optoeletronic properties (18) and has

significant effects on CNT agglomeration status and toxicity (15,99). For airborne CNTs, the

geometrical length affects the mobility and transport properties (100), and it is necessary to

model the lung deposition of fibrous particles or to assess filtration efficiency of CNTs

(61,62). For these reasons, determination of the geometrical length of CNTs aroused

significant interests.

Scanning electron microscopy (SEM) is commonly used to measure the geometrical

properties of nanoparticles. Recently Li et al. (100) investigated the theoretical model to

calculate geometrical length of cylindrical shaped particles. Lall and Freidlander (101)

provided a model which could calculate the geometrical property of chain aggregates as a

function of the mobility diameter. In the present study, the geometric lengths of CNTs were

empirically measured by SEM analysis and compared with models in previous studies. This

study demonstrated that the combination of the mobility measurement and model calculation

could give satisfactory results for CNT lengths.

A new method to measure the geometrical length of CNT is introduced in the study.

Filtration method was employed to calculate the length of CNTs. Seto et al. (60) and Wang et

al. (62) validated experimental data of penetration of CNTs through uniform screens, which

was characterized by using the single fiber filtration theory. Elongated aerosol particles

including CNTs and chain agglomerates have lower penetration than spheres of the same

electrical mobility or same volume, because they possess longer geometrical lengths and

stronger interception effect (22,62,102,103). Conversely, the geometrical length can be


calculated from penetration of airborne CNTs with the single fiber filtration theory. This

method shows possibility of fast measurement of geometrical parameter of CNTs with a

simple setup for filtration using the screen. Myojo (104) developed a method to determine the

length distribution of fibrous aerosol based on the penetration results through mesh screens.

His model included several simplifications, including neglecting the flow disturbance by the

screen wires and inertia of the fibers. The demonstrated the method was applicable for fibers

with lengths in the range of 5-80 µm. Introduced method enables measurement of the

geometrical length of elongated particles down to hundreds of nanometers.

The filtration methods using nanofiber filter samples with different solidities and the

Nuclepore filter were also investigated and the results were compared with measured

geometrical lengths of CNTs by scanning electron microscopy (SEM) analysis (105,106).

3.1. Electron microscopy analysis


A Collison type atomizer and electrospray (TSI 3480) were used to generate airborne CNTs.

The MWCNTs from the generators went through a diffusion dryer to ensure that liquid vapors

were removed. Single dispersed airborne CNTs were collected on silicon substrates to prepare

the samples for SEM analysis by using a nanometer aerosol sampler (TSI 3089, USA). A

large number of classified airborne MWCNTs by DMA were collected and lengths of them

were measured from the SEM images. The obtained SEM images were analyzed by an image-

processing software (ImageJ, National Institutes of Health, USA). The length distribution of

MWCNTs was fitted to a lognormal distribution and mean lengths of MWCNTs were

obtained. The obtained geometrical lengths were compared with several models provided by

Li et al. (100) and Lall and Freidlander (101). The lengths were also compared with calculated

lengths by the filtration method.


MWCNTs with diameters about 15-20 nm (Baytubes, BMS, Germany), which were

functionalized by the nitric acid refluxing method, and about 20-30 nm (Cheaptubes, USA),

which possessed a functional group COOH, were used in the experiments and dispersed in DI

water.

Relation between the geometrical length and electrical mobility diameter

The geometrical length of airborne MWCNTs was measured by SEM analysis and showed a

single mode distribution as shown in Figure 19. In order to define the mean length of

MWCNTs, the geometrical length distribution was fitted to a lognormal distribution. For

instance, Figures 19 and 20 show results of classified MWCNTs (Cheaptubes) with 75 nm

mobility size and the mean length was 221 nm. Measured lengths for different mobility sizes

were shown in Table 2. The relation between the electrical mobility diameter and geometrical

length of airborne MWCNTs from Cheaptubes was shown in Figure 21 where the x axis is the

mobility diameter (dm) divided by the Cunningham slip correction factor (Cc). Results of

MWCNTs from Baytubes showed similar results and were shown in Figure 22. According to

linear fitting of the data, Baytubes CNTs possess longer geometrical lengths than MWCNTs

from Cheaptubes with the same mobility diameter. This was attributed to that CNTs from

Baytubes have smaller tube diameters than those of CNTs from Cheaptubes, in agreement

with the models of Li et al. (100) and Lall and Friedlander (101).

Figure 23 shows comparison of measured geometrical lengths of CNTs with several previous

studies. Uncertainties of each study were calculated using the standard deviation of the

experiment data and shown in Figure 23. MWCNTs from Cheaptubes and Baytubes showed

similar length distributions as the results presented by Kim and Zachariah (107). Wang et al.

(62) and Seto et al. (60) showed considerably longer lengths of CNTs than present results and

Kim and Zachariah' results. This difference seems to be due to the sizes of diameters of used


CNTs. CNTs from Cheaptubes and Baytubes and CNTs used in Kim and Zachariah's study

possessed relatively small diameter values at 25 nm, 20 nm and 15 nm, respectively. However,

CNTs in the studies by Wang et al. (62) and Seto et al. (60) possessed 85 nm and 65 nm

diameters, respectively. Both models of Li et al. and Lall and Friedlander were derived for

free-molecular flows, which indicated that the Knudson number (= 2𝜆𝜆/𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶) should be

significantly larger than 1. It appears that the flows for CNTs in Wang et al. (62) and Seto et

al. (60) were not in the free-molecular flow regime. It might be part of the reason why the

particles, which possessed different geometrical lengths, were classified as same mobility size

particles.

Figure 19. Measured distribution of the length of MWCNTs with75 nm mobility size.


Figure 20. Fitting for measured length distribution of electrical mobility size of 75 nm and log

normal distribution.

Figure 21. Linear fitting for the length of CNTs and electrical mobility size of CNTs divided

by the slip correction factor, Cheaptubes.


Figure 22. Linear fitting for the length of CNTs and electrical mobility size of CNTs divided

by the slip correction factor, Baytubes.

Figure 23. Comparison of measured geometrical lengths of CNTs with previous studies.


3.2. Filtration method

Theoretical models

Li et al. (100) developed a model which described relation between the geometrical length

and electrical mobility diameter of cylinder shaped particles. They considered two cases,

random orientation and total aligned orientation.

For the random orientation, the geometrical length of a CNT can be expressed as,

𝐿𝐿𝐶𝐶𝐶𝐶𝐶𝐶 = 6𝜅𝜅𝜅𝜅𝑑𝑑𝑚𝑚𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑐𝑐

, (3-1)

where 𝜆𝜆 is the mean free path of gas, 𝑑𝑑𝑚𝑚 is the electrical mobility size, 𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶 is the CNT

diameter and 𝐶𝐶𝑐𝑐 is the Cunningham slip correction.

𝜅𝜅 = ln [(1+𝑎𝑎)/(1−𝑎𝑎)]2𝑎𝑎(𝑓𝑓+𝑘𝑘3) , (3-2)

where

𝑎𝑎 = 𝛽𝛽𝑘𝑘3𝛽𝛽(𝑓𝑓+𝑘𝑘3)+𝑘𝑘2

,𝑘𝑘1 = 𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝜋𝜋𝜋𝜋𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶

, 𝑘𝑘2 = 𝜋𝜋𝑓𝑓6

+ 43

,𝑘𝑘3 = 2 − 6−𝜋𝜋4𝑓𝑓 and 𝑑𝑑𝑚𝑚

𝐶𝐶𝑐𝑐= 𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶2𝑎𝑎(𝛽𝛽(𝑓𝑓+𝑘𝑘3)+𝑘𝑘2)

3𝜅𝜅 𝑙𝑙𝑛𝑛[(1+𝑎𝑎) (1−𝑎𝑎)⁄ ]. (3-3)

f is the momentum accommodation coefficient (=0.9), 𝛽𝛽 is the aspect ratio of CNT, 𝑛𝑛 is the

number of charge on CNT, 𝑒𝑒 is the elementary electrical charge, 𝐾𝐾𝑛𝑛 is the Knudson number

(=2𝜆𝜆/𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶) and 𝜇𝜇 is the gas viscosity.

For the total alignment,

𝐿𝐿𝐶𝐶𝐶𝐶𝐶𝐶 = 6𝜅𝜅𝑑𝑑𝑚𝑚𝑓𝑓𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝐶𝑐𝑐

. (3-4)

Lall and Friedlander (101) theoretically determined the number of primary particles for the

aggregates. The Lall and Friedlander model was developed for aggregates with open

structures. Considering the geometrical similarity between a CNT and a chain aggregate

composed of primary spheres with the same diameter, we applied the Lall and Friedlander

model to compute the CNT length. The assumption is that the CNT length can be


approximated as the length of a chain aggregate composed of primary spheres with the same

diameter as the CNT diameter 𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶.

The number of primary particles in an aggregate can be calculated as

𝑁𝑁 = 3𝜋𝜋𝑑𝑑𝑚𝑚𝜅𝜅𝑐𝑐∗(𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶 2⁄ )2𝐶𝐶𝑐𝑐

, (3-5)

where 𝑐𝑐∗ is the dimensionless drag force, which takes value as 9.17 for the random orientation

and 6.62 for the parallel orientation.

The geometrical length of CNT can be expressed as

𝐿𝐿𝐶𝐶𝐶𝐶𝐶𝐶 = 𝑁𝑁𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶. (3-6)

The single fiber theory was employed for CNT penetration through a screen filter, which was

used by Wang et al. (61), to calculate the geometrical lengths of CNTs. Penetration of the

airborne CNTs is calculated as

𝑃𝑃 = exp �− 4𝛼𝛼𝐸𝐸𝐶𝐶𝑡𝑡𝜋𝜋𝑑𝑑𝑓𝑓(1−𝛼𝛼)�, (3-7)

where 𝛼𝛼 is the solidity of the filter (=0.345), 𝑡𝑡 is the thickness of 20 layers of the screen filter,

𝑑𝑑𝑓𝑓 is the wire diameter in the screen (= 20𝜇𝜇𝜇𝜇) and the total single fiber efficiency 𝐸𝐸𝐶𝐶 is

summation of efficiency due to the diffusion, interception, impaction and interception of

particles undergoing diffusion

𝐸𝐸𝑡𝑡 = 𝐸𝐸𝐷𝐷 + 𝐸𝐸𝑅𝑅 + 𝐸𝐸𝐼𝐼 + 𝐸𝐸𝐷𝐷𝑅𝑅. (3-8)

The single fiber efficiency due to diffusion is expressed as

𝐸𝐸𝐷𝐷 = 2.7𝑃𝑃𝑒𝑒−2 3⁄ (3-9)

where

𝑃𝑃𝑒𝑒 = 𝑑𝑑𝑓𝑓𝑈𝑈0𝐷𝐷

= 𝑑𝑑𝑓𝑓𝑈𝑈03𝜋𝜋𝜋𝜋𝑘𝑘𝐶𝐶𝐶𝐶𝑐𝑐

𝑑𝑑𝑝𝑝. (3-10)

In the equations, 𝑃𝑃𝑒𝑒 is the Peclet number, 𝑈𝑈0 is the face velocity, 𝐷𝐷 is the particle diffusion

coefficient, 𝑘𝑘 is Boltzmann constant, 𝑇𝑇 is the temperature and 𝑑𝑑𝑝𝑝 is the particle diameter.


Wang et al. (62) showed the diffusion coefficient can be directly related to the electrical

mobility, thus 𝑑𝑑𝑝𝑝 can be approximated using the mobility diameter of CNTs.

The single fiber efficiency due to interception can be written as (108)

𝐸𝐸𝑅𝑅 = 1+𝑅𝑅2𝑛𝑛𝐾𝐾

[2 ln(1 + 𝑅𝑅) − 1 + 𝛼𝛼 + � 11+𝑅𝑅

�2�1 − 𝛼𝛼

2� − 𝛼𝛼

2(1 + 𝑅𝑅)2], (3-11)

where

𝑅𝑅 = 𝑑𝑑𝑝𝑝 𝑑𝑑𝑓𝑓⁄ and 𝐾𝐾𝐾𝐾 = − ln𝛼𝛼 2⁄ − 3 4⁄ + 𝛼𝛼 − 𝛼𝛼2 4⁄ . (3-12)

𝐾𝐾𝐾𝐾 is the Kuwabara hydrodynamic parameter. The particle diameter 𝑑𝑑𝑝𝑝 can be directly used

in the equation for spheres, however other assumptions are needed for 𝑑𝑑𝑝𝑝 of CNTs. Wang et

al. (61,62) computes the CNT interception length by using the geometrical length and the

orientation angle 𝜃𝜃 between the CNT and the screen wire surface. When random orientation

of the CNTs is assumed, the angle 𝜃𝜃 takes the average value of 40°. Then the interception

parameter for CNTs is

𝑅𝑅 = 𝐿𝐿𝐶𝐶𝐶𝐶𝐶𝐶 sin𝜃𝜃 𝑑𝑑𝑓𝑓⁄ . (3-13)

The efficiency due to inertial impaction can be written as

𝐸𝐸𝐼𝐼 = 1(2𝑛𝑛𝐾𝐾)2

[(29.6 − 28𝛼𝛼0.62)𝑅𝑅2 − 27.5𝑅𝑅2.8]𝑆𝑆𝑡𝑡𝑘𝑘, (3-14)

where 𝑆𝑆𝑡𝑡𝑘𝑘 = 𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶3𝜌𝜌𝐶𝐶𝐶𝐶𝐶𝐶𝑈𝑈018𝜋𝜋𝑑𝑑𝑓𝑓𝑅𝑅𝑎𝑎𝑎𝑎

and 𝜌𝜌𝐶𝐶𝐶𝐶𝐶𝐶 is the CNT density, 1.74 𝑔𝑔 𝑐𝑐𝜇𝜇3(22)⁄ . The aerodynamic

radius 𝑅𝑅𝑎𝑎𝑛𝑛 of a fibrous particle is depended on the aspect ratio and can be written as

𝑅𝑅𝑎𝑎𝑛𝑛 = 𝑅𝑅𝑎𝑎𝑛𝑛1 sin2 𝜓𝜓 + 𝑅𝑅𝑎𝑎𝑛𝑛2 cos2 𝜓𝜓, (3-15)

where

𝑅𝑅𝑎𝑎𝑛𝑛1 = 𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶𝛽𝛽3[ln(2𝛽𝛽)−0.5]

and 𝑅𝑅𝑎𝑎𝑛𝑛2 = 2𝑑𝑑𝐶𝐶𝐶𝐶𝐶𝐶𝛽𝛽3[ln(2𝛽𝛽)+0.5]

. (3-16)

In the equation 𝑅𝑅𝑎𝑎𝑛𝑛1 is for a paralle fiber orientation and 𝑅𝑅𝑎𝑎𝑛𝑛2 is for a perpendicular

orientation within the gas flow. For the random orientation angle 𝜓𝜓 takes the value 54.74°.


For interception of particles undergoing diffusion, single fiber efficiency is described as

𝐸𝐸𝐷𝐷𝑅𝑅 = 1.24𝑅𝑅2 3⁄

(𝑛𝑛𝐾𝐾𝐾𝐾𝑛𝑛)1 2⁄ . (3-17)

The geometrical lengths of CNTs were obtained with spreadsheet software which employed

the single fiber theory. Experimental results of CNT penetration through the screen were put

into the software to calculate the geometrical length of CNTs. As shown in the equations, the

penetration is a function of the diameter, density and length of CNTs and the diameter of

screen wire. Penetration is determined experimentally and most of the parameters, such as the

diameter and density of CNTs and diameter of screen wire, are known values. Only one

unknown value is the geometrical length of CNT and that can be calculated from the

equations.

For nanofiber fiber, different fiber diameters were applied for nanofiber and substrate fiber,

and values were 250 nm diameter and 20 µm, respectively. 3 layers of filters were considered

in the model and total thicknesses were 750 nm and 450 µm for nanofibers and substrate

fibers, respectively.

The single fiber efficiency due to diffusion for nanofiber can be written as (109)

𝐸𝐸𝐷𝐷 = 2.27𝐾𝐾𝐾𝐾−1 3⁄ 𝑃𝑃𝑒𝑒−2 3⁄ (1 + 0.62𝐾𝐾𝑛𝑛𝑃𝑃𝑒𝑒1 3⁄ 𝐾𝐾𝐾𝐾−1 3⁄ ). (3-18)

The efficiency due to interception is given by (110)

𝐸𝐸𝑅𝑅 = (1+𝑅𝑅)−1−(1+𝑅𝑅)+2(1+1.996𝑛𝑛𝑛𝑛)(1+𝑅𝑅)ln (1+𝑅𝑅)2(−0.75−0.5 ln𝛼𝛼)+1.996𝑛𝑛𝑛𝑛(−0.5−ln𝛼𝛼)

. (3-19)

The current study found that the aerodynamic diameter gave results in much better agreement

with the experiments. In this case, the interception parameter 𝑅𝑅 for CNTs is

𝑅𝑅 = 2𝑅𝑅𝑎𝑎𝑛𝑛 𝑑𝑑𝑓𝑓⁄ . (3-20)

The single fiber efficiency due to inertial impaction is written as (111)

𝐸𝐸𝐼𝐼 = 𝑆𝑆𝑡𝑡3

𝑆𝑆𝑡𝑡3+0.77𝑆𝑆𝑡𝑡2+0.22, (3-21)


The interaction term 𝐸𝐸𝐷𝐷𝑅𝑅 to consider interception of particles undergoing diffusion can be

written as (21),

𝐸𝐸𝐷𝐷𝑅𝑅 = 1.24∙𝑅𝑅2 3⁄

(𝑛𝑛𝐾𝐾∙𝐾𝐾𝑛𝑛)1 2⁄ . (3-22)

The capillary tube model was applied to filtration method for calculating the effective lengths

of CNTs from the obtained penetration through the Nuclepore filters. The capillary tube

models developed by Spurny et al. (112) and Manton (113,114) for spherical particle

penetration through Nuclepore filters have been modified for applications to soot and Ag

agglomerates by Chen et al. (115,116). In the models, the considered mechanisms of particle

deposition included the diffusions on filter surface and inside the pore, interception at the pore

opening and impaction on the filter surface. They found that the major differences of the

model application between spherical and agglomerate particles were on the interception and

impaction, especially on interception, whereas the diffusion deposition did not differ

significantly. In the interception model, the effective length should be used as the interception

length instead of the mobility diameter to account for the elongated shapes of agglomerates.

In present model for the Nuclepore filter, interception and diffusion were considered, because

previous studies on soot agglomerates showed impaction was negligible under the similar

situation (116). Interception length in the model without orientation angle was used due to

that the capillary tube was relatively long and the CNT might have enough time to rotate

during the time flying through the tube, so the effective length provided the highest possibility

for the CNT to be intercepted.

The partial efficiency of diffusion is expressed as (117),

𝜖𝜖𝐷𝐷 = 2.56𝑁𝑁𝐷𝐷2 3⁄ − 1.2𝑁𝑁𝐷𝐷 − 0.177𝑁𝑁𝐷𝐷

4 3⁄ , (3-23)

where the coefficient of diffusive collection, 𝑁𝑁𝐷𝐷 = 𝑡𝑡𝐷𝐷𝑡𝑡𝑟𝑟02𝑈𝑈0

, t is a filter thickness, ɛ is the filter

porosity, r 0 is a pore size, all parameters are in cm unit.


The interception efficiency can be calculated as (112),

ϵR = 𝑁𝑁𝑅𝑅(2 − 𝑁𝑁𝑅𝑅),𝑤𝑤ℎ𝑒𝑒𝑒𝑒𝑒𝑒 𝑁𝑁𝑅𝑅 = 𝐿𝐿𝐶𝐶𝐶𝐶𝐶𝐶2𝑟𝑟0

. (3-33)

According to the theory, when diameter and density of CNT and fiber diameter are known

and penetration is determined experimentally, the geometrical length of CNT can be

calculated from equations. The introduced equations in order to calculate nanoparticle

filtration efficiencies are also used in Chapter 5, the study regarding emission controls using

different filters.


