Control of Hexavalent Chromium Emission

1

CONTROL OF HEXAVALENT CHROMIUM EMISSION FROM WELDING USING SILICA PRECURSORS AS SHIELD GAS ADDITIVES

By

NATHAN E. TOPHAM

A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT

OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ENGINEERING

UNIVERSITY OF FLORIDA

2010

2

2010 Nathan Topham

3

To my family

4

ACKNOWLEDGMENTS

I would like to thank my graduate advisor and committee chair, Dr. Chang-Yu Wu,

for his unending patience and support during my studies. Without his guidance and hard

work, this project would not have been possible. I would also like to thank my committee

members, Dr. Jean-Claude Bonzongo and Dr. Vito Ilacqua, for their suggestions and

support. Additionally, Id like to thank my graduate mentor, Dr. Yu-Mei Hsu, for her

support during my studies. Her leadership helped inspire me to pursue a graduate

degree. I would also like to extend my gratitude to all of my fellow undergraduate and

graduate lab mates who helped throughout this research. Finally, Id like to thank my

better half, Chelsea, for supporting me during the difficult times during my studies and

helping me find the strength to continue striving towards my goal when I faltered.

5

TABLE OF CONTENTS page

ACKNOWLEDGMENTS .................................................................................................. 4

LIST OF TABLES ............................................................................................................ 7

LIST OF FIGURES .......................................................................................................... 9

LIST OF ABBREVIATIONS ........................................................................................... 10

ABSTRACT ................................................................................................................... 11

CHAPTER

1 INTRODUCTION .................................................................................................... 13

Fundamentals of Arc Welding ................................................................................. 13 Physical and Chemical Properties of Welding Fumes ............................................ 14 Health Effects of Welding Fume Exposure ............................................................. 15 Welding Fume Sampling and Analysis.................................................................... 18 Past Attempts to Control Welding Fume Exposure ................................................. 19 Sorbents in Combustion Systems ........................................................................... 20 Objective of Study ................................................................................................... 22

2 EXPERIMENTAL METHODS ................................................................................. 24

Overview of Experimental Methods ........................................................................ 24 GTAW Fume Generation and Sampling ................................................................. 25 GTAW Experimental Conditions ............................................................................. 26 GTAW Sample Analysis .......................................................................................... 26 GMAW Fume Generation and Sampling................................................................. 27 GMAW Experimental Conditions ............................................................................ 29 GMAW Sample Analysis ......................................................................................... 30

3 RESULTS AND DISCUSSION GAS TUNGSTEN ARC WELDING (GTAW) ....... 35

Results .................................................................................................................... 35 Discussion .............................................................................................................. 36 Cost Analysis .......................................................................................................... 37

4 RESULTS AND DISCUSSION GAS METAL ARC WELDING (GMAW) .............. 39

TMS Cr6+ and Particle Size Distribution Results ..................................................... 39 30 Lpm Shield Gas Flow Rate .......................................................................... 39 30 Lpm Shield Gas Flow Rate Using High Voltage .......................................... 42 25 Lpm Shield Gas Flow Rate .......................................................................... 43

6

20 Lpm Shield Gas Flow Rate .......................................................................... 44 Effects of Shield Gas Flow Rate ....................................................................... 45

TMS Cost Analysis ................................................................................................. 46 Regulatory Compliance........................................................................................... 47

5 CONCLUSIONS ..................................................................................................... 61

APPENDIX

A DETERMINATION OF THEORETICAL MINIMUM PRECURSOR CONCENTRATION IN FUME PARTICLE FORMATION AREA ............................. 63

B DETERMINATION OF TMS AND TEOS VAPOR PRESSURE .............................. 66

C DETERMINATION OF PRECURSOR CONCENTRATION IN SATURATED CARRIER GAS ....................................................................................................... 67

D COST ANALYSIS ................................................................................................... 69

GTAW Operating Cost of TEOS Addition ............................................................... 69 GMAW Operating Cost of TMS Cost Addition ........................................................ 69

E RAW DATA FOR HEXAVALENT CHROMIUM ....................................................... 71

F SCANNING MOBILITY PARTICLE SIZER PARTICLE SIZE DISTRIBUTION DATA ...................................................................................................................... 73