MWCNT penetrations through filters, such as metal screens, nanofibers and Nuclepore filters,

were determined empirically. Figure 24 shows a schematic of the CNT filtration measurement

system, which consists of an airborne CNT generation system, a size classification system and

a penetration measurement system. The MWCNTs from the generators were classified by the

DMA and the classified airborne CNTs were carried by the air flow with controlled flow rate.

A neutralizer (Kr-85) was used to avoid electrostatic effect in the filter. Concentration of

CNTs was measured by condensation particle counters (CPC, TSI 3775) at upstream and

downstream of the filter. Then the fraction of penetration through the filter was obtained.

One of the tested filters was the screen made of 635-mesh type 304 stainless steel Results

were compared with penetrations through screens made of 635-mesh type 304 stainless steel,

which have well-defined structures and were used in the diffusion battery (TSI, model 3040).

The screen wire was 20 μm in diameter and the opening dimension was also 20 μm. The

solidity α, which is the solid fraction in the filter compared to the total volume, is 0.345 (62).

The face velocity was 5 cm/s in the filtration tests and pressure drop was 67 Pa. CNTs

penetrating through the testing filter were removed by a high efficiency filter before the flow

went to the flow meter.


Nanofiber sample A with solidity 0.134, which was obtained by Wang et al. (118) in the

previous study as samples A, B, C and D with solidities 0.134, 0.104, 0.059 and 0.034,

respectively, was used for filtration method to calculated the geometrical length of MWCNTs.

The pressure drops for the sample A was 12 Pa with 5 cm/s face velocity on the filters. Figure

25 shows SEM images of nanofiber samples. All other nanofiber samples were used in

filtration studies, which are presented in Chapter 5. As shown in the figure, airborne PSL

particles and CNTs of different sizes were captured by nanofibers. Substrate micrometer

fibers possess about 20 µm diameters and nanofibers on the substrate fibers possess 150-300

nm diameters. In the experiments three layers of the same type nanofiber filters were used in

order to show the clear mobility size dependency.

Nuclepore filter with 1 µm pores was also used for investigating the filtration method. Its

pressure drop under the condition with 5 cm/s face velocity on the filters was 810 Pa and it

was relatively higher than other tested filters. The solidity and thickness of filter were 0.843

and 11 µm, respectively. Further results regarding filtration studies are discussed in Chapter 5.

Figure 24. Experimental system for CNT filtration tests.


Figure 25. SEM images of nanofiber filter samples A, B, C and D with solidities 0.134, 0.104,

0.059 and 0.034, respectively. PSL particles and CNTs were captured by nanofibers.

CNT Lengths measurement by the filtration method and comparison with

theoretical models

Penetration of MWCNTs were measured and applied to the model, which included the

particle capture mechanisms, to calculate the geometrical length of CNTs. Table 2 showed

penetration of both CNTs through the screen with flow velocity 5 cm/s. CNTs from Baytubes

showed lower penetrations than those from Chepatubes, because they possessed longer

geometrical lengths. We input the penetration as a parameter to the model. Uncertainties due

to the fluctuation of flow rate through the filter (± 3%) and due to variation of the penetration

among 4 trials of experiments (± 1%) were considered. Obtained results were shown in

Figures 26 and 27 for MWCNTs from Cheaptubes and Baytubes, respectively. The standard


deviation was used for error bars for the lengths of CNTs from the SEM measurement. The

filtration approach showed reasonable results, which were mostly within or close to the

uncertainty ranges of measured lengths of CNTs by SEM analysis.

The measured length distribution of MWCNTs was compared with theoretical models

including Li’s model (100) and Lall’s model (101). Comparison with our SEM results showed

that the models give satisfactory results with proper choices of parameters. Lall’s model for

the random orientation fits especially well with the data of Cheaptubes as shown in Figure 26.

The results of Baytubes lied between Li's and Lall's models with random orientation.

Measured and calculated values of all models are shown in Table 2.

Two approaches for fast determination of CNT geometrical lengths were evaluated. The first

one is to measure the CNT mobility size, then to compute the CNT geometrical length using

models relating the mobility diameter and length. Two models (100,101) were used for this

purpose. Our results demonstrated that the models agreed well with experimental results when

random orientation of CNTs was assumed. Both of the models were derived for free-

molecular flows, thus their application is limited to CNTs with small diameters. Both of the

models include adjustable parameters depending on particle orientation and accommodation

coefficients. The values of the parameters and coefficients working well for the CNTs in the

present study may not be optimal for other types of CNTs.

The developed approach (filtration method) for CNT length measurement based on filtration

through the screens showed reasonable agreement with the SEM analysis. The range of

uncertainty depends on the filtration experiments. A number of limitations exist for this

approach. Bending and curling of CNTs may occur when the aspect ratio is too high (61,62).

CNTs may align with the flow when the face velocity is too high (119,120). These effects are

not included in the filtration model and may lead to underestimation of the CNT length. On

the other hand, CNT agglomeration in the filter or loading on the filter may lead to

overestimation of the CNT length.


Figure 26. Comparison between the measured geometrical length of CNTs by SEM analysis

and obtained length of CNTs by filtration method (Screens, Cheaptubes).

Figure 27. Comparison between the measured geometrical length of CNTs by SEM analysis

and obtained length of CNTs by filtration method (Screens, Baytubes).


Table 2. Measured and calculated geometrical lengths (Screens) of CNTs

Cheaptubes

dm Measured

length from SEM (nm)

σg Penetration Calculated length (nm)

Minimum length considering

uncertainty (nm)

Maximum length considering

uncertainty (nm)

75 221 1.46 0.363 151 104 202

90 331 1.36 0.434 210 158 265

110 430 1.27 0.503 289 235 343

150 606 1.34 0.600 383 331 435

dm Li’s model randomly

rotating (nm)

Li’s model totally aligned (nm)

Lall’s model random

orientation (nm)

Lall’s model parallel orientation (nm)

75 162 374 230 319

90 231 516 318 441

110 327 730 450 624

150 546 1221 753 1043

Baytubes

dm Measured




uncertainty (nm)


uncertainty (nm)

65 207 1.45 0.228 441 352 469

80 294 1.43 0.303 446 390 501

95 373 1.43 0.362 487 433 540

110 429 1.37 0.411 516 464 567

dm Li’s model randomly

rotating (nm)

Li’s model totally aligned (nm)

Lall’s model random

orientation (nm)

Lall’s model parallel orientation (nm)

65 162 361 222 308

80 235 524 323 448

95 318 709 437 606

110 410 913 563 780

The filtration method was also evaluated by inputting penetration of CNTs through the

nanofiber filter sample A as a parameter to the model. The results were shown in Figures 28

and 29, for CNTs from Cheaptubes and Baytubes, respectively. The obtained geometrical


lengths of CNTs were compared with measured lengths by SEM analysis. The same

uncertainties considered in the length calculation using screen filter results were used. The

standard deviation was also used to insert the error bars for obtained lengths by SEM analysis.

The comparisons showed reasonable agreement in the range that we chose in the study,

however, results for the smaller size end in the range showed bigger discrepancy than larger

size end in the range. The reason might be that smaller size particles are mainly captured by

diffusion rather than interception, thus the model, which relies on the efficiency due to

interception to compute the geometrical length, showed a discrepancy in the smaller size

range.

Using the aerodynamic diameter of CNTs for the interception parameter gave rise to better

agreement between the model and experiments than using the measured geometrical length of

CNTs. The reason is that the consideration of aerodynamic diameter includes effects of

bending and curling of CNTs fortuitously, because when bending and curling happen on

CNTs the interception lengths of CNTs are shortened (62). However, the aerodynamic

diameter of CNTs used in the model for interception might be shorter than the actual

interception lengths of CNTs and led to a model overestimation of the penetration, especially

for Baytubes in the range of 65 - 110 nm. This also provides an explanation for the

discrepancy between the length calculation and measurement for Baytubes in Figure 29. The

better agreement for Cheaptubes shown in Figure 28 indicates that the aerodynamic diameter

is a better approximation to the interception length in this case.

In the model, condition of filters during the experiments as particle loading effect on the

filters is not considered. That might be another reason for the discrepancy between the model

and experimental results. The measured and calculated lengths, penetration through the

nanofiber filters and uncertainties for each case are listed in Table 3.


Figure 28. Comparison between calculated length of CNTs using filtration method and

measured geometrical length of CNTs by SEM analysis (Nanofiber filter A, Cheaptubes).

Figure 29. Comparison between calculated length of CNTs using filtration method and

measured geometrical length of CNTs by SEM analysis (Nanofiber filter A, Baytubes).


Table 3. Measured and calculated geometrical lengths (nanofiber sample A) of CNTs

Cheaptubes

dm Measured




uncertainty (nm)


uncertainty (nm)

75 221 1.46 0.303 288 272 304

90 331 1.36 0.314 352 337 367

110 430 1.27 0.315 434 419 447

140 563 1.34 0.306 533 521 547

Baytubes

dm Measured




uncertainty (nm)


uncertainty (nm)

65 207 1.45 0.269 327 308 343

80 294 1.43 0.289 393 376 409

95 373 1.43 0.296 459 444 474

110 429 1.37 0.296 522 507 536

In addition, the geometrical lengths of MWCNTs were calculated by the filtration method

using the Nuclepore filter and Table 4 shows the calculation results. The results were

compared with the average values determined by the SEM analysis. The geometrical lengths

of MWCNTs calculated by the filtration method using Nuclepore filter showed a good

agreement with SEM measured lengths. It is seen the average SEM measured length was very

close to that by model with a relative difference of less than 13%. To be noted, the calculated

length for 500 nm MWCNTs could not be obtained because the penetration was zero and the

calculated length could be any value larger than 1000 nm.


Table 4. The geometrical length (nm) of different mobility diameter MWCNTs determined

from SEM analysis and filtration method (Baytubes).

Mobility diameter (nm) Measured length by SEM (nm)

Calculated length from model (nm)

Relative difference, %

50 138±59 143 3.6 100 312±93 320 2.6 150 493±240 440 10.8 200 642±176 570 11.2 300 928±476 855 7.9 400 1092±397 970 12.6 500 1362±472 >1000 --

64 Wang, Bahk, Chen, Pui (2015) Carbon 48(10):997-1008

4. Physical characterization III: Effective density measurement

and Fractal model for CNT agglomerate

Characterization of CNT agglomerate is as crucial as that of single standing CNTs,

because CNTs tend to agglomerate due to their high aspect ratios and van der Walls forces.

The manufacturers often provide CNTs in the powder form, which can be composed of loose

agglomerates in the range of 0.1 – 1 mm (24). Such large agglomerates have limited mobility,

and are easier and safer to handle and transport (99). CNT agglomerates in aqueous media

have been widely encountered in toxicity studies and the agglomeration status affects the

resultant toxicity (15,121,122). CNT agglomerates in the airborne form are observed in many

exposure studies during CNT production or handling (36-38,123) and CNTs are seldom

present as individual straight nanofibers at workplaces (25). The literature on CNT density

showed substantial discrepancies, which may be related to the different types of density being

considered or due to a variety of different form of CNTs including individual CNTs, cylinder-

like, bended, coiled, CNT bundles and quasi-isometric shaped agglomerates (8,22-25,39).

In the present study, the DMA-CPMA and DMA-APM tandem measurement were performed

to obtain effective densities for CNTs with different mobility diameters and a fractal model

was developed to establish the relationship between the MWCNT properties including tube

diameters, length, intrinsic density and the agglomerate properties such as the mass, porosity,

effective density, surface area and characteristic diameters. The model developed to compute

the geometrical outer diameters and fractal dimensions of the agglomerates agreed well with

the results by the SEM image analysis and literature data. The model can be used for

monitoring CNT emissions at workplaces, real-time measurement and toxicity studies of

airborne CNTs (124). For example, the obtained density data were used in the lung deposition


model in Chapter 6. Furthermore, airborne MWCNT agglomerates were collected on silicon

substrate and analyzed by SEM. Outer diameters and 2D fractal dimensions for CNT

agglomerates were obtained and compared with the developed model.

4.1. Densities of MWCNTs


MWCNTs with outer diameter of 15–20 nm (Baytubes) were purchased from Bayer Material

Science (BMS), Germany. Another type of MWCNTs with an outer tube diameter of 10-20

nm was purchased from Timesnano (China). The properties of both types of MWCNTS

provided by the manufacturer are listed in Table 5. It can be seen that the inner and outer tube

diameters of them are in similar ranges. Both MWCTNs were then treated with the nitric acid

refluxing method, which gave rise to the carboxyl functional group (COOH) on the tube

surfaces and reduced tube lengths, and removed the metal catalysts in the MWCNTs (57). The

surface carboxyl group increased the hydrophilicity of the MWCNTs and resulted in better

dispersion in aqueous suspensions, which were then atomized to generate airborne MWCNTs.

Chapter 2 showed that the agglomeration status of the airborne MWCNTs could be adjusted

by controlling their concentration in the suspension, with lower concentration leading to more

individual airborne MWCNTs and higher concentration leading to more agglomerates. Two

types of MWCNTs were used in the study to provide more comprehensive data sets and to

demonstrate that the fractal-like characteristics reported here are not limited to MWCNTs of a

single type.


Table 5. Properties of Baytubes C 150 P by Bayer Material Science (24) and MWCNTs

TNM3 by Timesnano (125).

Property Baytubes C 150 P Timesnano TNM3 Value Unit Method Value Unit C-Purity > 95 % Elemental analysis > 95 % Free amorphous carbon Not detectable % TEM - - Number of walls 3 – 15 - TEM - - Outer mean diameter 13 – 16 nm TEM - - Outer diameter distribution 5 – 20 nm TEM 10 - 20 nm Inner mean diameter 4 nm TEM - - Inner diameter distribution 2 – 6 nm TEM 5 - 10 nm Length 1 – >10 μm SEM 10 - 30 μm Bulk density 140 – 160 kg/m3 EN ISO 60 220 kg/m3 Loose agglomerate size 0.1 – 1 mm Particle size

distribution - -

The intrinsic density was measured in two steps: a balance (Mettler Toledo AG, Switzerland)

to measure the mass of the MWCNTs and a gas displacement pycnometry system

(Micromeritics, AccuPyc II 1340) to measure the true volume occupied by the MWCNT walls.

Helium was used as the displacement medium because of its inertness and small atomic

number which helps the penetration into the hollow core in a MWCNT and pores in the

powder. Finally, the intrinsic density was calculated based on the mass and the true volume of

the MWCNTs.

Figure 30 shows a schematic diagram of the system for the mass measurement of airborne

MWCNTs. The functionalized MWCNTs were prepared with a concentration of 0.002 wt% in

water suspension, ultrasonicated for 30 minutes and then aerosolized by a homemade Collison

type atomizer. The same aerosolization system used in previous Chapters was used to

generate airborne MWCNTs. In order to determine the mass of the aerosolized MWCNTs,

they were first classified according to their mobility diameter using a DMA (model 3081, TSI,

USA) and then according to their mass to charge ratio using a CPMA (Cambustion, UK). The


particles exiting the DMA mostly carry one charge, thus the subsequent CPMA can select

particles according to their mass. The mass value corresponding to the highest particle

number concentration was determined to correspond to the selected particle size. The DMA-

CPMA tandem provided the mass of MWCNTs as a function of their mobility size. It is noted

that the particle number was used to identify the size-mass relationship, not the particle mass.

Therefore even if the mass distribution could be affected substantially by the multiply charged

particles due to the higher contribution to the mass by larger particles, they did not unduly

affect our results. In part of our experiments, the APM (Kanomax 3601, Japan) was used in

place of the CPMA for the mass measurements. Both instruments were calibrated using

standard PSL beads with known diameter and density to ensure the consistency of the mass

measurement. For example, Figure 31 shows the CPMA calibration using 200 nm diameter

PSL particles. Obtained particle mobility distribution by CPMA was compared with that by

SMPS and showed reasonable agreement.

Figure 30. A schematic diagram of the system for the mass measurement of airborne

MWCNTs. The CPC was used to quantify the classified particles.


Figure 31. CPMA calibration using standard PSL particle with 200 nm diameter. Obtained

distributions by SMPS and CPMA were compared.

The DMA-classified MWCNTs were collected on a piece of silicon wafer chip with a

nanometer aerosol sampler (model 3089, TSI, USA). Imaging was carried out using SEM

(Nova NanoSEM 230, FEI, Hillsboro, OR). Measurement of the MWCNT particle size was

performed using the image analysis software ImageJ (National Institutes of Health, USA).

MWCNT intrinsic density

The Baytube powder after the nitric acid refluxing treatment was measured 10 cycles by the

pycnometry system. The sample mass was 0.9986 g and the average volume was 0.4558 cm3,

giving rise to the intrinsic density of 2.19 g/cm3. This value was quite close to the theoretical

graphite density of 2.27 g/cm3, indicating that the surface –COOH groups did not

significantly impact the density. In addition, the nitric acid refluxing method could remove the

metal catalyst (57) thus the density was not affected by the metal catalyst. The manufacturer

reported true density of the Timesnano MWCNT was ~ 2.1 g/cm3, which was close to the


measured intrinsic density (2.19 g/cm3) of the Baytubes. For simplicity, ρi = 2.19 g/cm3 was

used in the model.

Effective density and mobility size relationship for airborne MWCNTs

The mass of aerosolized MWCNTs may be affected by the impurities in the aqueous

suspension. After evaporation of the aerosolized droplets, the impurities could attach onto the

MWCNTs. In our study the only possible impurities were from the DI water, since catalysts

for CNT synthesis were removed and cleaned during the refluxing process and functionalized

MWCNT suspension did not contain any surfactants. Thus the mass of aerosolized particles

of the DI water was measured to estimate the amount of impurities. The result showed that the

impurity mass corresponding to the mode (20.8 nm) of the impurity particle size distribution

was 0.0063 fg. On the other hand, the mode of the aerosolized CNT particle size distribution

was close to 80 nm as shown in Figure 1 and the mass was about 0.2 fg. Thus the mass of the

impurity was negligible compared to MWCNTs.

The measurement for airborne MWCNTs delivered the mass for the mobility size range from

50 nm to 500 nm, which are plotted in Fig. 32(a) and listed in Table 6. It can be seen in Fig.

32 that the data points including both APM and CPMA measurements fall nicely on fitted

curves. Details for the fitting curve, which based on mobility diameter, are presented in next

fractal model section. The effective densities corresponded to mobility diameters were listed

in Table 6. Discrepancies between effective densities and fitting curve for mobility diameters

< 150 nm are due to the higher fraction of individual MWCNTs than agglomerates in this

mobility size ranges as presented in Chapter 2.


(a) (b)

Figure 32. (a) The mass and (b) the effective density ρem as functions of the mobility size of

airborne MWCNTs, including both data from the APM and CPMA measurement.

Table 6. The mass m and the corresponding effective density ρem of airborne MWCNT

agglomerates with different mobility diameter dm. a data measured by APM; b data measured

by CPMA.

dm (nm) m (fg) ρem (g/cm3) dm (nm) m (fg) ρem (g/cm3)

50 0.0546a 0.834 200 2.4588a 0.587

80 0.200b 0.746 200 2.7227b 0.650

100 0.3382a 0.646 300 9.4012a 0.665

100 0.2985b 0.570 300 8.7650b 0.620

100 0.3503b 0.669 400 21.313a 0.636

120 0.5800b 0.641 400 20.5753b 0.614

150 0.9012b 0.510 500 41.2334a 0.630

4.2. Fractal model for CNT agglomerate

Theoretical model for fractal-like agglomerates of MWCNTs

The model treats MWCNTs as agglomerates of hollow tubes and establishes the relationship

between the MWCNT properties including tube diameters, length, intrinsic density and the

agglomerate properties such as the mass, porosity, effective density, surface area, and

characteristic diameters.