G CASCADE IMPACTOR CUT SIZE AND PARTICLE SIZE DISTRIBUTION DATA ...................................................................................................................... 87

LIST OF REFERENCES ............................................................................................... 89

BIOGRAPHICAL SKETCH ............................................................................................ 96

7

LIST OF TABLES

Table page 2-1 GMAW experimental conditions. ........................................................................ 34

2-2 TMS molar flow rates used during GMAW sampling. ......................................... 34

4-1 Qualitative elemental analysis of Figure 4-6 fume particles. ............................... 54

4-2 Fume generation rate using various TMS feed rates. ......................................... 55

4-3 Baseline particle size characteristics at different shield gas flow rates. .............. 59

A-1 Particle size data used to calculate minimum precursor concentration, based on particle size distribution from literature (Zimmer and Biswas 2001). .............. 64

B-1 Antoine coefficients for TMS and TEOS. ............................................................ 66

D-1 Cost analysis data for GMAW with TMS as a shield gas additive. ...................... 70

E-1 Raw Cr6+ data for baseline and TMS samples. ................................................... 71

F-1 SMPS particle size data for baseline 20 Lpm shield gas flow rate...................... 74

F-2 SMPS particle size data for 20 Lpm shield gas flow rate with 1.05% TMS carrier gas........................................................................................................... 74









8



F-13 Raw SMPS particle size distribution data for 20 Lpm shield gas flow rate. ........ 78



G-1 Impactor data used for cut size calculations. ...................................................... 87

G-2 Particle size data measured using cascade impactor. ........................................ 88

9

LIST OF FIGURES

Figure page 1-1 Welding torches. ................................................................................................. 23

1-2 Formation of welding particles with and without TMS addition. .......................... 23

2-1 GTAW fume generation and sampling system. .................................................. 31

2-2 GTAW torch modified to allow TEOS addition. ................................................... 32

2-3 GMAW fume generation and sampling system................................................... 33

3-1 GTAW Cr6+ and nitrate data before and after TEOS addition. ............................ 37

3-2 TEM images of GTAW fume particles. ............................................................... 38

4-1 Average Cr6+ mass as a function of TMS carrier gas flow rate at 30 Lpm total shield gas flow rate. ............................................................................................ 47

4-2 GMAW fume particle size distributions at 30 Lpm shield gas flow rate. .............. 48

4-3 GMAW fume particle size distribution under 4.2% TMS flow rate measured with cascade impactor. ....................................................................................... 49

4-4 GMAW fume agglomerate without SiO2 coating. ................................................ 50

4-5 GMAW fume agglomerate with SiO2 coating indicated by arrows. ..................... 51

4-6 GMAW primary particles coated with SiO2. ........................................................ 52

4-7 Amorphous SiO2 agglomerate fume particle. ...................................................... 53

4-8 Intercoagulation between metal particles and SiO2 agglomerate. ...................... 54

4-9 Average Cr6+ mass in GMAW fumes using globular metal transfer mode. ......... 55





4-14 Particle size trends at different shield gas and TMS carrier gas flow rates. ........ 60

10

LIST OF ABBREVIATIONS

AWS American Welding Society

cfm Cubic feet per minute

Cr6+ Hexavalent chromium

EDS Energy dispersive x-ray spectroscopy

GMAW Gas metal arc welding

GTAW Gas tungsten arc welding

IC Ion chromatography

ICP-AES Inductively-coupled plasma atomic emission spectrometry

Lpm Liters per minute

mmHg Millimeter mercury

mM Millimolar

MMD Mass median diameter

NIOSH National Institute for Occupational Safety and Health

nm Nanometer

OSHA Occupational Safety and Health Administration

ppm Parts per million

SMPS Scanning mobility particle sizer

TEM Transmission electron microscopy

TEOS Tetraethyloxysilane (CAS # 78-10-4)

TMS Tetramethylsilane (CAS # 75-76-3)

g Microgram

g/m3 Microgram per cubic meter of air

m Micrometer

UV Ultraviolet

11

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the

Requirements for the Degree of Master of Engineering

CONTROL OF HEXAVALENT CHROMIUM EMISSION FROM WELDING USING SILICA PRECURSORS AS SHIELD GAS ADDITIVES

By

Nathan E. Topham

August 2010

Chair: Chang-Yu Wu Major: Environmental Engineering Sciences

Hexavalent chromium (Cr6+) emitted from arc welding poses serious health risks to

workers exposed to welding fumes in occupational settings. Stainless steel gas

tungsten arc welding (GTAW) and gas metal arc welding (GMAW) produce aerosols

that contain Cr6+, nickel, manganese, and gaseous pollutants such as ozone and

nitrogen oxides (NOx) that lead to a number of respiratory and neurological ailments as

well as cancer. Traditional welding technologies employ shield gas or material

incorporated into the welding wire to minimize fume formation. However, significant

amounts of hazardous air pollutants (HAPs) are generated during gas shielded welding

processes.

In this study, tetraethyloxysilane (TEOS) and tetramethylsilane (TMS) were added

to GTAW and GMAW shield gas to provide a two-fold approach at controlling HAPs

produced during stainless steel welding. Silica precursors acted as reducing agents

when they decomposed in the high temperature welding arc, thereby limiting ozone and

Cr6+ formation. Additionally, an amorphous film of silica (SiO2) was deposited on

welding fume particles. This film insulated the aerosols, preventing subsequent

12

oxidation of surface chromium and masking the toxic effects of all of the metals

contained in the particles.

Cr6+ in GTAW fumes was reduced by 45% when 26.3% of the shield gas was used

as TEOS carrier gas. A 53% reduction in NOx emissions was observed, indicating a

reduction in reactive oxygen species. Tranmission electron microscope (TEM) imagery

showed a film of SiO2 on welding fume particles. The SiO2 film could insulate GTAW

aerosols, masking the toxic effects of all metals within the fume particles.

Experimental results demonstrated that total shield gas flow rate impacted the

effectiveness of TMS as a GMAW shield gas additive. Low shield gas flow rates led to

premature thermal decomposition of TMS and no observable change in Cr6+ mass in

fumes. Increasing shield gas flow rate led to significant reductions in Cr6+ mass when

TMS was used. When 4.2% of the 30 liter per minute (Lpm) shield gas flow rate was

used as TMS carrier gas, Cr6+ mass in GMAW fumes was reduced by at least 92.9%.

Adding TMS to GMAW shield gas increased fume particle size beyond the

nanometer range in all shield gas flow rates tested. Geometric mean particle size

increased from 40-60 nanometers (nm) under baseline conditions to 180-300 nm when

TMS was added. SiO2 particles formed from decomposition of silica precursors

scavenged nanoparticles through intercoagulation. TEM imagery provided visual

evidence of an amorphous film of SiO2 on some fume particles along with the presence

of large amorphous SiO2 agglomerates. These results showed that vapor phase

sorbents may be capable of reducing the health risks posed by welding fumes through

elimination of nanoparticles and reduction of Cr6+.

13

CHAPTER 1 INTRODUCTION

Fundamentals of Arc Welding

Gas metal arc welding (GMAW) and gas tungsten arc welding (GTAW) are

commonly used welding processes that use mild or stainless steel wire as filler material

to join pieces of metal. Shield gas is used to protect the superheated weld site from

gaseous species in air that degrade the mechanical properties of the weld and

destabilize the welding arc. The operator controls welding parameters such as voltage,

shield gas composition, shield gas flow rate, and wire feed speed to obtain desired weld

characteristics. As shown in Figure 1-1, the GTAW electrode is non-consumable

tungsten and a separate rod is used for filler material while in GMAW the consumable

steel wire acts as the electrode and filler material.

Operating parameters influence the properties of the welding arc and welded

material. Voltage and metal transfer mode are related, with low voltage conditions

creating short circuit and high voltage creating globular and spray transfer (Soderstrom

and Mendez 2008). During globular metal transfer mode, droplet detachment from the

electrode is chaotic and large amounts of spatter and fumes can result from operating

under this metal transfer mode (Simpson 2009; Pires et al. 2007). Choice of shield gas

can also impact the characteristics of the weld. The addition of carbon dioxide (CO2) to

GMAW shield gas requires higher voltage to maintain arc stability and weld penetration

due to loss of arc heat by conduction by CO2 (Pires et al. 2007). The operating

parameters impact fume formation as well as weld characteristics which creates a

challenge for industry to balance maximum product quality with minimum emissions.