MWCNTs consist of multiple layers of rolled graphene sheets. In each layer, the carbon

atoms are arranged in a honeycomb lattice with separation of 0.142 nm, and the distance

between layers is close to that in graphite, approximately 0.335 nm (126). The graphite

density estimated from the above parameters and the carbon atomic weight of 12 Dalton is

about 2.27 g/cm3.

Figure 33. An example SEM image of a CNT agglomerate with illustrations for the tube inner

(di) and outer (do) diameters, and agglomerate outer (dout), mobility (dm) and volume

equivalent (de) diameters.

The MWCNTs with the outer diameter do, inner diameter di and the intrinsic density ρi, and

the total length of the all the MWCNTs in one agglomerate L (Figure 33) were considered, the

mass of this agglomerate is then

( )2241

ioi ddLm −π××ρ= . (4-1)

When m is measured, L can be obtained as:

( )22

4

ioi ddmL

−πρ= . (4-2)

This relationship holds regardless how many individual MWCNTs are in the agglomerates,

how they bend, coil or tangle. The number of layers of the MWCNT wall n is related to the

diameters by


( ) 2/)1( io ddn −=∆×− , (4-3)

where ∆ is the distance between two graphene layers, approximately 0.335 nm. The total tube

surface area S of all the MWCNTs in the agglomerate is

LdS oπ= , (4-4)

where the surface area of the tube ends is ignored. S provides the value of the available

surface exposed to the surrounding gas molecules or for interaction with cells upon MWCNT

deposition in lungs.

It was assumed that certain MWCNT agglomerates can be treated as fractal-like agglomerates

satisfying

fdckLm = , (4-5)

where Lc is a characteristic length of the MWCNT agglomerate, k is a proportionality

parameter, and df is the fractal dimension. The experiments delivered data sets for m of

airborne MWCNTs vs. the mobility diameter dm, which satisfied the fractal-like relationship

nicely:

fmdmmdkm = . (4-6)

The experimental data and values of the parameters are given in next result section. The

effective density based on the mobility diameter is

( )33

61

6 −

π=

π=ρ fmd

mm

mem dk

dm

. (4-7)

Combining (4-2) and (4-6), equation (4-8) is obtained as:

( )fmd

mioi

m ddd

kL 22

4−πρ

= . (4-8)

Equation (4-8) gives the relationship between the total length of MWCNTs in the agglomerate

and the mobility diameter of the agglomerate.


It is of interest to calculate the outer diameter dout of the agglomerate, i.e. the diameter of the

smallest circle which can enclose the whole agglomerate. For this end, the drag force on the

agglomerate was considered. The measurement of the electrical mobility of the agglomerate

indicates that the drag force Dg on the agglomerate is equal to that on a spherical particle with

the diameter dm, therefore

)(3

mc

mg dC

UdD πm= , (4-9)

where μ is the air viscosity, U is the relative velocity between the agglomerate and

surrounding air, Cc is the slip correction for spheres. On the other hand, the drag force can be

obtained by considering the drag on unit length of a fiber, Fd,

LFD dg = . (4-10)

Calculation of Fd is more complicated than the drag experienced by a circular fiber in a

perpendicular flow, because the MWCNTs have random orientations in the agglomerate, and

the neighboring tubes affect each other. This is analogical to fibrous filters, for which the drag

force and pressure drop are well studied. Simple force balance gives the relationship between

the pressure drop ∆p and Fd:

td

Fp d 24πα

=∆ , (4-11)

where α is the solid fractions of fibers in the filter, t is the thickness of the filter, and d is the

fiber diameter. With the analogy between the MWCNT agglomerate and a fibrous filter, d is

equivalent to the MWCNT outer diameter do.

One widely used equation for ∆p, obtained from empirical data and incorporating the effects

of random fiber orientations and neighboring interference, is given by Davis (127),

[ ] 235.1 /)561(64 dUtp α+αm=∆ for 0.006 < α < 0.3. (4-12)


When d is comparable or smaller than the air mean free path λ, the slip effect on the fiber

surface becomes important and a correction is needed. The correction derived by Pich (128) in

the transition regime approaching free molecular flow behaviors was used,

( )nc

n

c k

pppk

pp 998.0ln5.075.057.0+α−−

∆−∆

=∆ , (4-13)

where the Knudson number is kn = 2λ /d, p is the ambient pressure, ∆pc is the pressure drop

for the filter were it at the continuum flow regime and equation (4-12) was used for ∆pc. In

present case, the pressure drop induced by a MWCNT agglomerate in its normal motion is

clearly negligible compared to p, thus equation (4-13) was simplified and combined with

equations (4-11) and (4-12) to obtain

( ) [ ])561(64998.0ln5.075.04

57.0 35.1 α+αm+α−−α

π= Uk

kF n

nd . (4-14)

Now from (4-9), (4-10) and (4-14), equation (4-15) is obtained as:

( )[ ])(

3)561(64/996.1ln5.075.0/8

57.0 35.1

mc

mo

o dCdLd

d=α+αl+α−−

αl. (4-15)

With the DMA-CPMA measurement, dm and L are determined, and do of MWCNTs may be a

given, then the solid fraction α is the only unknown in (4-15). Once α is solved from (4-15),

the porosity is ε = 1 – α. The solid fraction is related to the outer diameter of the agglomerate

through

361

241

out

o

ddL

ππ

=α , (4-16)

which then leads to dout. Thus the relationship between the agglomerate mass and outer

diameter is obtained and fit into a power law:

fodoutodkm = , (4-17)


where ko and dfo are the proportionality parameter and fractal dimension based on the outer

diameter, respectively. The effective density based on the outer diameter is

( )33

61

6 −

π=

π=ρ fod

outo

outeo dk

dm . (4-18)

The dynamic shape factor χ is used to correct the drag force on a non-spherical particle (15),

defined as the ratio between the actual drag force Dg on the particle and the drag force on a

spherical particle having the same volume and velocity:

)(/3 ece

g

dCUdD

πmc = , (4-19)

where the volume equivalent diameter de is obtained from

iedm ρπ= 3

61 . (4-20)

Combining (4-9) and (4-19), χ is obtained for the MWCNT agglomerate

)()(

mce

ecm

dCddCd

=c . (4-21)

Furthermore, the aerodynamic diameter da is calculated as the diameter of a spherical particle

having unit density (ρ0 = 1 g/cm3) and the same terminal settling speed as the MWCNT

agglomerate, which can be important for determination of the agglomerate transport, e.g. the

impaction deposition and gravitational settling in the lung. The expression of da is as follows

cρρ

0

)()( ecieaca

dCddCd = . (4-22)

Mass and fractal-like relationship for airborne MWCNTs

The measured mass for the mobility size range from 50 nm to 500 nm are plotted in Figure 34.

The data laid on a power law curve and can fit into a fractal-like relationship,


9.27108.4 mdm −×= , (4-23)

where the unit of m is fg, and the unit of dm is nm. Surprisingly the exponent dfm is 2.9,

indicating a very compact, sphere-like morphology. However, it should be cautioned that the

mobility diameter is not a linear representation of the geometrical size of agglomerates as

shown in the next section. Thus the exponent dfm is higher than the fractal dimension based on

the geometrical outer diameter. In addition, the fractal relationship is intended to describe

self-similar patterns with increasing scale (129). This assumption was not satisfied by the

aerosolized MWCNTs, as the small particles might contain individual or very few MWCNTs

which were possibly relatively straight or lightly bent, whereas the large particles contained a

significant amount of MWCNTs which tangled heavily. Therefore, even though Eq. (4-23)

gives a truthful description of the experimental data for how the particle mass changes with

increasing size, it cannot be concluded that the small particles containing very few MWCNTs

possessed compact and sphere-like morphology. This is similar to agglomerates formed by

spherical primary particles, for which the fractal dimension obtained from the mass vs. size

relationship is usually not applicable to small particles composed of one single or very few

primary particles. In experiments, it has been observed that a substantial fraction of the

atomized MWCNTs with dm < 150 nm could be individual MWCNTs and possessed linear

structure as shown in Chapter 2, and even for dm of 200 – 500 nm, individual airborne

MWCNTs were obtained (60,62). It is clear that the diameter, aspect ratio and stiffness of the

MWCNTs play important roles for the agglomeration status. The Baytubes and Timesnano

tubes used in this study, had tube diameters in the range of 10 – 20 nm, thus they were more

flexible and prone to agglomeration formation. The MWCNTs in Seto et al. (60) and Wang et

al. (62) have tube diameters of about 65 and 85 nm, respectively, thus were more stiff and

likely to stay as individual fibers. The fractal-like relationship of Eq. (4-23) is derived for

MWCNTs with 10 – 20 nm diameter and may not be applicable for other CNTs. The high

value of dfm should not be applied to particles with individual or very few MWCNTs.


(a) (b)

Figure 34. (a) The mass and (b) the effective density ρem as functions of the mobility size of

airborne MWCNTs, including both data from the APM and CPMA measurement.

Modeling results for MWCNT agglomerates

The calculation is carried out based on the following parameters: the MWCNT outer diameter

do = 16 nm, the number of layers of the MWCNT wall n = 15, inner diameter di = 6.62 nm.

These parameters are consistent with the information provided by the manufacturer (Table 5).

The intrinsic MWCNT density used is ρi =2.19 g/cm3. The total MWCNT length L, the total

tube surface area S, the solid fraction in an agglomerate α, the outer diameter of an

agglomerate dout, the effective density based on the outer diameter ρeo, the dynamic shape

factor χ, the volume equivalent diameter de, and the aerodynamic diameter da are computed

according to the described model and are listed in Table 7.

m = 4.8E-07dm2.9

R² = 0.9974

0.0100

0.1000

1.0000

10.0000

100.0000

10.0 100.0 1000.0

m(f

g)

dm (nm)

APM dataCPMA data

ρem = 0.92dm-0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 100 200 300 400 500 600

ρ em

(g/c

m3 )

dm (nm)

APM dataCPMA data


Table 7. Modeling results of the total MWCNT length L, the total tube surface area S, the

solid fraction in an agglomerate α, the outer diameter of an agglomerate dout, the effective

density based on the outer diameter ρeo, the dynamic shape factor χ, the volume equivalent

diameter de, and the aerodynamic diameter da.

dm (nm) m (fg) L (nm) S (nm2) Α dout (nm) ρeo (g/cm3) χ de (nm) da (nm) 50 0.0546 149.5 7514 0.191 67.0 0.3465 1.83 36.2 47.5 80 0.2000 547.7 27531 0.166 108.2 0.3014 1.92 55.9 69.6

100 0.3290 901.0 45291 0.160 129.3 0.2909 2.09 65.9 76.1 120 0.5800 1588.4 79840 0.147 160.7 0.2667 2.03 79.7 93.7 150 0.9012 2468.3 124069 0.143 187.8 0.2599 2.26 92.3 99.3 200 2.5908 7095.5 356657 0.115 287.3 0.2087 1.95 131.2 156.4 300 9.0831 24876.4 1250426 0.089 475.3 0.1615 1.81 199.3 245.7 400 20.9439 57360.3 2883241 0.075 666.3 0.1352 1.76 263.3 325.5 500 41.2334 112928.0 5676381 0.064 880.2 0.1155 1.71 330.0 411.6

The total length L increases rapidly with the mobility size. It is only about 150 nm when dm =

50 nm, but approaches 1 μm when dm = 100 nm, and is almost 113 μm when dm = 500 nm.

This is due to the fact that L is proportional to the mass m, thus L has a power law relationship

with dm with an exponent of 2.9. The outer diameter dout is larger than dm and their

relationship can be fitted to a power law

1.173.0 mout dd = , (4-25)

as seen in Figure 35. This relationship is valid in the range of 50 ≤ dm ≤ 500 nm. Similar

relationship between the outer and mobility diameters of agglomerates composed of primary

spheres has been observed. Kim et al. (130) investigated silver agglomerates produced by

tandem-furnace system and reported the exponent in Eq. (4-25) to be 1.268 and 1.088 under

two different conditions.


Figure 35. The total MWCNT length L, the outer diameter dout and aerodynamic diameter da

of an agglomerate as functions of the mobility diameter dm.

Figure 35 shows that L is slightly larger than dout for small particles around dm = 50 nm,

indicating these MWCNTs are relatively straight or slightly bent. For larger agglomerates, the

ratio between L/dout rapidly increases, indicating that the MWCNTs are significantly coiled

and bent. At dm = 500 nm, a total length of 113 μm is packed in an agglomerate with dout =

880 nm. However, the solid fraction α is actually low and it decreases with the agglomerate

size, from 19.1% at dm = 50 nm down to 6.4% at dm = 500 nm by Eq. (4-16). The decrease of

α is expected for fractal-like agglomerates, since the fractal dimension is less than 3 and the

increase of the solid volume cannot be as fast as the agglomerate volume. The solid fraction

around 6% is commonly encountered in agglomerates of primary spheres (131). The solid

fraction α for all the measured particles is within the applicable range of the pressure drop

equation (4-12).

The aerodynamic diameter da is larger than de, but slightly smaller than dm. The ratio between

dm and the volume equivalent diameter de is in the range 1.3 to 1.6 for 50 ≤ dm ≤ 500 nm,

which leads to the dynamic shape factor χ in the range 1.7 to 2.3 (Eq. (4-21)). In comparison,

dout = 0.73dm1.1

R² = 0.9947

10

100

1000

10000

100000

1000000

10 100 1000

leng

th (n

m)

dm (nm)

Total length L

Outer diameter dout

Aerodynamic diameter da


Chen et al. (23) obtained the dynamic shape factor χ in the range 1.94 to 2.71 for fiber-like

MWCNTs by aerodynamic measurement.

The relationship between the agglomerate mass and outer diameter was considered in the

study. The data are plotted in Figure 36(a) and can be fit into the following power law

6.26103.1 outdm −×= . (4-26)

The fractal dimension based on the outer diameter is dfo = 2.6, smaller than that based on the

mobility diameter dfm = 2.9. The two fractal dimension values are related by the exponent in

Eq. (4-25), which is 1.1. The value of dfo = 2.6 shows that the MWCNT agglomerates have a

rather compact geometrical structure, thought it is distinguishable from spheres. Interestingly,

the value of dfo agrees well with the fractal dimension of bulk CNT assemblies. Staszczuk et

al. (19) used thermo-gravimetric analysis, gas adsorption, and atomic force microscopy to

measure several CNT samples grown from xylene-ferrocene reaction and obtained the fractal

dimension in the range of 2.33 to 2.64. Smajda et al. (40) used gas adsorption and reported the

fractal dimension 2.49 for their MWCNT assemblies prepared by filtration. Tripol’skii et al.

(41) used electron microscopy to analyze CNTs produced by the thermal decomposition of

ethylene and obtained fractal dimension of 2.65 ± 0.05. It seems the fractal dimensions around

2.4 – 2.6 represent a rather stable agglomeration structure for CNTs.

(a) (b) Figure 36. (a) The mass and (b) the effective density ρeo as functions of the outer diameters of

airborne MWCNT agglomerates.

m = 1.3E-06dout2.6

R² = 0.9995

0.01

0.1

1

10

100

10 100 1000

m(f

g)

dout (nm)

ρeo = 2.4dout-0.4

R² = 0.9822

0

0.1

0.2

0.3

0.4

0 200 400 600 800 1000

ρ eo

(g/c

m3 )

dout (nm)


The effective density based on the outer diameter is shown in Figure 36(b) and the data can be

fit to

4.04.2 −=ρ outeo d . (4-27)

ρeo is about 0.35 g/cm3 for the small MWCNTs with dout around 60 nm, then decreased to

about 0.11 g/cm3 for the 880 nm agglomerates. The bulk density of Baytubes is 0.14 – 0.16

g/cm3 (Table 5), comparable to those of the airborne 400 – 700 nm agglomerates. Thus the

smaller airborne agglomerates have more compact structures and lower porosities than the

bulk Baytube powder, whereas the larger ones can be more porous than the bulk powder.

Microscopic studies

Collection and imaging was carried out on DMA classified airborne CNT particles.

Figure 37 shows typical SEM images of CNT particles with dm = 200, 300, 400 and 500 nm.

At 200 nm, some straight, branched or coiled chain-like CNT particles exist; the fraction of

such structures diminishes with the increasing mobility size. At 400 and 500 nm, all the CNT

particles have compact structures with individual CNTs occasionally sticking out the

agglomerate surface. These images support the high fractal dimension dfo = 2.60 from the

model, though the 2D images do not provide direct comparison for the 3D dfo. The

distributions of the outer diameters of the CNT particles measured from the SEM images are

shown in Table 8 for different mobility sizes. The values of dout from the model are within

half of the standard deviation from the mean measured dout values and the relative error

ranges from -12% to 12%.

The 2D image fractal dimension can be obtained using the box algorithm. The analyzed

fractal dimensions for CNT particles with different mobility diameters are shown in Table 8,

and the value increases from 1.73 for dm = 200 nm to 1.83 for dm = 500 nm. The 2D fractal

dimension has been measured in the literature for bulk CNT assemblies, e.g, 1.82 in Smajda


et al. (40) and 1.65 in Tripol’skii et al. (41). These values are in good agreement with our

image analyses for aerosolized CNT agglomerates.

200 nm

300 nm

400 nm

500 nm

Figure 37. Typical SEM images of the DMA classified CNT particles with dm = 200, 300, 400

and 500 nm.


Table 8. Measured agglomerate outer diameter and 2D image fractal dimension for CNT

particles with different mobility diameters (around 50 sample images were analyzed for each

mobility size).

Mobility diameter (nm) 200 300 400 500 Mean value of the measured outer diameter (nm) 325.8 496.8 609.4 785.7

Standard deviation for the measured outer diameter (nm) 104.9 101.3 122.2 238.3

Calculated outer diameter (nm) 287.3 475.3 666.3 880.2 Relative error between the measured and

calculated outer diameter -12% -4% 9% 12%

Measured fractal dimension 1.73 1.77 1.81 1.83 Standard deviation for fractal dimension 0.06 0.04 0.05 0.04

84 Bahk et al. (2013) Aerosol Sci. Tech. 47(7):776-734; Bahk et al. (2014) Aerosol Air Qual. Res. 14(5):1352-1359;

Chen, Wang, Bahk, Fissan, Pui (2014) Aerosol Sci. Tech. 48(10):997-1008

5. Emission control: Filtration of air-borne CNTs

Filtration is an effective method to control nanoparticle emissions, thus NIOSH (132)

suggested the respirators as the final personal protective equipment for handling

nanomaterials. A NIOSH survey found that 77% carbonaceous nanomaterial companies were

using respirator to mitigate the exposure of workers (63,133). However, the performance of

filter media was usually determined by challenging with spherical or sphere-like particles

such as dioctylpthalate (DOP) and sodium chloride (NaCl) (134-138). Therefore, Vo and

Zhuang (63) of NIOSH studied the penetration of pristine MWCNTs aerosolized by a

nebulizer through six commercialized facepiece respirators (two of each were NIOSH

approved N95, N99, and N100). The authors concluded that the penetration of the elongated

MWCNTs was lower than the designated, or the allowed penetration of these respirators. This

finding was in agreement with the results for elongated caffeine rods in electret filters (139)

and for CNTs in mechanical filters (60-62,99). The reduced penetration was attributed to the

higher interception deposition efficiency of the MWCNTs than the sphere-like particles.