14

Physical and Chemical Properties of Welding Fumes

The intense energy of the welding process results in the formation of fumes

containing a high number concentration of nanosized and sub-micron particles

containing toxic metals as well as a number of gaseous species including ozone and

nitrogen oxides (Jenkins et al. 2005; Hewett 1995a; Liu et al. 2007). Fume aerosols

generated during arc welding processes are typically less than 1 micrometer (m) in

diameter and the primary particles are often in the nanometer range (Jenkins et al.

2005; Biswas and Wu 2005; Zimmer et al. 2002; Stephenson et al. 2003). The fume

characteristics are influenced by a variety of parameters, including shield gas

composition, welding wire composition, voltage, and metal transfer mode (Zimmer et al.

2002; Hovde and Raynor 2007). Increasing the shield gas flow rate yields higher

hexavalent chromium (Cr6+) formation while increasing the amount of CO2 in shield gas

reduces Cr6+ formation (Dennis et al. 1997). Fume formation increases as welding

voltage is increased (Hovde and Raynor 2007). GMAW processes are more commonly

used and have much higher fume formation rates than GTAW (Serageldin and Reeves

2009). Thus, fume formation from GMAW has been studied much more thoroughly than

GTAW.

The chemical composition of welding aerosols depends on the composition of the

welding wire used as well as the operating parameters. Mild steel welding fumes

generally consist of iron and manganese oxides (Minni et al. 1984; Jenkins and Eagar

2005). Stainless steel welding fumes contain manganese and iron; however, they also

contain chromium and nickel oxides (Castner and Null 1998; Heung et al. 2007). About

1-5% of the chromium found in GMAW fumes is in the hazardous hexavalent state

(Heung et al. 2007; Serageldin and Reeves 2009). These metals are layered in a

15

core-shell structure. Fume particles typically contain a core made up primarily of iron

with other metals making up the outer layers along with surface enrichment by lighter

elements such as silicon, chlorine, and fluorine (Konarski et al. 2003a,b; Maynard et al.

2004).

Welding uses a shield gas to maintain arc stability, produce desirable weld

penetration, and reduce fume formation rate (Dennis et al. 1997; Ebrahimnia et al.

2009). The shield gas chosen has an impact on ultraviolet (UV) light, ozone generated,

particle size distribution of aerosols, and the amount of Cr6+ (Zimmer et al. 2002; Dennis

et al. 1997). The use of shield gas in GMAW decreases overall fume formation rate;

however, it may increase the formation rate of some of the toxic species in welding

fumes. Dennis et al. (1997) studied the effect of shield gas flow rate on formation rates

of Cr6+, UV light, ozone, and total fume. As the shield gas flow rate increased, the total

fume formation rate decreased. However, the formation rates of UV light, ozone, and

Cr6+ all increased with shield gas flow rate. The absence of shield gas leads to low UV

light production and rapid decomposition of ozone through reactions with nitrogen. As

ozone formation increases, the Cr6+ generation rate also increases as ozone is

considered the dominant oxidizer of chromium. The toxic species present in welding

fumes lead to many adverse health effects following occupational exposure.

Health Effects of Welding Fume Exposure

Professional welders are a group that has been the subject of many

epidemiological studies linking occupational exposures to aerosols with adverse health

effects including a wide variety of respiratory and cardiovascular ailments (Lillienberg et

al. 2008; Loukzadeh et al. 2009; Fang et al. 2009; Antonini 2003). A number of the

metallic species present in welding fumes are potentially detrimental to worker health

16

and ambient air quality. Hexavalent chromium causes decreased lung function,

asthma, and a number of other respiratory ailments (Bagchi et al. 2002; Pascal and

Tessier 2004). Cr6+ is also a known human carcinogen, with extensive human and

animal data available (Holmes et al. 2008; IARC 1990; Seel et al. 2007). Studies of

welding emissions in California found that welding is the primary source of airborne Cr6+

in the state (Chang et al. 2004). Nickel is also a known human carcinogen (IARC 1990),

which is linked with oxidative stress after inhalation (Luo et al. 2009).