However, the details of the deposition mechanisms, such as the effects of shape (bending and

curling), and length in mechanical filters, of the MWCNTs have not been investigated.

In filtration studies evaluating the effectiveness of filter and understand filtration mechanisms,

various different filters have been tested against spherical nanoparticles and elongated

particles such as asbestos fiber and CNTs (60,61,63,64,118,140). Wang et al. (61) evaluated

stainless screens, known as a diffusion battery, in filtration tests against airborne CNTs. They

performed numerical calculations of penetration of CNTs through the filter screen to

understand the capture mechanisms of elongated particles by the screen. The results showed

the importance of geometrical length in the filtration study of elongated particles. The

dependence of interception increases due to longer effective interception lengths of elongated

particles than those of spherical particles. Nanofiber filter media has been introduced as a

85 Bahk et al. (2013) Aerosol Sci. Tech. 47(7):776-734; Bahk et al. (2014) Aerosol Air Qual. Res. 14(5):1352-1359; Chen, Wang, Bahk, Fissan, Pui (2014) Aerosol Sci. Tech. 48(10):997-1008

promising media, which possess a higher efficiency for submicron particles due to their larger

specific surface area to collect and a lower pressure resistance for air flow than conventional

filter media. The most common method to produce the nanofibers is using electrospinning

method, which can generate fibers with diameters from several nanometers to a few

micrometers depending on fabrication conditions (118,141-143).

In the present study, the stainless screens with 20 µm wires and nanofiber media with fiber

diameters in the range 150-300 nm were investigated against airborne CNTs to understand

their capture mechanism. Furthermore, a mechanistic study was performed to measure the

penetration of CNTs through the Nuclepore filter and it can provide complementary

information to understand the results of the respirator media.

The single fiber theory was widely used for predicting the particle penetration in the fibrous

filtration. However, discrepancies between the prediction and data were commonly observed

(60,144,145). For example, Kim et al. (145) found about 15% penetration difference between

the theory and data at most penetrating particle size (MPPS) for silver spheres penetrating

through the H&V HE1073 fiberglass filter (Hollingsworth & Vose, East Walpole, MA, USA).

The authors believed that microscale inhomogeneity and random orientations of fibers in the

filters were important reasons for the discrepancy. Besides, a discrepancy was also found by

Seto et al. (60) in the MWCNT filtration through a fiberglass filter (2.8 mm fiber diameter,

0.049 solidity, and 0.38 mm thickness). They introduced orientation angles for the MWCNTs

in the theory for fulfilling the comparison but the assumption needed independent verification

and the authors suggested investigating the rotation dynamics of the MWCNTs in the

filtration with fibrous filters in the future. Therefore, there remain research interests to clarify

the reasons for the discrepancy. In this study, Nuclepore filter was used to study the

mechanical (diffusion, interception, and impaction) deposition mechanisms of MWCNTs,

since the Nuclepore filters possess uniform and well-defined microstructures, which can be



accurately described by a modified capillary tube model (115,116). The modified capillary

tube model predicted the agglomerate penetration through Nuclepore filter very well.

Therefore, usage of the Nuclepore filter allows this study to clarify and further understand the

mechanical filtration mechanism of MWCNTs.

A number of studies regarding the characterization and synthesis of CNTs have shown useful

results for a lot of practical applications, however, a few studies concerned CNT deposition,

transport behavior in liquid and filtration through porous media (146,147). Although the

exposure to the wet engineered nanoparticles such as CNTs in liquid is rarer than that to

emitted dry nanoparticles (148), previous studies showed toxicity of CNT suspension in

specific conditions (15,149-151). Therefore, control of liquid-borne CNT emissions is also

important to prevent releasing to the environment such as soil and groundwater which possess

potential to make secondary exposure situations. The filtration is also an attractive technique

to control the particle emissions to the environments for suspended particles in liquid (152-

155). It is widely used in different areas such as chemical engineering, pharmaceutical, food

and beverage industries and for water treatment of drinking water (156-158). Since clean

water is important not only for the environmental aspects, but also for industries such as

pharmaceutical and semi-conductor industries (159). Here, filtration studies for liquid-borne

MWCNTs through the Nuclepore filter were performed and the results were presented in

Appendix. Two filtration systems using a filtration flask or peristaltic pump and electrolytes

effects on the filtration efficiencies were investigated.


Figure 38 shows a schematic diagram of CNT filtration system. The same experimental set-up

used in filtration method session to determine the geometrical length of MWCNTs, Chapter

3.2, was employed to perform the filtration tests. The controlled face velocity on the filter was


controlled and in most of the cases the velocity was 5 cm/s. The stainless screens, 4 different

nanofiber filters and Nuclepore filter were tested against airborne MWCNTs and detailed

information for filters can be found in Chapter 3.2.

Different size ranges for airborne MWCNTs were tested for different filters. For instance,

with stainless screens and nanofiber filters, 65-200 nm and 75-200 nm mobility size ranges

were chosen for Baytubes and Cheaptubes, respectively, to test single standing CNTs without

residual particles and agglomerated CNTs. Size range from 20 nm to 500 nm were chosen for

the Nuclepore filter to investigate the effects of both single standing CNTs and CNT

agglomerates. In order to compare the results of CNT filtration with those of spherical

particles’, PSL particles with 51.4, 95.6 and 193.3 nm diameters, were used for the screens

and nanofiber samples. NaCl was used as sphere-like nanoparticles for the Nuclepore filter

tests. The obtained penetrations of CNTs were compared with the model calculation and also

those of PSL particles.

Figure 38. Experimental system for filtration tests.



Filtration of airborne MWCNTs

PSL particles were adopted to verify the filtration model used in this study. Penetrations of

PSL particles were obtained by measuring particle concentrations upstream and downstream

of the filter. The comparison between penetrations calculated by the model and measured in

experiments showed reasonable agreement as shown in Figure 39. PSL particles with three

different sizes, such as 50 nm, 95 nm and 190 nm, were used in the experiments and tested

against the metal screens and nanofiber sample A. In order to compare two different kinds of

filters, the pressure drop for each filter was measured. Nanofiber filters showed a relatively

lower pressure drop than the screens showed and obtained values were 36 Pa and 67 Pa for

nanofibers and screen filters, respectively. Although nanofiber sample A possesses lower

solidity (0.134) than the diffusion screen (0.345), higher filtration efficiency was obtained

with 95nm and 190 nm PSL particles. In the equations of single fiber efficiency, the

efficiency due to the interception increases when the fiber size decreases. Thus the nanofiber

filters showed higher efficiency than screens in the interaction regime, because of their

smaller fiber size. The results showed a good agreement with the theory, which represents the

relations between the fiber size and the minimum efficiency and most penetrating particle size

(MPPS). The minimum efficiency increases and the MPPS decreases when the fiber size

decreases (21,160). The calculated MPPSs with the experiment conditions were 487 nm and

96.1 nm for the screens and nanofiber filter sample A, respectively. The equation for MPPS

was derived by Lee and Liu (160) and expressed as:

�̂�𝑑p = 0.885�� 𝐾𝐾𝐾𝐾1−𝛼𝛼

� �√𝜆𝜆𝑘𝑘𝑘𝑘𝜇𝜇� �

𝑑𝑑𝑓𝑓2

𝑈𝑈0��

2 9⁄

, (5-1)

where k is Boltzmann constant and T is the temperature.


Figure 39. Penetrations and model calculations of PSL particles with 50nm, 95 nm and 190

nm diameters through the screens and nanofiber sample A.

The screens and nanofibers were also tested against Cheaptubes and Baytubes and the

obtained penetrations by the DMA and CPC measurement upstream and downstream of the

filters were compared with the filtration models introduced in Chapter 3. The comparisons in

Figures 40 and 41 show reasonable agreements in smaller mobility size range, however,

measured penetrations for the larger size end in the range showed bigger discrepancy with the

calculated ones. Although fraction of single standing CNTs were high in smaller mobility size

range, main deposition on the filter was due to diffusion, which is dependent on the mobility

size, for smaller size particles. It may lead good agreement with the models. On the other

hand, effective interception lengths were important for MWCNT deposition in larger size

range, but a fraction of MWCNT agglomerates, bended and curled MWCNTs increases when

the mobility size of MWCNT is above 150 nm as shown in Chapter 2-4. It might be another

reason to see the discrepancies in both models. Because of longer effective interception length



of MWCNTs than that of spherical particle, MWCNTs showed lower penetration than

spherical particles in the entire range, especially, in the larger mobility size range. In addition,

penetrations of MWCNTs in larger size range showed bigger discrepancies between two

filters than that of spherical particles. Similar with PSL particles’ results, nanofiber filter

showed better effectiveness with lower pressure drop than metal screens due to higher chance

of interception collection than the metal screens. Even if the nanofiber filter possesses lower

solidity, the nanofiber filter contents more fibers than metal screens. It can increase the

chance of CNT depositions through interception mechanism.

Figure 40. Penetrations and model calculations of Baytubes with mobility sizes from 65nm to

200 nm through the screens and nanofiber sample A.


Figure 41. Penetrations and model calculations of Cheaptubes with mobility sizes from 75nm

to 200 nm through the screens and nanofiber sample A.

Nanofiber filter samples with four different solidities were tested against airborne MWCNTs

and PSL particles. The size range, we chose for the experiments, was already in the transition

regime from the diffusion dominant regime to interception dominant regime, and the smaller

size end could be more dependent on diffusion, the other end could be more dependent on

interception. In the transition regime, the higher solidity samples showed higher collection

efficiency for airborne CNTs as shown in Figure 42. Penetrations of both types of CNTs

showed lower values than those of PSL particles due to the longer effective lengths of CNTs

for interception than diameters of PSL particles, which were used as the effective interception

length for spheres. It is also explainable by the single fiber theory. The particles possessing

longer effective lengths can have more chances to be collected by the nanofibers. The lower

solidity sample possessed lower fraction of nanofibers, thus discrepancy between penetrations



of CNTs and PSL particles was decreased. It can be seen that discrepancies were decreased

with decreasing solidities of filter samples in the figures. Same penetrations were obtained

with sample D for all particles used in the experiments, since it possessed lower solidity and

the expected effect of nanofibers was insignificant.

Figure 42. Penetrations of Baytubes, Cheaptubes and PSL particles with a mobility size range

from 50nm to 200 nm through different nanofiber samples A to D with different solidities

0.134, 0.104, 0.059 and 0.034, respectively.

In order to better understand the mechanical filtration mechanisms of MWCNTs, Nuclepore

filters and the corresponding capillary tube model were adopted. Figure 43 shows the

comparison of experimental MWCNT penetrations with the theoretical results for spherical

particles (Spurny sphere model) at different face velocities from 5 to 25 cm/s. Through the

comparison with spherical particles more information on filtration dynamics of the MWCNTs

could be obtained. In the model, the effective length of MWCNT determined from SEM

analysis in Chapter 3 was used as the interception length.

The tested face velocities higher than 10 cm/s were for examining whether the alignment of

MWCNTs occurred in the filtration and the results showed that it occurred. To be noted,


alignment effect was not observed for soot and silver agglomerates at the face velocity lower

than 7.5 cm/s (145), therefore, it is expected that the MWCNTs also would not experience the

effect at the face velocity lower than 7.5 cm/s. For instance, the penetration for 50 nm

MWCNTs was very close to that of spherical particles at 5 cm/s, which was attributed to that

diffusion mechanism dominated the deposition while interception was much weaker. The

penetration difference then increased with increasing face velocity at 10 cm/s when diffusion

efficiency decreased, resulting the observation of the penetration difference. As the face

velocity was increased further to 15 and 25 cm/s the difference was decreased to 52.1 and

41 %, respectively. This is attributed to the occurrence of alignment. The alignment effect,

similar trend that found for 50 nm MWCNTs, was also observed for 80 and 100 nm

MWCNTs. However, the alignment was not significant for larger size MWCNTs due to a

high level of agglomeration as shown in Figure 37. These MWCNT agglomerates did not

have as high aspect-ratio as smaller ones therefore the alignment effect was not obvious.

The comparison of the penetration of MWCNTs with that of spheres showed that they are

close at only 30 nm when diffusion mechanism dominated the particle collection. The

penetration difference increased with increasing particle size from 30 nm to about 200 nm and

then decreases slightly to 500 nm. This was due to the enhanced interception deposition of

MWCNTs. It is also seen the penetration of MWCNTs increased with increasing face velocity;

at the same time, the most penetration particle size shifted to smaller particle size.



Figure 43. Comparison of experimental and theoretical penetrations of MWCNTs in

Nuclepore filters at different face velocities. Penetrations of spheres were also shown for

comparison purpose with MWCNTs.

95

6. Exposure modeling: Lung deposition model for air-borne

CNTs

Aerosol deposition in the human airway is significant not only in occupational health,

but also in pharmaceutical research for drug delivery to the deep inside the lung. Airborne

particle depositions have been estimated either empirically or numerically, however, due to

ethical reasons, investigations are limited to laboratory animal tests. For this reason, a number

of studies have been published regarding developments of lung deposition models for

particles in the human airway into the lung. Yeh and Schum (67) have developed the model of

human lung airways and the model led to the accurate prediction of regional particle

deposition in the lung. A multiple path model for particle depositions in the rat lung has been

developed by Anjilvel and Asgharian (68) and they have extended the model to the particle

deposition in the human lung (161). They also showed the models for deposition of inhaled

fibrous particles such as asbestos in both rat and human lung (162,163). Ding et al. (69) has

revised Asgharian’s model and proposed combined deposition of impaction and interception

of fibers in the nasal region.

CNTs possess a cylindrical shape with a high aspect ratio as asbestos fibers. Thus

toxicological concerns on CNTs lead interests in CNT deposition in the human lung. Several

studies showed that CNT has cytotoxicity and causes inflammations in the lung cell.

96

Furthermore, it has potential toxicity effect on lung cancers (7-13,15). Recent studies have

presented numerical calculations of CNT deposition in the lung using different deposition

models. For instance, Strum (65) introduced the theoretical approach for a risk assessment of

carbon nanotubes in the human alveoli by employing the model using aerodynamic diameter.

On the other hand, Erderly et al. (64) used a mouse model to assess the workplace exposure of

CNTs for an inhalation toxicological study. For an empirical estimation of inhaled CNTs, Su

and Cheng (66) used a human respiratory tract replica and obtained the local deposition

fraction of CNTs in the airways in the lung. Existing studies regarding model calculation of

CNT deposition have considered CNTs as a single standing cylindrical particle in the models.

However, presented results in Chapters 2, 3 and 4 showed that CNTs can exist in a form of

either singly standing tubes or agglomerates, or both formations at the same time. Thus the

fraction of agglomerate state CNT particles in CNT samples was investigated and obtained by

the SEM analysis. The obtained fraction was applied in the lung deposition model, which

includes revised models of previous studies.

Lung deposition model for CNTs

The lung structures used in different studies were all based on the morphometric measurement

by Raabe et al. (164) and Yeh and Schum (67) presenting results for simplified geometries of

entire lung. The morphological parameters, such as length, diameter, branching angle, number

97

of airway tubes and gravity angle, from Yeh’s study, were employed in the present study. The

used data are presented in Table 9 and it includes information of every generation from

respiratory bronchioles, alveolar ducts to alveoli. As shown in Figure 44, the human lung has

been described as two main assembled parts, i.e. the conducting zone and transitional and

respiratory zones, and they also termed tracheobronchial (TB) and alveolar region,

respectively. The tracheobronchial includes generations from 1 to 16 and alveolar region

includes generations from 17 to 23.

Figure 44. Weibel’s lung diagram which describes bronchiole and alveolar generations of the

lung. (165)

98

Table 9. Typical path Lung Model for human (67)

Generation (n)

Number of tubes L (cm) dair (cm) θ φ

1 1 10 2.01 0 0

2 2 4.36 1.56 33 20

3 4 1.78 1.13 34 31

4 8 0.965 0.827 22 43

5 16 0.995 0.651 20 39

6 32 1.01 0.574 18 39

7 64 0.890 0.435 19 40

8 128 0.962 0.373 22 36

9 256 0.867 0.322 28 39

10 512 0.667 0.257 22 45

11 1024 0.556 0.198 33 43

12 2048 0.446 0.156 34 45

13 4096 0.359 0.118 37 45

14 8192 0.275 0.092 39 60

15 16384 0.212 0.073 39 60

16 32768 0.168 0.060 51 60

17 65536 0.134 0.054 45 60

18 131072 0.120 0.050 45 60

19 262144 0.092 0.047 45 60

20 524288 0.080 0.045 45 60

21 1048576 0.070 0.044 45 60

22 2097152 0.063 0.044 45 60

23 4194304 0.057 0.043 45 60

The lung deposition model contains depositions in head (Nasal), tracheobronchial and

alveolar regions. The particle depositions in the airways are mainly caused by diffusion,

99

sedimentation, impaction and interception. For single standing CNT deposition, theory for the

fiber deposition was taken into account in the present model.

Due to the complex structure and lack of knowledge on the anatomical structure and flow

field in the nasal region, prediction of deposition in nasal region was obtained empirically.

Zhang and Yu (166) have established the particle deposition efficiency in nasal region based

on experimental data from Raabe et al. (167,168). The deposition efficiency Nimp due to

impaction is expressed as:

𝑁𝑁𝑖𝑖𝑖𝑖𝑖𝑖 = � �𝑑𝑑𝑎𝑎2𝑄𝑄�𝑎𝑎

𝑐𝑐+�𝑑𝑑𝑎𝑎2𝑄𝑄�𝑎𝑎�𝑏𝑏, (6-1)

where a, b and c are constants and da is the aerodynamic diameter of particles in µm and Q is

the air flow rate in cm3/s. a and b for human are 1.257 and 0.609, respectively, c is 10000 for

all animals. Q was obtained from average weight of human, 70 kg, using Guyton formula

(169) and the formula was explained as:

𝑄𝑄 = 0.15 𝑊𝑊34, (6-2)

where W is weight of mammal in grams.

For fiber deposition in the nose, the effect of fiber orientation on particle Stokes number was

considered and applied to equation 6-1.

𝑁𝑁𝑖𝑖𝑖𝑖𝑖𝑖 = � (𝜒𝜒)𝑎𝑎

𝑐𝑐+(𝜒𝜒)𝑎𝑎�𝑏𝑏, (6-3)

where

100

𝜒𝜒 = 13

(𝑏𝑏1 + 2𝑏𝑏2)𝜌𝜌𝑓𝑓𝑖𝑖𝑏𝑏𝑓𝑓𝑓𝑓𝑑𝑑𝑣𝑣2𝑄𝑄𝑖𝑖, (6-4)

where dv is the fiber equivalent volume diameter, ρCNT is the CNT density, Qi is the

inhalation flow rate, b1 and b2 are the normalized fiber mobilities for parallel and

perpendicular movement to the flow direction, expressed as:

𝑏𝑏1 = 34

𝛽𝛽13�2𝛽𝛽

2−1

�𝛽𝛽2−1ln�𝛽𝛽+�𝛽𝛽2−1�−𝛽𝛽�𝑐𝑐1

𝛽𝛽2−1, (6-5)

𝑏𝑏2 = 38

𝛽𝛽13�2𝛽𝛽

2−1

�𝛽𝛽2−1ln�𝛽𝛽+�𝛽𝛽2−1�+𝛽𝛽�𝑐𝑐2

𝛽𝛽2−1, (6-6)

where β is the aspect ratio of CNTs and slip coefficients c1 and c2 were derived by Asgharian

and Yu (170) and 1.142 and 0.558, respectively.