Manganese (Mn) exposure can cause a number of adverse neurological effects,

including a Parkinsons-like disorder known as manganism (Smargiassi et al. 2000;

Yuan et al. 2006; Antonini et al. 2006; Bowler et al. 2006; Halatek et al. 2005; Bowler et

al. 2007; Flynn and Susi 2009). Although this disorder creates symptoms similar to

Parkinsons disease, it is a distinct disease and exposure to welding fumes is not linked

with increased incidence of Parkinsons disease (Park et al. 2006a; Stampfer 2009).

The prevalence of neurological effects among workers exposed to welding fume

manganese is as high as one in three (Park et al. 2006b). The high risk of developing

serious neurological impairment from exposure to manganese in welding fumes will

likely lead to tightening regulations for this metal. Elimination of metals in welding fumes

from the body is a slow process. The half lives of chromium and nickel in urine are 730

days and 610 days, respectively (Schaller et al. 2007). The health effects from welding

aerosols are compounded by the fact that exposure to multiple potentially toxic

constituents usually occurs at the same time.

The particulate metal species are not the only harmful constituents of welding

fumes. Ozone is present in welding fumes and enhances the toxic effects of the metals

17

in welding fumes by inhibiting defense mechanisms in the human respiratory system

(Cohen et al. 2003). Ozone is produced within a few seconds when welding starts and

persists up to 10 minutes after welding is stopped (Liu et al. 2007). This complicates the

use of local ventilation because pumps are often turned off after welding stops since

particles are only generated during welding while ozone persists beyond the cessation

of welding.

The size and shape of welding aerosols also play a key role in toxicity. A large

percentage of inhaled welding fume particles can deposit in the lungs, typically in the

alveoli (Hewett 1995b). Few welding fume particles deposit in the upper respiratory

system, while deposition in the lower regions of the respiratory system occurs more

readily (Yu et al. 2000; Kleinstreuer et al. 2008). Toxicity of nanoparticles is unique

because the solid particles can be translocated across pathways other than the

respiratory system that are typically not considered for larger particles (Biswas and Wu

2005). For example, manganese nanoparticles can be translocated through the

olfactory nerve directly to the brain where manganese expresses neurotoxic effects

(Elder et al. 2006). Once inhaled, the behavior of the species present in welding fume

particles in biological systems becomes important.

Solubility is an important factor in determining toxicity. Welding fumes have been

divided into soluble and insoluble fractions, and the toxic effects of each fraction have

been examined separately. The soluble fraction of welding fumes is responsible for

toxicity due to oxidative stress and free radical production (McNeilly et al. 2004; Taylor

et al. 2003). It also causes more loss of lung macrophage viability than the insoluble

components (Antonini et al. 1999). The insoluble fraction of welding particles is also

18

responsible for pulmonary damage. Welding particles differ from other metal aerosols in

that the insoluble and soluble fractions of the fumes are both significant contributors to

toxicity (Antonini et al. 2004). The serious adverse health effects of welding fumes have

led regulatory agencies to develop a number of sampling and analytical techniques to

measure the toxic constituents of welding fumes in occupational settings.

Welding Fume Sampling and Analysis

The process of sampling and analyzing welding fumes is complicated with many

variables that must be controlled. The American Welding Society (AWS) has developed

a method of generating and collecting welding fumes (AWS 1999). This technique uses

a welding torch fixed inside a chamber in which fumes are generated, and welding

fumes are collected on a glass fiber filter. The collection efficiency of the glass fiber

filters used in this method has been tested and verified (Quimby and Ulrich 1999). The

AWS hood allows for a number of operating parameters to be fixed, making it ideal for

generating welding fumes while only varying certain operating parameters.

Personal samplers are a common method for collecting welding fumes in actual

occupational settings. These methods typically involve a filter placed near the breathing

zone that is connected to a personal pump. Recent research has shifted to placing

personal samplers inside welding masks to gather samples that accurately represent

particles present in the welders breathing zone (Lidn and Surakka 2009). While there

are a number of National Institute for Occupational Safety and Health (NIOSH) and

Occupational Safety and Health Administration (OSHA) methods for personal sampling

of Cr6+, there is no significant difference between the results obtained by these methods

(Boiano et al. 2000).