The deposition efficiency Nimp+int due to combined mechanisms of impaction and interception

was derived by Ding et al. (69) and written as:

𝑁𝑁𝑖𝑖𝑖𝑖𝑖𝑖+𝑖𝑖𝑖𝑖𝑖𝑖 = � (𝜒𝜒+3.5𝐿𝐿𝐶𝐶𝐶𝐶𝐶𝐶)𝑎𝑎

𝑐𝑐+(𝜒𝜒+3.5𝐿𝐿𝐶𝐶𝐶𝐶𝐶𝐶)𝑎𝑎�𝑏𝑏, (6-7)

where LCNT is the length of CNTs.

For very fine particles as CNTs in the chosen range, 50-500 nm, in the present study, the

diffusion effect is dominant for deposition. Cheng et al. (171) proposed the expression for the

particle deposition due to diffusion as:

𝑁𝑁𝑑𝑑𝑖𝑖𝑓𝑓 = 1 − exp (−𝐶𝐶𝐷𝐷12𝑄𝑄𝑖𝑖

−18), (6-8)

101

where C is a constant and was obtained by Gerde et al. (172) as 12.65 for human species. D is

a particle diffusion coefficient and for CNT can be expressed as:

𝐷𝐷 = 𝑘𝑘𝑘𝑘3𝜋𝜋𝜋𝜋𝑑𝑑𝑣𝑣

�13𝑏𝑏1 + 2

3𝑏𝑏2�, (6-9)

in which k is the Boltzmann constant, T is the absolute temperature, µ is the viscosity of air.

The total efficiency Nnasal in the nasal region is expressed as the following:

for spherical particles,

𝑁𝑁𝑖𝑖𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 = 1 − �1 −𝑁𝑁𝑖𝑖𝑖𝑖𝑖𝑖��1 −𝑁𝑁𝑑𝑑𝑖𝑖𝑓𝑓�, (6-10)

for CNTs,

𝑁𝑁𝑖𝑖𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 = 1 − �1 −𝑁𝑁𝑖𝑖𝑖𝑖𝑖𝑖+𝑖𝑖𝑖𝑖𝑖𝑖��1 − 𝑁𝑁𝑑𝑑𝑖𝑖𝑓𝑓�. (6-11)

For airway in the lung, deposition due to sedimentation is considered additionally. The

deposition model shares equations for tracheobronchial and alveolar region with different

input parameters. The deposition efficiency due to diffusion was derived by Ingham (173) as:

η𝑑𝑑 = 1 − 0.819 exp(−14.63∆) − 0.097 exp(−89.22∆) − 0.0325 exp(−228∆) −

0.059exp (−125.9∆2/3), (6-12)

in which,

∆= 𝐷𝐷𝐿𝐿4𝑈𝑈𝑅𝑅2

, (6-13)

where L and R are the airway length and radius, respectively, U is the average flow velocity in

the airway and D is the diffusion coefficient of CNTs.

102

The deposition efficiency in the airways due to sedimentation adopting Pich’s expression (174)

was derived as:

η𝑛𝑛 = 2𝜋𝜋�2𝜀𝜀�1 − 𝜀𝜀

13 − 𝜀𝜀

13�1 − 𝜀𝜀

23 + sin−1 𝜀𝜀

13�, (6-14)

in which,

𝜀𝜀 = 38𝐿𝐿𝑅𝑅𝑉𝑉𝑠𝑠𝑈𝑈

, (6-15)

where Vs is the terminal settling velocity of CNTs and expressed as:

𝑉𝑉𝑛𝑛 = 𝜌𝜌0𝑑𝑑𝑎𝑎,𝐶𝐶𝐶𝐶𝐶𝐶2 𝑔𝑔𝑔𝑔𝑐𝑐18𝜋𝜋

, (6-16)

where ρ0 is unit density, da,CNT is the aerodynamic diameter for CNT, g is the gravitational

acceleration and Cc is the Cunningham slip correction.

Cai and Yu (175) derived the deposition efficiency for impaction by:

η𝑖𝑖 = 8𝑛𝑛𝑖𝑖𝑖𝑖𝑠𝑠𝑓𝑓1𝑅𝑅/𝑅𝑅0𝑓𝑓0

𝑆𝑆𝑆𝑆𝑆𝑆, (6-17)

where

𝑓𝑓0 = π − �π4

+ �54𝜋𝜋 − 8

3� cos2 𝛼𝛼� � 𝑅𝑅

𝑅𝑅0�2, (6-18)

𝑓𝑓1 = 1 + �− 13

+ �𝜋𝜋 − 113� cos2 𝛼𝛼 − 𝑛𝑛𝑖𝑖𝑖𝑖𝑠𝑠

3� � 𝑅𝑅

𝑅𝑅0�2

+ ��23− 𝜋𝜋

8� cos2 𝛼𝛼 + sin2 𝑠𝑠

5+ �6 −

158𝜋𝜋� cos4 𝛼𝛼 + � 7

15− π

8� sin2 𝛼𝛼 cos2 𝛼𝛼� � 𝑅𝑅

𝑅𝑅0�2, (6-19)

and α is the bifurcation angle and Stk is the particle Stokes number which is given by

Asgharian and Anjilvel (163) as:

103

𝑆𝑆𝑆𝑆𝑆𝑆 = 1.142𝜌𝜌𝑑𝑑𝑣𝑣2𝛽𝛽𝑈𝑈36𝜋𝜋𝑅𝑅0

. (6-20)

The total efficiency ηtotal can be expressed as:

η𝑖𝑖𝑡𝑡𝑖𝑖𝑛𝑛𝑛𝑛 = 1 − (1 − η𝑑𝑑)(1 − η𝑛𝑛)(1− η𝑖𝑖). (6-21)

In the calculations for CNTs, obtained physical parameters in Chapters 3 and 4 were used.

Estimation of particle deposition in the airways to the human lung

Particle depositions in the airways to the human lung were calculated by the presented models

in the previous section with different particles such as spherical particles with unit density,

CNT agglomerates, single standing CNTs and particle model for mixture of agglomerated and

single standing CNTs. Three different breathing scenarios such as sitting breathing, light

working breathing and heavy working breathing were applied and the air flow rates for each

condition were 646, 1291 and 1937 ml/s, respectively. The particles mobility size range from

50 nm to 500 nm was chosen for estimation, because the typical size distribution of used

airborne CNTs in the present study showed the same size range and the size range includes

the size ranges containing only single standing CNTs or CNT agglomerates.

Particle deposition in each airway generation was obtained by varying the breathing scenarios

and sizes of particles. The results were shown in Figures 45, 46 and 47. Particles with smaller

sizes than 100 nm showed almost same depositions for all particles up to 10th generation. It

104

can be explained by that small nanoparticles are more dependent of diffusion than other

deposition mechanisms and sizes of airways in these generations are much larger than

nanoparticles, thus sedimentation and impaction affect less. The deposition fraction increases

when the generation increases for entire size range and three different breathing scenarios

with spherical particles and CNTs. The spherical particles and CNT agglomerates deposited

more than single standing CNTs in higher airway generations, which means they possess a

higher possibility to deposit deep inside the lung. Orientation and geometry concerns for

CNTs in the model led the lower collection efficiency of single standing CNTs in the airway

generations in the lung. As presented in Chapter 5, single standing CNTs can align in the

direction of air flow into the lung when flow velocity is high. For example, up to 14th, 16th

and 17th generation in the sitting breathing, light working and heavy working scenarios,

respectively, the velocity in airway is above 10 cm/s, which the velocity showed alignment

effects of CNTs in air flow. Another reason for the lower deposition efficiency of single

standing CNTs is the lower effective density of CNTs and it affects to deposition due to

impaction in the lung airway.

Similarly, smaller particles as 50 nm size particles showed the highest deposition amount in

entire airways and relatively higher than those of larger size particles. For particles larger than

400 nm, spherical particles showed a little higher penetration than others due to higher effect

of impaction, however, difference was negligible. The deposition of particles slightly

105

decreased with increasing of the air flow rate, because increasing of velocity led decreasing of

deposition due to the diffusion effect.

Figure 45. Deposition of particles with different mobility sizes in the lung as a function of

airway generation. The air flow rate was 646 ml/s for the sitting breathing condition.

106

Figure 46. Deposition of particles with different mobility sizes in the lung as a function of

airway generation. The air flow rate was 1291 ml/s for the light working breathing condition.

107

Figure 47 Deposition of particles with different mobility sizes in the lung as a function of

airway generation. The air flow rate was 1937 ml/s for the heavy working breathing condition.

Airborne CNTs were collected and analyzed by SEM in order to obtain the fraction of single

standing CNTs in DMA classified samples. A number of images have been taken by SEM and

agglomerated and single standing CNTs counted by the image analyzing software (ImageJ).

108

Obtained data are presented in Table 10 and the data was fitted to the exponential line as

shown in Figure 48 to define relationship between the mobility size and the fraction of

agglomerated and single standing CNTs. As presented in Chapter 2, agglomerate’s fraction

was increased with increasing of mobility size of airborne CNTs. The obtained relation was

applied to the deposition model and equation is expressed as:

𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝐹𝑆𝑆𝐹𝐹𝐹𝐹𝐹𝐹 = 2.00exp (−0.01𝑑𝑑𝑖𝑖), (6-22)

where dm is in µm.

Table 10. Fraction of single standing CNT in classified airborne CNT samples as a function of

mobility diameter.

Mobility diameter (nm) 75 90 110 150 200 300 400 500

Single standing CNT fraction 0.97 0.86 0.77 0.58 0.20 0.08 0.04 0.02

109

Figure 48. Fraction of single standing CNTs as a function of classified mobility diameters.

The fitting shows the relation between single standing CNT fractions and mobility sizes.

Regional particle deposition in the airways and lung was calculated by different particle

models including different deposition mechanisms. The entire region was mainly divided to

three parts such as a head (nasal), tracheobronchial and alveolar regions. Addition to the three

particles, which are spherical particles with unit density, agglomerated and single standing

CNTs, the agglomerate fraction applied model was employed to estimate particle deposition

precisely. The calculated results are shown in Figures 49, 50 and 51, and each figure presents

results for sitting breathing, light working breathing and heavy working breathing conditions,

respectively. The nasal deposition increases with increasing of flow rate of breathing for all

110

particle models, since depositions by diffusion for small particles and impaction for large

particles were increased with higher air flow rate into the head airways. On the other hand,

particle depositions in tracheobronchial and alveolar regions slightly decreased due to

increasing of air flow velocity, which led decreasing of particle collections by diffusion in the

regions. When the fraction of agglomerates increases, agglomerate fraction applied model

smoothly moves close to the deposition line of agglomerated CNTs. More fractions of CNTs

can be captured deep inside the lung with increasing of agglomerated CNTs fraction.

111

Figure 49. Regional particle deposition models for different particles such as spherical

particles with unit density, CNT agglomerates, Single standing CNTs and mixed particle

model of agglomerated and single standing CNTs, under the sitting breathing condition.


particles with unit density, CNT agglomerates, Single standing CNTs and mixed model of

agglomerated and single standing CNTs, under the light working breathing condition.

112


particles with unit density, CNT agglomerates, Single standing CNTs and mixed model of

agglomerated and single standing CNTs, under the heavy working breathing condition.

Total regional deposition of particle was obtained by model calculations and results were

shown in Figures 52, 53 and 54 presenting results for different breathing scenarios. Total

deposition of each region was calculated with consideration of air flow direction from the

nasal to alveolar and deposited amount of particles in previous region. The efficiencies for the

tracheobronchial and alveolar region can be expressed as:

𝐸𝐸𝑘𝑘𝑇𝑇 = η𝑘𝑘𝑇𝑇(1 − 𝐸𝐸𝑖𝑖𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛) 𝐹𝐹𝐹𝐹𝑑𝑑 𝐸𝐸𝑛𝑛𝑛𝑛𝑣𝑣𝑓𝑓𝑡𝑡𝑛𝑛𝑛𝑛𝑓𝑓 = η𝑛𝑛𝑛𝑛𝑣𝑣𝑓𝑓𝑡𝑡𝑛𝑛𝑛𝑛𝑓𝑓(1 − 𝐸𝐸𝑖𝑖𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 − 𝐸𝐸𝑘𝑘𝑇𝑇). (6-23)

113

In the sitting breathing scenario, CNT agglomerates showed higher fraction of alveolar

deposition especially for smaller sizes than 100 nm mobility size. Furthermore, small single

standing CNTs, which possess structural similarity with asbestos fiber, showed high

deposition fraction in the alveolar region. This is due to the low flow velocity (4.3-0.1 m/s for

17th to 23th generations, respectively, under the sitting breathing scenario) in Alveolar region

and the low velocity promotes increasing of diffusion deposition of particles.

Tracheobronchial depositions for all particle types showed comparably lower fraction than

other regions under this low flow rate condition, because a high air flow velocity in big

airway helps particle penetration throughout airways in the tracheobronchial region. For the

increasing of air flow rate, for example, light working breathing scenario showed that

deposition fraction in the head region for all particles increased and lung depositions

decreased for both agglomerated and single standing CNTs. In heavy working breathing

condition, nasal deposition increased even more and depositions in the lung including

tracheobronchial and alveolar regions decreased dramatically in the entire particle size range.

114

Figure 52. Total efficiency of regional deposition model for different particles such as unit

density spherical particles, CNT agglomerates, Single standing CNTs and mixed model of

agglomerated and single standing CNTs, under the sitting breathing condition.

115

Figure 53. Calculation results of regional deposition models for different particles such as unit


agglomerated and single standing CNTs, under the light working breathing condition.

116

Figure 54. Calculation results of regional deposition models for different particles such as unit


agglomerated and single standing CNTs, under the heavy working breathing condition.

117

7. Conclusions

In the present study, carbon nanotubes (CNTs), which possess an extremely high

aspect ratio and potential toxicity concerns, have been characterized, since effect of their

physical properties such as a shape, size, length, chemistry, functionalization, etc. on their

toxicity are not fully understood. Thus the efficacy of current characterization methods for

CNTs has been investigated. For example, the geometrical length and effective density of

MWCNTs have been obtained by particle measurement systems, such as the Scanning

Mobility Particle Sizer (SMPS), Centrifugal Particle Mass Analyzer (CPMA) and Scanning

Electron Microscopy (SEM). Additionally the developed filtration method to obtain the

geometrical length of CNTs in real-time showed the possibility of fast measurement for the

geometrical length of elongated particles. The DMA-CPMA tandem measurement method

was used to obtain effective densities for CNTs with different mobility diameters, and the

fractal model to establish the relationship between the MWCNT properties including tube

diameters, length, intrinsic density and the agglomerate properties such as the mass, porosity,

effective density, surface area and characteristic diameters was developed. The model

developed to compute the geometrical outer diameters and fractal dimensions of the

agglomerates agreed well with the results obtained by the SEM image analysis and literature

data. The model can be used for monitoring CNT emissions and their fate and transport at

workplaces and can also provide useful data for toxicity studies of airborne CNTs.

118

To characterize CNTs, a good understanding of CNT dispersion is critical, because properties

of CNTs depend on whether they agglomerate or stand individually. Therefore, the

relationship between the concentration and agglomerate status of CNT suspensions was

investigated and a hydrogen bond effect on the enhancement of CNT dispersion in liquid

suspensions was discovered. The concentration of airborne CNTs was reduced when the

concentration decreased in the suspension and smaller mode sizes were obtained from the

suspensions with lower concentrations. Well dispersed MWCNTs or agglomerated MWCNTs

with a certain degree of control were successfully generated. Common electrolytes, such as

sodium chloride (NaCl), potassium chloride (KCl), etc., promote agglomeration of

nanoparticles in aqueous solutions. On the contrary, the acetic electrolyte effect on

enhancement of the dispersion of multi-walled carbon nanotubes (MWCNTs) with a carboxyl

functional group through the strong hydrogen bond was discovered. The results were

confirmed by the UV-Vis spectrometry and dispersion observations. The aerosolization-

quantification method using the relationship between the mobility through the model

membrane filter and dispersion stability of nanoparticles in suspensions was also employed.

When concentrations of acetate electrolytes such as ammonium acetate (CH3CO2NH4) and

sodium acetate (CH3CO2Na) were lower than a certain threshold, MWCNT suspensions

showed a better dispersion and had a higher mobility in porous media. The discovered effects

by the acetic environment are also applicable to other nanoparticles with the carboxyl

119

functional group, which was demonstrated with polystyrene latex (PSL) particles as an

example.

Filtration is an effective method to control nanoparticle emissions and more efficient for

elongated particles, because the deposition due to interception is more effective with

cylindrical particles which possess long effective lengths. Different filters such as metal

screens, nanofibers and Nuclepore filters were tested against either airborne or liquid borne

CNTs. When mobility diameters of CNTs were small, which means CNTs were short and

well-dispersed, filtration efficiencies for CNTs were higher than those for sphere-like

particles. On the other hand, the CNTs with big mobility diameters, which means CNTs were

either aligned to the flow or formed agglomerates, showed the lower filtration efficiencies

than those for sphere-like electrolytes.

Obtained physical characteristics were applied to the lung deposition models including

different deposition mechanisms and different particle models. The models for spherical

particles and fiber were revised and showed reasonable results with given parameters. The

model calculations showed that the CNT agglomerate status plays an important role on CNT

deposition in airways into the lung. For example, under quiet breathing condition,

agglomerate CNTs in the entire size range and small single standing CNTs can be deposited

the deep inside the lung and it may cause the toxic effect on the lung as shown in previous

120

literatures. The modified model can be applied to assess exposure of CNTs or elongated

particles at workplaces.

121

8. Outlook

There are remained extendable research branches regarding the present study.

Filtration studies using different types of filters have been performed and experimental results

still showed discrepancies with the model calculation for airborne CNT filtration, because the

existing model considered CNTs as cylindrical or ellipsoid particles. However, as shown in

the present study, airborne CNTs contains a fraction of agglomerates and it must be

considered in the model. The fractal model, which developed in the present study, can be a

key to improve the model prediction.

Detection of airborne CNTs is important in occupation health to prevent inhalation of released

airborne CNTs at workplaces. However, there is no effective method to detect CNTs in air so

far, because CNTs mixed in air with other nanoparticles, which is a common situation in the

nature, are hardly distinguishable from the other particles. In the long term, development of

detection method for released airborne CNTs is interesting and valuable research. Tandem

measurement of airborne measuring instruments can be one possibility and optical or

electrical sensing methods can be the other. In addition, the research can be expanded to a

nanoparticle sensor development.

Regarding the lung deposition study, the present study showed that agglomerate fraction

applied model to assess CNT exposure at workplaces. However, none of existing study

performed experiment with this concern, thus it is difficult to verify the developed model.

122

Therefore, animal tests including the degree of control on CNT agglomeration will be useful

to improve the model and can be applied for an assessment of CNT exposure.

123

Bibliography

1. Mossman BT, Churg A (1998) Mechanism in the pathogenesis of asbesotisi and silicosis.

Am. J. Respir. Crit. Care Med. 157:1666-1680.

2. Davis JMG, Addison J Bolton RE, Donaldson K, Jones AD, Smith T (1986) The

pathogenicity of long versus short fibre samples of amosite asbestos administered to rats by

inhalation and intraperitoneal injection. Br. J. exp. Path. 67:415-430.

3. Donaldson K, Tran CL (2004) An introduction to the short-term toxicology of respirable

industrial fibres. Mutat. Res. 553:5-9.

4. Muller J, Delos M, Panin N, Rabolli V, Huaux F, Lison D (2009) Absence of carcinogenic

response to multiwall carbon nanotubes in 2-year bioassay in the peritoneal cavity of the rat.

Toxicol. Sci. 110:442–448.

5. Poland CA, Duffin R, Kinloch I, Maynard A, Wallace WAH, Seaton A, Stone V, Brown S,

Macnee W, Donaldson K (2008) Carbon nanotubes introduced into the abdominal cavity of

mice show asbestos-like pathogenicity in a pilot study. Nat. Nanotechnol. 3:423–428.

6. Takagi A, Hirose A, Nishimura T, Fukumori N, Ogata A, Ohashi N (2008) Induction of

mesothelioma in p53+/- mouse by intraperitoneal application of multi-wall carbon nanotube.

J. Toxicol. Sci. 33:105–116.

124

7. Pauluhn J (2010) Subchronic 13-week inhalation exposure of rats to multiwalled carbon

nanotubes: toxic effects are determined by density of agglomerate structures, not fibrillar

structures. Toxicol. Sci. 113:226–42.

8. Ma-Hock L, Strauss V, Treumann S, Küttler K, Wohlleben W, Hofmann T, Groters S,

Wiench K, van Ravenzwaay B, Landsiedel R (2013) Comparative inhalation toxicity of

multi-wall carbon nanotubes, graphene, graphite nanoplatelets and low surface carbon black.

Part. Fibre. Toxicol. 10(1):23.

9. Shvedova AA, Kisin E, Murray AR, Johnson VJ, Gorelik O, Arepalli S, Hubbs AF, Mercer

RR, Keohavong P, Sussman N, Jin J, Stone S, Chen B, Deye G, Maynard A, Castranova V,

Baron PA, Kagan V (2008) Inhalation versus aspiration of single walled carbon nanotubes

in C57BL/6 mice: Inflammation, fibrosis, oxidative stress and mutagenesis. Am. J. Physiol,

Lung Cell Mol. Physiol. 295:L552–565.

10. Li JG, Li WX, Xu JY, Cai XQ, Liu RL et al. (2007) Comparative study of pathological

lesions induced by multiwalled carbon nanotubes in lungs of mice by intratracheal

instillation and inhalation. Environ. Toxicol. 22:415-421.

11. Ma-Hock L, Treumann S, Strauss V, Brill S, Luizi F et al. (2009) Inhalation toxicity of

multiwall carbon nanotubes in rats exposed for 3 months. Toxicol. Sci. 112:468-481.

125

12. Porter DW, Hubbs AF, Chen BT, Mckinney W, Mercer RR, Wolfarth MG et al. (2012)

Nanotoxicol. 7:1179-1194.

13. Sargen L, Porter DW, Lowry DT, Battelli LA, Siegrist K et al. (2013) Multiwalled carbon

nanotube-induced lung tumors. Toxicologist 132:98.

14. Erdely A, Hulderman T, Salmen R, Liston A, Zeidler-Erdely PC et al. (2009) Cross-talk

between lung and systemic circulation during carbon nanotube respiratory exposure.

Potential biomarkers. Nano Lett. 9:36-43.

15. Wick P, Manser P, Limbach LK, Dettlaff-Weglikowska U, Krumeich F, Roth S, Stark WJ,

Bruinink A (2007) The degree and kind of agglomeration affect carbon nanotube

cytotoxicity. Toxicol. Lett. 168:121-131.

16. Murphy FA, Poland CA, Duffin R, Donaldson K (2013) Length-dependent pleural

inflammation and parietal pleural responses after deposition of carbon nanotubes in the

pulmonary airspaces of mice. Nanotoxicology 7:1157-1167.

17. Erdely A, Dahm M, Chen BT, Zeidler-Erdely PC, Fernback JE, Birch ME et al. (2013)

Carbon nanotube dosimetry: from workplace exposure assessment to inhalation toxicology.

Part. Fibre. Toxicol. 10:53.

126

18. Wang X, Li Q, Xie J, Jin Z, Wang J, Li Y, et al. (2009) Fabrication of ultralong and

electrically uniform single-walled carbon nanotubes on clean substrates. Nano. Lett.

9(9):3137–3141.

19. Staszczuk P, Matyjewicz M, Kowalska E, Radomska J, Byszewski P, Kozlowski M (2003)

Studies of surface properties and fractal dimensions of carbon nanotubes using complex

methods. Rev. adv. Mater. Sci. 5:471-476.

20. Kim SH, Zachariah MR (2006) In-flight kinetic measurements of the aerosol growth of

carbon nanotubes by electrical mobility classification. J. Phys. Chem. 110:4555–4562.

21. Hinds WC (2012) Aerosol technology: properties, behavior, and measurement of airborne

particles (Second edition). John Wiley & Sons, New York, Chapter 9.

22. Kim SH, Mulholland GW, Zachariah MR (2009) Density measurement of size selected

multi-walled carbon nanotubes by mobility-mass characterization. Carbon 47(5):1297–1302.

23. Chen BT, Schwegler-Berry D, McKinney W, Stone S, Cumpston JL, Friend S, Porter DW,

Castranova V, Frazer DG (2012) Multi-walled carbon nanotubes: Sampling criteria and

aerosol characterization. Inhal. Toxicol. 24:798-820.

24. Bayer Material Science (2012) Product Information Baytubes C 150 P. Available at

http://www.baytubes.com/product_production/baytubes_data.html.

127

25. Oberdörster G (2012) Nanotoxicology: in vitro–in vivo dosimetry. Environ. Health Persp.

120(1): A13.

26. Premkumar T, Mezzenga R, Geckeler KE (2012) Carbon nanotubes in the liquid phase:

addressing the issue of dispersion. Small 8(9):1299-1313.

27. Wu L, Gao B, Tian Y, Muñoz-Carpena R, Zigler KJ (2013) DLVO interactions of carbon

nanotubes with isotropic planar surfaces. Lagnmuir 29:3976-3988.

28. Balasubramanian K, Burghar, M (2005) Chemically functionalized carbon nanotubes.

Small 1(2):180-192.

29. Smith B, Wepasnich K, Schrote KE, Bertele AR, Ball WP, O’Melia C, Fairbrother DH

(2009) Colloidal properties of aqueous suspensions of acid-treated, multi-walled carbon

nanotubes. Environ. Sci. Technol. 43:819-825.

30. Zhang J, Zou H, Qing Q, Yang Y, Li Y, Liu Z, Guo X, Du Z (2003) Effect of chemical

oxidation on the structure of single-walled carbon nanotubes. J. Phys. Chem. B 107, 3712-

3718.

31. Kennedy AJ, Hull MS, Steevens JA, Dontsova KM, Chappell MA, Gunter JC, Weiss Jr

CA (2008) Factors influencing the partitioning and toxicity of nanotubes in the aquatic

environment. Environ. Toxicol. Chem. 27(9):1932-1941.

128

32. Hyung H, Fortner JD, Hughes JB, Kim JH (2007) Natural organic matter stabilizes carbon

nanotubes in the aqueous phase. Environ. Sci. Technol. 41:179-184.

33. Parks GA (1990) Surface energy and adsorption at mineral-water interfaces: An

introduction. Rev. Mineral Geochem. 23:133-175.

34. Schwarzenbach RP, Gschwend PM, Imboden DM (1993) Environmental organic

chemisty. John Wiley & Sons, New York.

35. Askari M, Abouie J (2011) A closed form for the electrostatic interaction between two

rod-like charged objects. J. Phys. Condens. Matter 23(15):155103-155108.

36. Methner M, Hodson L, Dames A, Geraci C (2010) Nanoparticle emission assessment

technique (NEAT) for the identification and measurement of potential inhalation exposure to

engineered nanomaterials part B: Results from 12 field studies. J. Occup. Environ. Hyg.

7(3):163–176.

37. Dahm MM, Evans DE, Schubauer-Berigan MK, Birch ME, Fernback JE (2012)

Occupational exposure assessment in carbon nanotube and nanofiber primary and secondary

manufacturers. Ann. Occup. Hyg. 56:542–556.

38. Hedmer M, Isaxon C, Nilsson PT, Ludvigsson L, Messing ME, Genberg J, Skaug V,

Bohgard M, Tinnerberg H, Pagels JH (2014) Exposure and emission measurements during

129

production, purification, and functionalization of arc-discharge-produced multi-walled

carbon nanotubes. Ann. Occup. Hyg. 58:355-79.

39. Maynard AD, Ku BK, Emery M, Stolzenburg M, McMurry PH (2007) Measuring particle

size-dependent physicochemical structure in airborne single walled carbon nanotube

agglomerates. J. Nanopart. Res. 9(1):85-92.

40. Smajda R, Kukovecz Á, Kónya Z, Kiricsi I (2007) Structure and gas permeability of

multi-wall carbon nanotube buckypapers. Carbon 45(6):1176-1184.

41. Tripol’skii AI, Serebrii TG, Lemesh NV, Ivashchenko TS, Strizhak PE (2009) Fractal

analysis of carbon nanotube agglomerates obtained by chemical vapor decomposition of

ethylene over nickel nanoparticles. Theor. Exp. Chem. 45(2):103-107.

42. NIOSH (2013). NIOSH Current Intelligence Bulletin 65: Occupational Exposure to

Carbon Nanotubes and Nanofibers. Cincinnati, OH: U.S. Department of Health and Human

Services, Centers for Disease Control, National Institute for Occupational Safety and Health,

DHHS (NIOSH) Publication No. 2013–145.

43. Kagan VE, Konduru NV, Feng W, Allen BL, Conroy J, Volkov Y, Vlasova II, Belikova

NA, Yanamala N, Kapralov A, Tyurina YY, Shi J, Kisin ER, Murray AR, Franks J, Stoiz D,

Gou P, Klein-Seetharaman J, Fadeel B, Star A, Shvedova AA (2010) Carbon nanotubes

130

degraded by neutrophil myeloperoxidase induce less pulmonary inflammation. Nat.

Nanotechnol. 5:354–358.

44. Kanaoka C, Emi H, Otani Y, Iiyama T (1987) Effect of charging state of particles on

electret filtration, Aerosol Sci. Technol. 7:1–13.

45. Lam CW, James JT, McCluskey R, Hunter RL (2004) Pulmonary toxicity of single-wall

carbon nanotubes in mice 7 and 90 days after intratracheal instillation. J. Toxicol. Sci.

77:126–134.

46. Liu X, Guo L, Morris D, Kane AB, Hurt RH (2008) Targeted removal of bioavailable

metal as detoxification strategy for carbon nanotubes. Carbon 46:489–500.

47. Osmond-McLeod MJ, Poland CA, Murphy F, Waddington L, Morris H, Hawkins SC,

Clark S, Aitken R, McCall MJ, Donaldson K (2011) Durability and iinflammogenic impact

of carbon nanotubes compared with asbestos fibres. Part. Fibre Toxicol. 8:15–32.

48. Shvedova AA, Kisin ER, Mercer R, Murray AR, Johnson VJ, Potapovich AI, Tyurina YY,

Gorelik O, Arepalli S, Schwegler-Berry D (2005) Unusual inflammatory and fibrogenic

pulmonary responses to single walled carbon nanotubes in mice. Am. J. Physiol. Lung Cell

Mol. Physiol. 289:L698–708.

49. Warheit DB (2006) What is currently known about the health risks related to carbon

nanotube exposures? Carbon 44:1064–1069.

131

50. Federal Register (2010) Multi-walled carbon nanotubes and single-walled carbon

nanotubes significant new use rules. Federal Register, 75 (180), September 17, 2010/ Rules

and Regulations.

51. Federal Register (2011) Multi-walled carbon nanotubes significant new use rule. Federal

Register, 76 (88), May 6, 2011/ Rules and Regulations.

52. Federal Register (2013) Functionalized multi-walled carbon nanotubes significant new

use rule. Federal Register, 78 (90), May 9, 2013/ Rules and Regulations.

53. Chen J, Hamon MA, Hu H, Chen Y, Rao AM, Eklund PC, Haddon RC (1998) Solution

properties of single-walled carbon nanotubes. Science 282:95–98.

54. Sun YP, Fu K, Lin Y, Huang W (2002) Functionalized carbon nanotubes: Properties and

applications. Acc. Chem. Res. 35:1096–1104.

55. Zhang Y, Bai Y, Yan B (2010) Functionalized carbon nanotubes for potential medicinal

applications. Drug Discov. Today 15:428–435.

56. Chen RJ, Bangsaruntip S, Drouvalakis KA, Kam NWS, Shim M, Li Y, Kim W, Utz PJ,

Dai H (2003) Noncovalent functionalization of carbon nanotubes for highly specific

electronic biosensors. Proc. Natl. Acad. Sci. USA 100:4984–4989.

57. Hu H, Zhao B, Itkis ME, Haddon RC (2003) Nitric acid purification of single-walled

carbon nanotubes. J. Phys. Chem. B 107 :13838–13842.

132

58. Bottini M, Bruckner S, Nika K, Bottini N, Bellucci S, Magrini A, Bergamaschi A,

Mustelin T (2006) Multi-walled carbon nanotubes induce T lymphocyte apoptosis. Toxicol.

Lett. 160:121–126.

59. Zhao X, Liu R (2012) Recent progress and perspectives on the toxicity of carbon

nanotubes at organism, organ, cell, and biomacromolecule levels. Environ. Int. 40:244–256.

60. Seto T, Furukawa T, Otani Y, Uchida K, Endo S (2010) Filtration of multi-walled carbon

nanotube aerosol by fibrous filters. Aerosol Sci. Technol. 44:734–740.

61. Wang J, Kim SC, Pui DYH (2011) Measurement of multi-wall carbon nanotubes

penetration through a screen filter and single-fiber analysis. J. Nanopart. Res. 13:4565–4573.

62. Wang J, Kim SC, Pui DYH (2011) Carbon nanotube penetration through a screen filter:

Numerical modeling and comparison with experiments. Aerosol Sci. Technol. 45:443–452.

63. Vo E, Zhuang Z (2013) Development of a new test system to determine penetration of

multi-walled carbon nanotubes through filtering facepiece respirators. J. Aerosol Sci. 61:50–

59.

64. Cheng YS, Holmes TD, Fan B (2006) Evaluation of respirator filters for asbestos fibers. J.

Occup. Environ. Hyg. 3:26–35.

65. Sturm R (2014) Deposition of carbon nanotubes in human alveoli – A theoretical

approach for risk assessment. SOP Transactions of Nano-technology 1:21-31.

133

66. Su WC, Cheng YS (2015) Estimation of carbon nanotubes deposition in a human

respiratory tract replica. J. Aerosol Sci. 79:72-85.

67. Yeh HC, Schum GM (1980) Models of human lung airways and their application to

inhaled particle deposition. Bull. Math. Biol. 42:461-480.

68. Anjilvel S, Ascharian B (1995) A multiple-path model of particle deposition in the rat

lung. Fundam. Appl. Toxicol. 28(1):41-50.

69. Ding JY, Yu CP, Zhang L, Chen YK (1997) Deposition modeling of fibrous particles in

rats: Comparisons with available experimental data. Aerosol Sci. Tech. 26(5):403-414.

70. Lee S, Lee JH, Bae GN (2009) Size response of an SMPS-APS system to commercial

multi-walled carbon nanotubes. J. Nanopart. Res. 12:501-512.

71. Ku B, Kulkarni, P (2009) Morphology of single-wall carbon nanotube aggregates

generated by electrospray of aqueous suspensions, J. Nanopart. Res. 11:1393-1403.

72. Jennerjohn N, Eiguren-Fernandez A, Priknodko S, Fung DC, Hirakawa KS, Zavala-

Mendez JD, Hinds W, Kennedy NJ (2010) Design, demonstration and performance of a

versatile electrospray aerosol generator for nanomaterial research and applications.

Nanotechnol. 21(25):255603.

134

73. Bahk YK, Buha J, Wang J (2013) Determination of geometrical length of airborne carbon

nanotubes by electron microscopy, model, calculation, and filtration method. Aerosol Sci.

Technol. 47(7):776-784.

74. Bahk YK, He X, Kuo YY, Kim N, Gitsis E, Wang J (2015) Enhanced dispersion stability

and mobility in porous media of carboxyl functionalized carbon nanotubes in aqueous

solutions through strong hydrogen bonds. J. Nanopart. Res. 17:396.

75. Steiner T (2002) The hydrogen bond in the solid state. Angew. Chem. Int. Ed. Engl.

41(1):49-76.

76. Bhattacharjee S, Elimelech M, Borkovec M (1998) DLVO interaction between colloidal

particles: Beyond Derjaguin’s approximation. Croat. Chem. Acta. 71:883-903.

77. Masliyah JH, Bhattacharjee S (2006) Electrokinetic and colloid transport phenomena.

John Wiley & Sons, Hoboken, NJ.

78. Bunkin NF, Kiseleva OA, Lobeyev AV, Movchan TG, Ninham BW, Vinogradova OI

(1997) Effect of salts and dissolved gas on optical cavitation near hydrophobic and

hydrophilic surfaces. Langmuir 13:3024-3028.

79. Grasso D, Subramaniam K, Butkus M. Strevett K, Bergendahl J (2002) A review of non-

DLVO interactions in environmental colloidal systems. Rev. Environ. Sci. Bio-Tech. 1:17-38.

135

80. Israelachvili JN, Wennerström H (1996) Role of hydration and water structure in

biological and colloidal interactions. Nature 379:219-224.

81. Jeffrey GA (1997) An introduction to hydrogen bonding. Oxford University Press, New

York.

82. Ninham BW, Kurihara K, Vinogradova OI (1997) Hydrophobicity, specific ion adsorption

and reactivity. Colloids. Surf. A: Physicochem. Eng. Aspects 123-124: 7-12.

83. Pauling L (1960) The nature of the chemical bond. Cornell University Press, Ithaca, NY.

84. Bentz J, Nir S (1981) Aggregation of colloidal particles modeled as a dynamical process.

Proc. Natl. Acad. Sci. USA 78(3):1634-1647.

85. Bohidar HB, Bhakat P, Sharma J, Saxena A (2000) Anomalous colloidal stability of

protein coated polystyrene latex beads studies by small angle light scattering. Int. J. Biol.

Macromol. 27:111-116.

86. Fukasawa T, Adachi Y (2010) Direct observation on the Brownian coagulation of PSL

particles through optical microscope in the regime near critical coagulation concentration

(CCC). J. Colloid. Interface Sci. 344(2):343-347.

87. Takayasu MM, Galembeck F (1993) Polystyrene latex coagulatio: A study using

centrifugation in density gradients. J. Colloid. Interface Sci. 155:16-22.

136

88. Peng X, Jia J, Gong X, Luan Z, Fan B (2009) Aqueous stability of oxidized carbon

nanotubes and the precipitation by salts. J. Hazard Mater. 165:1239-1242.

89. Ling TY, Wang J, Pui DYH (2010) Measurment of retention efficiency of filters against

nanoparticles in liquids using an aerosolization technique. Environ. Sci. Technol. 44:774-

779.

90. Shieh Y, Liu G, Wu H, Lee C (2007) Effect of polarity and pH on the solubility of acid-

treated carbon nanotubes in different media. Carbon 45:1880-1890.

91. Marcus Y (1991) Thermodynamics of solvation of ions: Part 5. Gibbs free energy of

hydration at 298.15K. J. Chem. Soc. Faraday Trans. 87(18):2995-2999.

92. Tansel B, Sager J, Rector T, Garland J, Strayer R, Lanfang L, Roberts M, Hummerick M,

Bauer J (2006) Significance of hydrated radius and hydration shells on ionic permeability

during nanofiltration in dead end and cross flow modes. Sep. Purif. Technol. 51:40-47.

93. Israelachvili JN, Pashley RM (1983) Molecular layering of water at surfaces and origin of

repulsive hydration forces. Nature 306:249-250.

94. Goel A (2006) Colloidal chemistry. Discovery publishing house, New Delhi.

95. Jorgensen TC, Weatherley LR (2003) Ammonia removal from wastewater by ion

exchange in the presence of organic contaminants. Water Res. 37:1723-1728.

137

96. Jaisi DP, Elimelech M (2009) Single-walled carbon nanotubes exhibit limited transport in

soil columns. Environ. Sci. Technol. 43:9161-9166.

97. Shieh YT, Chen JY, Twu YK, Chen WJ (2012) The effect of pH and ionic strength on the

dispersion of carbon nanotubes in poly (acrylic acid) solutions. Polym. Int. 61:554–559.

98. Heister E, Lamprecht C, Neves V, Tilmaciu C, Datas L, Flahaut E, Soula B, Hinterdorfer

P, Coley HM, Silva SRP, McFadden J (2010) Higher dispersion efficacy of functionalized

carbon nanotubes in chemical and biological environments. ACS Nano 4:2615-2626.

99. Wang J, Pui DYH (2013) Dispersion and filtration of carbon nanotubes (CNTs) and

measurement of nanoparticle agglomerates in diesel exhaust. Chem. Eng. Sci. 85:69-76.

100. Kim SH, Mulholland GW, Zachariah MR (2007) Understanding ion-mobility and

transport properties of aerosol nanowires. J. Aerosol Sci. 38:823-842.

101. Lall AA, Friedlander SK (2006) On-line measurement of ultrafine aggregate surface area

and volume distributions by electrical mobility analysis: I. Theoretical analysis. J. Aerosol

Sci. 37(3), 260-271.

102. Lange R, Fissan H, Schmidt-Ott A (1999) Predicting the collection efficiency of

agglomerates in fibrous filter: Part. Part. Syst. Charact. 16:60-65.

103. Wang J (2012) Effects of particle size and morphology on filtration of airborne

nanoparticles. KONA Powder Part. J. 30:256-266.

138

104. Myojo, T. (1999) A simple method to determine the length distribution of fibrous

aerosols. Aerosol Sci. Technol. 30:30-39.

105. Bahk YK, Wang J (2014) Filtration and length determination of airborne carbon

nanotubes in the submicrometer range using nanofiber filters. Aerosol Air Qual. Res. 14

(5):1352-1359.

106. Chen SC, Wang J, Bahk YK, Fissan H, Pui DYH (2014) Carbon nanotube penetration

through fiberglass and electret respirator filter and Nuclepore filter media: experiments and

models. Aerosol Sci. Tech. 48(10):997-1008.

107. Kim SH, Zachariah MR (2005) In-flight size classification of carbon nanotubes by gas

phase electrophoresis. Nanotechnol. 16:2149-2152.

108. Wang J, Pui DYH (2009) Filtration of aerosol particles by elliptical fibers: a numerical

study. J. Nanopart Res 11:185-196.

109. Pich J (1965) The filtration theory of highly dispersed aerosols. Staub Reinhalt. Luft

5:16-23.

110. Pich J (1966) The effectiveness of the barrier effect in fiber filters at small Knudsen

numbers. Staub Reinhalt. Luft 26:1-4.

111. Landahl HD, Herrmann RG (1949) Sampling of liquid aerosols by wires, cylinders, and

slides, and the efficiency of impaction of the droplets. J. Colloid Sci. 4:103-136.

139

112. Spurny KR, Lodge JP, Frank ER, Sheesley DC (1969) Aerosol filtration by means of

Nuclepore filters: Structural and filtration properties. Environ. Sci. Technol. 3:453–464.

113. Manton MJ (1978) The impaction of aerosols on a Nuclepore filter. Atmo. Environ.

12:1669–1675.

114. Manton MJ (1979) Brownian diffusion of aerosols to the face of a Nuclepore filter. Atmo.

Environ. 13:525–531.

115. Chen SC, Wang J, Fissan H, Pui YHD (2013) The use of Nuclepore filters for ambient

and workplace nanoparticle exposure assessment spherical particles. Atmos. Environ.

77:385–393.

116. Chen SC, Wang J, Fissan H, Pui YHD (2013) Exposure assessment of nanosized

engineered agglomerates and aggregates using Nuclepore filter. J. Nanoparticle Res.

15:1955–1969.

117. Gormley PG, Kennedy M (1949) Diffusion from a stream flowing through a cylindrical

tube. Proc. Roy. Irish Acad. Sect. A 52:163-169.

118. Wang J, Kim SC, Pui DYH (2008) Investigation of the figure of merit for filters with a

single nanofiber layer on a substrate. J. Aerosol. Sci. 39:323-334.

140

119. Gallily I, Schiby D, Cohen AH, Hollander W, Schless D (1986) On the inertial

separation of nonspherical aerosol particles from laminar flows. I. The cylindrical case.

Aerosol Sci. Tech. 5:267-286.

120. Gradoń L, Grzybowski P, Piłaciński W (1988) Analysis of motion and deposition of

fibrous particles on a single filter element. Chem. Eng. Sci. 43:1253-1259.

121. Davoren M, Herzog E, Casey A, Cottineau B, Chambers G, Byrne HJ, Lyng FM (2007)

In vitro toxicity evaluation of single walled carbon nanotubes on human A549 lung cells.

Toxicol. In Vitro 21:438–448.

122. Fraczek A, Menaszek E, Paluszkiewicz C, Blazewicz M (2008) Comparative in vivo

biocompatibility study of singleand multi-wall carbon nanotubes. Acta Biomat. 4:1593-1602.

123. Wang J, Asbach C, Fissan H, Hülser T, Kuhlbusch TAJ, Thompson D, Pui DYH (2011)

How can nanobiotechnology oversight advance science and industry: examples from

environmental, health and safety studies of nanoparticles (nano-EHS). J. Nanopart. Res.

13:1373 – 1387.

124. Wang J, Bahk YK, Chen SC, Pui DYH (2015) Characteristics of airborne fractal-like

agglomerates of carbon nanotubes. Carbon 93:441-450.

125. www.timesnano.com, Technical data for TNM3, 2015.

126. Delhaes P (2001) Graphite and precursors. CRC Press, Amsterdam.

141

127. Davies CN (1973) Air filtration. Academic Press, London.

128. Pich J (1971) Pressure characteristics of fibrous aerosol filters. J. Colloid. Interf. Sci.

37:912-917.

129. Mandelbrot B (1967) How long is the coast of Britain? Statistical self-similarity and

fractional dimension. Science 156(3775):636–638.

130. Kim SC, Wang J, Emery M, Shin WG, Mulholland G, Pui, DYH (2009) Structural

property effect of nanoparticle agglomerates on particle penetration through fibrous filter.

Aerosol Sci. Tech 43:344–355.

131. Vanni M (2000) Creeping flow over spherical permeable aggregates. Chem Eng Sci

55:685-698.

132. NIOSH (2009) Approaches to safe nanotechnology: Managing the health and safety

concerns with engineered nanomaterials. Cincinnati, OH: U.S. Department of Health and

Human Services, Centers for Disease Control, National Institute for Occupational Safety and

Health, DHHS (NIOSH) Publication No. 2009–125.

133. Dahm M, Yencken M, Schubauer-Berigan M (2011) Exposure Control Strategies in the

Carbonaceous Nanomaterial Industry. J. Occup. Environ. Med. 53:S68–S73.

134. NIOSH (1995) Approval of respiratory protective devices. Code of Federal Regulations

Title 42, Part 84.

142

135. Ozkan Y, Bradford PD (2013) Aligned carbon nanotube sheet high efficiency particulate

air filters. Carbon 64:295–304.

136. Bałazy A, Toivola M, Reponen T, Podgorski A, Zimmer A, Grinshpun SA (2006)

Manikin-based performance evaluation of N95 filtering facepiece respirators challenged

with nanoparticles. Ann. Occup. Hyg. 50:259–69.

137. Rengasamy S, Eimer BC, Shaffer R (2009) Comparison of nanoparticle filtration

performance of NIOSH-approved and CE-marked particulate filtering facepiece respirators.

Ann. Occup. Hyg. 53:117–128.

138. Brown RC (1993) Air filtration. Pergamon Press, London.

139. Stenhouse JIT, Gradon L, Marijnissen JCM (1999) Electret filters, production and

properties. Delft University Press, The Netherlands:119–128.

140. Wang J, Kim SC, Pui DYH (2008) Figure of merit of composite filters with micrometer

and nanometer fibers. Aerosol. Sci. Technol. 42: 722-728.

141. Gradoń L, Bałazy A, Podgórski A (2006) Nanofibrous media—promising tools for

filtration of nanosized aerosol particles. presented in The 7th international aerosol

conference; St. Paul, Minnesota.

142. Graham K, Ouyang M, Raether T, Grafe T, McDonald B, Knauf P (2002) Polymeric

nanofibers in air filtration applications. presented in Proceedings of the fifteen annual

143

technical conference and expo of the American filtration and separations society; Galveston,

Texas.

143. Podgórski A, Bałazy A, Gradoń L (2006) Application of nanofibers to improve the

filtration efficiency of the most penetrating aerosol particles in fibrous filters. Chem. Eng.

Sci. 61: 6804-6815.

144. Wang J, Chen DR, Pui DYH (2007) Modeling of filtration efficiency of nanoparticles in

standard filter media. J. Nanoparticle Res. 9:109–115.

145. Kim SC, Wang J, Emery M, Shin WG, Mulholland G, Pui DYH (2009) Structural

property effect of nanoparticle agglomerates on particle penetration through fibrous filter.

Aerosol Sci. Technol. 43:344 –355.

146. Lecoanet HF, Bottero JY, Wiesner MR (2004) Laboratory assessment of the mobility of

nanomaterials in porous media. Environ. Sci. Technol. 38(19):5164–5169.

147. Wang Y, Li Y, Fortner JD, Hughes JB, Abriola LM, Pennell KD (2008) Transport and

retention of nanoscale C60 aggregates in water-saturated porous media. Environ. Sci.

Technol. 42(10):3588–3594.

148. Kuhlbusch TA, Asbach C, Fissan H, Göhler D, Stintz M (2011) Nanoparticle exposure at

nanotechnology workplaces: A review. Part. Fibre Toxicol. 8(22):1–18.

144

149. Kang S. Mauter MS, Elimelech M (2009) Microbial cytotoxicity of carbon-based

nanomaterials: Implications for river water and wastewater effluent. Environ. Sci.

Technol.43:2648-2653.

150. Alpatova AL, Shan W, Babica P, Upham BL, Rogensues AR, Masten SJ, Drown E,

Mohanty AK, Alocija EC, Tarabara VV (2010) Single-walled carbon nanotubes dispersed in

aqueous media via non-covalent functionalization: Effect of dispersant on the stability,

cytotoxicity, and epigenetic toxicity of nanotube suspensions. Water Res. 44(2):505-520.

151. Madani SY, Mandel A, Seifalian AM (2013) A concise review of carbon nanotube's

toxicology. Nano Rev. 4:21521.

152. Lee JK, BYH Liu, Rubow KL (1993) Latex sphere retention by microporous membranes

in liquid filtration. J. IEST 36:26–36.

153. Bowen WR, Jenner F (1995) Theoretical descriptions of membrane filtration of colloids

and fine particles: An assessment and review. Adv. Colloid Interface Sci. 56: 141–200.

154. Madaeni SS and Fane AG (1996) Microfiltration of very dilute colloidal mixtures. J.

Membr. Sci. 113(2):301-312.

155. Grant DC, BYH Liu (2004) Sieving capture of liquidborne particles by microporous

membrane filtration media. Part. Part. Syst. Charact. 8(1-4):142–150.

145

156. Oganesyan VR, Orelovich OL, Yanina IV, Apel PY (2001) Study of retention ability of

nucleopore membranes. Colloid J. 63(6):755–761.

157. Savage N, Diallo MS (2005) Nanomaterials and water purification: Opportunities and

challenges. J. Nanopart. Res. 7(4-5):331–342.

158. Guo H, Wyart Y, Perot J, Nauleau F, Moulin P (2010) Low-pressure membrane integrity

tests for drinking water treatment: A review. Water Res. 44(1):41–57.

159. Ling TY, Wang J, Pui DYH (2011) Measurement of filtration efficiency of Nuclepore

filters challenged with polystyrene latex nanoparticles : experiments and modeling. J.

Nanopart. Res. 13:5415–5424.

160. Lee KW, Liu BYH (1980) On the minimum efficiency and the most penetrating particle

size for fibrous filters. J. Air Pollut. Control. Assoc. 30:377-381.

161. Asgharian B, Hofmann W, Bergmann (2010) Particle deposition in a multiple-path

model of the human lung. Aerosol Sci. Technol. 34(4):332-339.

162. Asgharian B, Yu CP (1988) Deposition of inhaled fibrous particles in the human lung. J.

Aerosol Medicine 1(1):37.

163. Asgharian B, Anjilvel S (1998) A multiple-path model of fiber deposition in the rat lung.

Toxicol. Sci. 44:80-86.

146

164. Raabe OG, Yeh HC, Schum GM, Phalen RF (1976) Tracheobronchial geometry: Human,

dog, rat, hamster-A compilation of selected data from the project respiratory tract deposition

models. Report LF-53. Lovelace Foundation, Albuquerque, NM.

165. Weibel EB (1966) Morphometry of the human lung. Biometrische Zeitschrift 8(1-2):

143-144.

166. Zhang L, Yu CP (1993) Empirical equations for nasal deposition of inhaled particles in

small laboratory animals and humans. Aerosol Sci. Technol. 19:51-56.

167. Raabe OG, Yeh HC, Newton GJ, Phalen RF, Velasquez DJ (1977) Deposition of inhaled

monodisperse aerosols in small rodents, in haled paricles IV: Part I. (Walton WH, Mcgovern

B, eds.), pp. 1-21, Pergamon Press, Oxford.

168. Raabe OG, Al-Bayati MA, Teague SV, Rasolt A (1988) Regional deposition of inhaled

monodisperse coarse and fine aerosol particle in small laboratory animals, in Inhaled

Particles IV (Dodgson J, McCallum RI, Bailey MR, Fisher DR, eds.) pp. 53-63, Pergamon

Press, Oxford.

169. Guyton AC (1947) Measurement of the respiratory volumes of laboratory animals. Am. J.

Physiol. 150:70–77.

170. Asgharian B, Yu CP (1989) Deposition of fibers in the rat lung. J. Aerosol. Sci. 20:355-

366.

147

171. Cheng YK, Yu CP (1993) Particle deposition from laminar flows by combined

mechanisms. Aerosol Sci. Technol. 19:389-395.

172. Gerde P, Cheng YS, Medinsky MA (1991) In vivo deposition of ultrafine aerosol in the

nasal airway of the rat. Fund. Appl. Toxicol. 16:330-336.

173. Ingham DB (1975) Diffusion of aerosols from a stream flowing through a cylindrical

tube. J. Aerosol Sci. 6:125-133.

174. Pich J (1972) Theory of gravitational deposition of particles from laminar flows in

channels. J. Aerosol Sci. 3:351-361.

175. Cai FS, Yu CP (1988) Inertial and interceptional deposition of spherical particles and

fibers in bifurcating airways. J. Aerosol. Sci. 19:679-688.

148

Appendix

A1. Comparison of generation methods for airborne MWCNTs

Airborne MWCNTs were generated by either an atomizer or electrospray. Size distributions

of generated aerosols by both methods were compared with each other. The electrosprayed

MWCNTs showed particle size distributions as in Figure A1-1. Distribution of MWCNTs

was certainly differentiated from the distribution of residual particles in a buffer solution.

However, due to the clogging problem in the capillary tube low concentration of MWCNT

suspension was required in the electrospray. Thus the particle concentration of electrosprayed

MWCNTs was not suitable for a study which needed high concentration of airborne

MWCNTs. For instance, it was difficult to obtain sufficient samples for SEM or Transmission

Electron Microscope (TEM) analysis with the electrosprayed CNTs, because of the low

concentration. On the other hand, the atomizer generated relatively high concentration

airborne MWCNTs from the liquid suspension. As shown in Figure A1-2, the size distribution

of airborne MWCNTs was differentiated from the residual particle distribution under

carefully controlled conditions. The concentration of airborne MWCNTs, which were

generated by an atomizer, was much higher than that of electrosprayed MWCNTs.

149

Figure A1-1. Mobility size distributions of aerosolized MWCNTs and residual particles,

which were generated by an electrospray.

Figure A1-2. Mobility size distributions of aerosolized MWCNTs and residual particles,

which were generated by an atomizer.

150

A2. DLVO model for C-MWCNTs and PSL particles

The DLVO model was employed to understand adhesion between particles and particles on

filter surface. The interaction between C-MWCNTs and the polycarbonate surface was

simulated based on the model developed by Wu et al. (27). They calculated DLVO interaction

between a surface modified MWCNT and isotropic planar surface. In this study, the model

included not only the filter front surface facing the liquid flow, but also the wall surface in the

cylindrical pore, since both surfaces provided deposit sites for C-MWCNTs and PSL.

Although the wall surface in the cylindrical pore was a curved surface, the simplified case of

planar surface was used in the calculation because of the distinctly smaller sizes of C-

MWCNTs in the range of 100 nm to 200 nm and the 100 nm PSL particles than the pore sizes

(3 µm and 1 µm, respectively). Parameters in the model include the length and number of

layers of C-MWCNTs, contacting angle between a C-MWCNT and the planar surface, etc.

Three different angles, 0°, 30° and 60° were applied to the model in order to investigate effect

of the angle (inset in Figure A2-1) and three ionic concentrations, 0.01, 0.02 and 0.03 M, were

used. The calculated results for C-MWCNTs were shown in Figure A2-1.

151

Figure A2-1. Results of the DLVO model for interaction between the C-MWCNT and

isotropic planar surface with different ionic strengths and orientation angles of the C-

MWCNT to the planar surface.

Figure A2-2 shows the calculated interaction energy between PSL particles and Nuclepore

filter under different ionic conditions from the DLVO theory. DLVO calculation revealed that

normal electrolytes reduced the energy barrier and for 0.03 M ionic strength, the secondary

minimum appeared around 18 nm distance and the net interaction energy was attractive which

152

might lead to the attachment of PSL on the filter surface, even though the attachment could be

reversed.

Figure A2-2. Results of DLVO model for interaction between a 100 nm PSL particle and

isotropic planar surface with different ionic strengths.

153

A3. Filtration of liquid-borne CNTs


The 47 mm Whatman track-etched-polycarbonate membrane filters (Nuclepore filter, GE

Healthcare, USA) with 1 μm and 3 μm pore diameters were used in this study for PSL

filtration and CNTs filtration experiments, respectively. Nuclepore (Capillary pore)

membrane filters are manufactured from polycarbonate film of high quality and consist of an

array of pores of uniform diameter which are perpendicular to the filter surface. The porosities

of the filters were measured at 0.16 (2×107 pores/cm3) for the 1 μm pore filter and 0.14 (2×106

pores/cm3) for the 3 μm filter. Their thickness was 11 μm and 9 μm for the 1 μm and 3 μm

filter, respectively (145).

Functionalized MWCNTs (Baytubes) with a carboxyl group and 15-20 nm diameters were

used for the filtration experiments. 50, 100, 200 and 500 nm size PSL particles were also

tested to demonstrate penetrations of spherical particles through the filters and the results

were compared with those of MWCNTs. PSL particles and MWCNTs were dispersed in DI

water. In addition, different electrolytes such as NaCl, KCl, CH3CO2NH4 and CH3CO2Na

were added in the suspensions to investigate their effect on the filtration efficiencies.

Figure A3-1 shows a schematic of the filtration system. The system consisting two different

flow delivery methods, such as using a filtering flask and peristaltic pump, was investigated.

Driven suspensions were passed through the Nuclepore filter and collected for further analysis.

154

The UV-Vis spectrophotometry and aerosolized method were employed to define penetrations

of PSL and CNT particles through the filter and the values were obtained with collected

samples at the upstream and downstream of the filter. With the aerosolized method, the

penetrations were calculated from obtained size distributions by the SMPS.

Figure A3-1. Experimental setup for filtration test and penetration determination system.

Filtration of liquid-borne MWCNTs using the Nuclepore filter

Figure A3-2 shows SEM images of deposited PSL particles on 1 µm pore Nuclepore filter

surface and depositions due to the interception and diffusion were observed. In particular, the

SEM images clearly show the deposited PSL particles by the surface diffusion, diffusion

towards the pore walls and interception. 500 nm size PSL particles were also used and most

of the particles collected by interception and sieving after the firstly deposited particles

155

clogged the pores. Thus the penetrations showed inconsistent results and were excluded in the

following discussions.

Figure A3-2. SEM images of different size of PSL particles challenging a 1 μm Nuclepore

filter.

The penetration experiments were performed using MWCNTs in DI water and KCl and

CH3CO2NH4 were added to the solution in order to investigate the effect of electrolyte to the

filtration efficiencies. The filtering flask was used as flow delivery system with a house

156

vacuum. The penetration of each sample was calculated with the upstream and downstream

concentrations of MWCNT suspensions determined by the UV-Vis spectrophotometry.

CH3CO2NH4 samples showed higher penetrations than deionized water and KCl samples and

that indicates MWCNTs were better dispersed than original DI water suspensions. This agrees

well with the discovery of the enhancement of CNT dispersion stability by acetic electrolytes

presented in Chapter 2. As we discussed in Chapter 2, single standing MWCNTs have higher

penetrations than agglomerates due to their higher mobility, smaller geometrical sizes and

possible alignment to the liquid flow. On the other hand, as shown in Figure A3-3, MWCNTs

in KCl solutions showed slightly lower penetrations than those in DI water and the result is

consistent with results in the literature (96) and can be explained by the DLVO calculation

presented in Chapter 2. However, delivered liquid flow through the filtering flask using a

vacuum flow was not stable during filtration process, because of rapid changing pressure drop

on the filter with increasing amount of deposited CNTs. Therefore the peristaltic pump was

employed for further filtration experiments.

157

Figure A3-3. Penetrations of CNTs in the suspensions with different electrolytes such as KCl

and CH3CO2NH4 determined by the UV-Vis spectrophotometry.

The peristaltic pump provided consistent results in several times of trials and the results were

shown in Figure 14. UV-Vis spectrophotometer was used to determine MWCNT

concentrations upstream and downstream of the filter. In Figure 14(a), the penetration

increases with increasing of CH3CO2NH4 concentration, however, when the ionic

concentration increases above the threshold the penetration decreases immediately. These

changes are related to acetic electrolyte effects as introduced in Chapter 2. According to the

results, either penetration or collections can be selectable with control the concentration of

electrolytes in the suspensions depending on the purpose, for example, dispersion for CNT

applications or filtration for the waste water treatment. The aerosolization-quantification

158

method showed a good agreement with UV-Vis spectrophotometry and it is shown in Figure

14(b). Adding CH3CO2NH4 into the suspensions brought increasing penetrations of COOH-

MWCNTs and good agreement with the better dispersity shown in Chapter 2. DI water result

crossed with result for 0.021 M, and that might be caused by electrolyte effect. Since

CH3CO2NH4 affects not only to CNT agglomerates but also to the interaction between CNT

and filter surface, the CNT size is dominant without any electrolyte effect for DI water sample.

Thus difference between 100 nm and 200 nm mobility sizes is bigger for DI water sample

than 0.021 M sample with CH3CO2NH4. However, although two results were crossing each

other, they were reasonable, because overall efficiencies for two samples were similar as

shown in Figure 14(a). It can be seen that penetration is also possible to control with

controlling electrolytes in the suspension as this study presented.

159

Yeon Kyoung Bahk

Academic Education

Doctor of Science (Passed the final PhD examination on October 21, 2015; approved on December 9, 2015)

Swiss Federal Institute of Technology in Zurich (ETHZ)

Supervisors: Dr. Jing Wang; Professor, Institute of Environmental Engineering, ETHZ, Switzerland

Dr. Sotiris E. Pratsinis; Professor, Institute of Process Engineering, ETHZ, Switzerland

Committees: Dr. Alfred Weber; Professor, Institute of Particle technology, TU Clausthal, Germnay

Dr. Jasmin Aghassi; Professor, Institute of Nanotechnology, KIT, Germany

Thesis: Physical characterizations of carbon nanotubes for the emission control and exposure modeling

Master of Mechanical Engineering, August 2008

Pusan National University (PNU)

Supervisor: Dr. Jeung Sang Go; Associate Professor, School of Mechanical Engineering, PNU

Thesis: Study on abrasive focusing nozzle for precision machining by using high pressure water

Bachelor of Mechanical Engineering, August 2006 (Magna Cum Laude)

Pusan National University (PNU)

Thesis: Direct writing of waveguide in the glass slide by using UV laser

Experiences 12.2010-Present: Research and teaching assistant in ETH Zurich/Scientific Guest at EMPA (Swiss Federal Institute

of Material)

03.2011-10.2015: PhD candidate in ETH Zurich

09.2008-11.2010: Research assistant in Next MEMS lab.

09.2006-08.2008: Master course in PNU

03.1999-08.2006: Bachelor’s course in PNU (including military service; 01.2001-03.2003)

Date of birth : May 26, 1980

Citizenship : Republic of Korea

Address : Höglerstrasse 43, CH-8600 Dübendorf, Switzerland

Tel : +41 78 635 9124

Email : [email protected]

160

Research Interests Nano/Micro particle mechanic, characterization, synthesis, toxicology; Nano/Micro-Thermo-fluidic; Nano/Micro

technology for nanoparticle applications and bio-sensors; Harsh Environment Monitoring Systems; Biophysics.

Publications A. Publications included in the PhD thesis

[1] Bahk YK, Chen SC, Pui DYH, Wang J. Improved filtration models for carbon nanotube agglomerates. Under

preparation process to submit for publication.

[2] Bahk YK, Sachinidou P, Gitsis E, Wang J. Lung deposition model of airborne multi-walled carbon nanotubes

for inhalation exposure assessment. Under preparation process to submit for publication.

[3] Bahk YK, He X, Kuo YY, Kim N, Gitsis E, Wang J (2015) Enhanced dispersion stability and mobility in porous

media of carboxyl functionalized carbon nanotubes in aqueous solutions through strong hydrogen bonds. Journal of

Nanoparticle Research 17:396.

[4] Wang J, Bahk YK, Chen SC, Pui DYH (2015) Characteristics of airborne fractal-like agglomerates of carbon

nanotubes. Carbon 93:441-450.

[5] Chen SC, Wang J, Bahk YK, Fissan H, Pui DYH (2014) Carbon nanotube penetration through fiberglass and

electret respirator filter and Nuclepore filter media: experiments and models. Aerosol Science and Technology

48(10):997-1008.

[6] Bahk YK, Wang J (2014) Filtration and length determination of airborne carbon nanotubes in the

submicrometer range using nanofiber filters. Aerosol and Air Quality Research 14(5):1352-1359.

[7] Bahk YK, Buha J, Wang J (2013) Determination of geometrical length of airborne carbon nanotubes by electron

microscopy, model, calculation, and filtration method. Aerosol Science and Technology 47(7):776-784.

B. Others

[1] Sachinidou P, Bahk YK, Wang J. An integrative model for the filtration efficiencies in realistic tests with

consideration of filtration velocity profile and challenging particle size distribution. Under revision process to

publish in Journal of Aerosol Science.

[2] He X, Nowack B, Mitrano DM, Bahk YK, Schreiner C, Figi R, Wang J. Comparison of the synthetic natural

landfill leachates and their impacts on the agglomeration behaviors of spiked TiO2 nanoparticles. Under preparation

process to submit to Environmental Pollution.

[3] Kim N, Zhao S, Bahk YK, Carboni M, Kim SM, Yoon S, Wang J, Koebel MM. Synthesis of highly

homogeneous porous NiO/YSZ powder via epoxide-initiated sol-gel route. Under preparation process to submit for

publication.

[4] He X, Brem B, Bahk YK, Kuo YY, Wang J. Effects of humidity and types of loading particles on the

performance and service life of cabin air filters in automobiles. Accepted to publish in Aerosol Science &

Technology

[5] Schlagenhauf L, Kuo YY, Bahk YK, Nüesch F, Wang J. (2015) Decomposition and particle release of a carbon

nanotube/epoxy nanocomposite at elevated temperatures. Journal of Nanoparticle Research 17:440.

161

[6] Johnson TJ, Olfert JS, Symonds JPR, Johnson M, Rindlisbacher T, Swanson JJ, Boies AM, Thomson K,

Smallwood G, Walters D, Sevcenco Y, Crayford A, Dastanpour R, Rogak SN, Durdina L, Bahk YK, Brem B and

Wang J (2015) Effective density and mass-mobility exponent of aircraft turbine particulate matter. Journal of

Propulsion and Power 31(2):573-582.

[7] Bahk YK, Kim HH, Park DS, Chang SC and Go JS (2011) A new concept for efficient sensitivity amplification

of a QCM based immunosensor for TNF-α by using modified magnetic particles under applied magnetic field.

Bulletin 32(12):4215-4220.

[8] Yoon DH, Bahk YK, Kwon BH, Kim SS, Kim YD, Arakawa T, GO JS and Shoji S (2011) Improvement of

filtration performance using self-tuning of flow resistance. Japanese Journal of Applied Physics 50(1):017201(6

pages).

[9] Shin BS, Park KS, Bahk YK, Park SK, Lee JH, Go JS, Kang MC and Lee CM (2009) Rapid manufacturing of

SiC molds with micro-sized holes using abrasive water jet. The Transactions of Nonferrous Metals Society of China

19:178-182.

[10] Yoon DH, Ha JB, Bahk YK, Arakawa T, Shoji S and Go JS (2009) Size-selective separation of micro beads by

utilizing secondary flow in a curved rectangular microchannel. Lab Chip 9(1):87-90.

[11] Lee JH, Ha JB, Bahk YK, Yoon SH, Arakawa T, Ko JS, Shin BS, Shoji S and Go JS (2008) Microfluidic

centrifuge of nano-particles using rotating flow in a microchamber. Sensors and Actuators B: Chemical 132(2):525-

530.

Conferences A. Oral presentations

[1] Bahk YK, Wang J, Chen SC, Pui DYH. Fractal geometry and effective density for agglomerates of airborne

carbon nanotubes. EAC 2015, Milano, Italy.

[2] Schlagenhauf L, Kuo YY, Bahk YK, Nüesch F, Wang J. Thermal decomposition of a carbon nanotube/epoxy

nanocomposite and resultant particle release. EAC 2015, Milano, Italy.

[3] Bahk YK, Gitsis E, Wang J. Characterization and multiple path particle dosimetry model calculation of airborne

MWCNTs. IAC 2014, Busan, South Korea.

[4] Chen SC, Wnag J, Bahk YK, Fissan H, Pui DYH. Carbon nanotube penetration through fiberglass and electret

respirator filter media and Nuclepore filter: experiments and models. Conference on Aerosol Technology 2014,

Karlsruhe, Germany.

[5] Chen SC, Wang J, Bahk YK, Fissan H, Pui DYH. Carbon nanotube penetration through different respirator

filters and nuclepore filters: Models and experiments. AAAR 2013, Portland, US.

[6] Bahk YK, Buha J, Kim SC, Pui DYH, Wang J. CNT penetration through a screen filter and nano-fiber filter:

Numerical model-ing and comparison with experiments. WFC11, Graz, Austria.

[7] YK Bahk, Ahn C, Shin BS, Ko JS, Go JS. Microchannel and hole fabrication method on a bioglass using the

focusing flow of microparticles, Korean MEMS conference 2009. Jeju, South Korea.

[8] Yoon DH, Bahk YK, Arakawa T, Go JS, Shoji S. Filtration of a Fluid from Mixed Microparticles by Using Self-

Tuning of Flow Resistance Microfilter. 9th International Symposium on Microchemistry and Microsys-tems: ISMM

2009, Kanazawa, Japan.

162

B. Poster presentations

[1] Chen SC, Wang J, Bahk YK, Fissan H, Pui DYH. Carbon nanotube penetration through fiberglass and electret

respirator filter and nuclepore filter media: experiments and models. IAC 2014, Busan, South Korea.

[2] He X, Bahk YK, Wang J. Comparison between Electrical Mobility Based Aerosolization Technique and

Spectrophotometry in the Measurement of Filtration Efficiency of Liquid-borne CNTs. Conference on Aerosol

Technology 2014, Karlsruhe, Germany.

[3] Bahk YK, Wang J. Determination of geometrical length of airborne carbon nanotubes by filtration method. EAC

2013. Prague, Czech.

[4] Bahk YK, Wang J. Filtration of airborne carbon nanotubes with nanofiber filters. EAC 2012, Granada, Spain.

[5] Bahk YK, Buha J, Liu Z, Pui DYH, Wang J. Relationship between the geometrical length and mobility diameter

of airborne carbon nanotubes. EAC 2011, Manchester, UK.

[6] Bahk YK, Chang SC, Kim BH, Lee DS, Kim HH, Go JS. A new concept for efficient sensitivity amplification

of a QCM based immunosensor for TNF-α by using modified magnetic particles under applied magnetic field.

Biosensors 2010, Glasgow, UK.

[7] Ha JB, Yoon DH, Park SY, Bahk YK, Arakawa T, Shoji S, Go JS. Size-selective separation of glass beads using

the secondary flow in a curved microchannel. Micro-TAS 2007, Paris, France.

[8] Ha JB, Bahk YK, Yoon SH, Lee JH, Jeong EH, Yoon SY, Arakawa T, Ko JS, Shin BS, Kim KC, Boo JS, Shoji

S, Go JS. Microfludic centrifuge of nanoparticles using rotating flow in microchamber. Transducers 2007, Lion,

France.

[9-15] 7 Abstracts have been submitted on Korean domestic conferences (in 2006-2008, 3 papers on KMEMS, 4

papers on KSPE (Korean Society for Precision Engineering))

Patents [1] B.K. Park, J.S. Go, K.S. Park, Y.K. Bahk, B.S. Shin, G.S. Kim, C.M. Lee, J.H. Jang, Water jet cutting device

capable of reducing surface roughness by increasing concentration degree of abrasives, Samsung Electro-Mechanics,

June 2008: KR 1020070072839.

[2] B.K. Park, J.S. Go, K.S. Park, Y.K. Bahk, B.S. Shin, G.S. Kim, C.M. Lee, J.H. Jang: Water jet cutting device,

Samsung Electro-Mechanics January 2009: CN 200710308329.

[3] J.S. Go, S.C. Chang, Y.K. Bahk, H.H. Kim, J.M. Cha, B.H. Kwon, C.H. Ahn, Improvement method of measure

minute mass, Pusan National University, March 2010: KR 1020100029230.

Scholarships and Fellowships A. Research assistant (PhD candidate at ETHZ, Assistant researcher in EMPA, Switzerland and Next MEMS

Lab in PNU, Korea, Master’s course at PNU)

[1] PhD candidate at ETHZ and assistant researcher in EMPA (12.2010-present)

1. Novel techniques for nanoparticles and nanofiber filtration and applications in pollution control and health

protection, Sponsored by the Center for Filtration Research in University of Minnesota (12.2010-present).

2. Particulate matter and gad phase emission measurement of aircraft engine exhaust, Sponsored by the Swiss

Federal Office of Civil Aviation (04.2012-08.2013).

163

3. Methodology to determine effectiveness of filtration media against nanoparticles in the size range of 3 to 500

nanometer, in the framework of the EU mandate M/461 “Standardization activities regarding nanotechnologies

and nanomaterials”, Sponsored by the European Committee for Standardization (06.2013-present).

[2] Assistant researcher in Next MEMS Lab. in PNU and Master’s course at PNU (09.2006-11.2010)

1. Development of PCB cutting technique by using water-jet, Sponsored by Samsung Electro-mechanics

(09.2006~08.2007).

2. Technique of removing air bubble from epoxy in optical ferrule, Sponsored by Shinhan Photonics

(11.2006~10.2007).

3. Design of TV tilting structure on the refrigerator R-door, Sponsored by LG Electronics (05.2007~04.2008).

4. Design of abrasive focusing nozzle in micro water-jet, Sponsored by Samsung Electro-mechanics & Small and

Medium Business Administration (11.2007~11.2009).

5. Research on reduction of flow resistance using micro bubble attached on a plate surface, Sponsored by

Ministry of Science and Technology (06.2008~07.2010).

6. Development of High Temperature & High Pressure Sensor for Harsh Environment, Sponsored by Rolls-

Royce UTC in Pusan National University (12. 2007~07.2010).

B. Teaching assistant - 5 semesters;

1. Nanoparticle Mechanics: 4 semesters (1st 2012, 1st 2013, 1st 2014 and 1st 2015, ETHZ)

2. Technology and Environment and Computer Laboratory: 4 semesters (1st 2012, 1st 2013, 1st 2014 and 1st 2015,

ETHZ)

3. Air Quality and Health Impact: 1 semester (1st 2014, ETHZ)

4. Micro Fluidics: 1 semester (2nd 2006, PNU).

C. Scholarship by NURI BEAM (New University for Regional Innovation-Busan Educational Alliance of

Mechanical Engineering) - 3 semesters (1st and 2nd 2007, 1st 2008; Master’s course, PNU)

D. Scholarship for excellent grade students - 4 semesters (1st and 2nd 2004, 1st 2005, 1st 2006; Bachelor’s course,

PNU)

Supervision of Student Projects 1. Emmanouil Gitsis (Master project), Density measurement of powders composed of nanoparticles. 2012

2. Emmanouil Gitsis (Master thesis), Penetration of carbon nanotubes through nuclepore filters in liquids. 2013.

3. Xu He (Master thesis), Filtration of liquid-borne nanoparticles by micro-structural devices and mechanism study.

2013.

4. Viktor Carp (Bachelor thesis), Fabrication of nanofiber on the micro-chip for biological and environmental

applications. 2015.

Services: Peer reviews

1. Journal of Membrane Science

2. Powder Technology

3. Journal of Visualized Experiments

164

Skills Nanoparticle Measurement (SMPS, CPMA, FMPS, OPC, APS, NSAM, etc.), Measuring Equipment (SEM,

Microscopy, PIV, QCM, UV-Vis, DLS, etc.), MEMS Process (Lithography, Wet etch, DRIE, E-beam evaporating,

Sputtering, etc.), Fabrication Process (UV Laser, Water-jet, Mechanical machining, etc.).

Computer Skills

Multi physics analysis (CFD-ACE, Fluent, ANSYS, ANSYS Workbench), Matlab, CAD (CATIA, AutoCAD).

Academic Courses

Nanoparticle Mechanics and Technology, Micro- and Nanoparticle Technology, Fluid mechanics, Microfluidics,

Microfluidic systems, Computational fluid mechanics, Thermodynamics, Heat transfer, Thermal system engineering,

Internal combustion engine, Solid mechanics, Application of electronics, Physics, Chemistry, Micro/Nano

engineering, MEMS process, etc.

Referees 1. Title: Professor

Name: Jing Wang

Institution: Institute of Environmental Engineering, ETHZ

Address: HIF C46.2, Schafmattstr. 6, CH-8093 Zürich, Switzerland

Phone: +41 44 633 36 21

Email: [email protected]

2. Title: Professor

Name: David Y.H. Pui

Institution: Department of Mechanical Engineering, University of Minnesota

Address: Address: ME 3101F, 111 Church Street SE, Minneapolis, MN 55455, USA

Phone: +1 612-625-2537


3. Title: Professor

Name: Jeung Sang Go

Institution: School of Mechanical Engineering, Pusan National University

Address: San30, Jangjeon-dong, Geumjeong-gu, Busan 609-735, Republic of Korea

Phone: +82-51-510-3512


rights / license: research collection in copyright - non ... · 2 filtration is an effective method...

Documents