19

Analytical techniques used to measure components of welding fumes focus on

measuring Cr6+. NIOSH has developed multiple methods for analyzing Cr6+ in

occupational air that rely on ion chromatography (IC). NIOSH Method 7604 (NIOSH

1994) uses ion chromatography with conductivity detection while Method 7605 (NIOSH

2003) uses ion chromatography with UV detection. Method 7605 is an improvement

over Method 7604 in that the detection limit is reduced from 3.5 micrograms (g) per

sample using 7604 to 0.02 g per sample using 7605. However, metals such as iron

and nickel that are present in welding fumes can potentially interfere with Method 7605

and lead to overestimation of Cr6+ concentration (NIOSH 2003). OSHA Method ID-215

attempts to address the possible interferences by precipitating interfering metals out of

the extraction solution while employing the same analytical technique as NIOSH Method

7605 (Ku and Eide 1998). The sampling and analysis methods for Cr6+ in welding fumes

are necessary to protect workers because of significant adverse health effects that can

result from exposure.

Past Attempts to Control Welding Fume Exposure

OSHA recently reduced the permissible exposure limit for Cr6+ in occupational air

from 52 micrograms per cubic meter of air (g/m3) to 5 g/m3 (OSHA 2006). This

change placed pressure on industry to develop new control technologies to limit

emissions of Cr6+ from arc welding. The simplest method of controlling welding fume

exposure is removing the fumes from the breathing zone of the welder. Studies have

shown that using large vacuum pumps can reduce exposure to metals in welding fumes

by about half if the hose for the pump is placed 2-3 inches from the welding torch

(Meeker et al. 2007). This technology can be invasive and inconvenient for workers and

only removes some fume particles (Flynn and Susi 2009).

20

Previous research has demonstrated that the use of reducing agents as shield

gas additives can reduce ozone formation in the welding fume (Dennis et al. 2002).

Similarly, the addition of reactive metals, such as zinc and aluminum, to welding wires

can reduce Cr6+ formation (Dennis et al. 1996). However, the use of reactive metals was

not effective for controlling Cr6+ formation in some operating conditions. Reactive metal

additives caused an increase in Cr6+ formation in some high voltage operating

conditions. These technologies have not controlled formation of Cr6+ effectively enough

to gain acceptance in industry. Furthermore, nickel and manganese are present in

different oxidation states that have varying solubility and toxicity (Minni et al. 1984;

Jenkins and Eagar 2005). Therefore, traditional shield gas additives that act solely as

reducing agents may not be as effective at limiting the toxicity of these metals in welding

fumes. Mruczek et al. (2008) attempted to reduce manganese exposure by developing

welding wire with reduced Mn content. This study found that Mn content in welding

fumes could be reduced with this process modification; however, the change in

composition of the wire resulted in unacceptable degradation of weld properties. This

study demonstrates the difficulty in controlling hazardous components in welding fumes;

they cannot simply be removed without degrading the needed mechanical properties of

the weld. Therefore, it is necessary to develop a control technology that will limit

exposure to these pollutants while maintaining the integrity of the welded metal.

Sorbents in Combustion Systems

Sorbent injection into combustion systems has been examined as a means for

removing trace metals. Sorbents are capable of scavenging a variety of vapor phase

metals in combustion systems to prevent nucleation of metallic aerosols (Biswas and

Wu 1998). Thermodynamic analyses have been carried out that have shown sorbents

21

to be an effective means for removing vanadium, molybdenum, and arsenic metal

vapors in combustion environments (Lee and Wu 2002; Cho and Wu 2004; Wu and

Barton 2001; Iida et al. 2003). Injection of bulk solid phase sorbents is not feasible for

welding systems. Fortunately, vapor phase sorbents have recently emerged as an

alternative to solid phase materials. Numerous studies have tested the use of vapor

phase silica precursors for control of metal emissions in generic combustion systems.

During combustion, the precursor molecule is broken apart at high temperature. This

process leads to the formation of silica (SiO2) along with carbon dioxide and water.

Silica, formed from pyrolysis and subsequent oxidation of precursor chemicals,

has been proven effective as a sorbent for lead produced during combustion (Biswas

and Zachariah 1997; Owens and Biswas 1996a,b). Silica is capable of forming an

amorphous web that effectively captures ultrafine metal aerosols and increases their

particle size (Biswas and Zachariah 1997; Owens and Biswas 1996a,b; Lee et al. 2005).

The increase in aggregate particle size is beneficial for improved performance of

filtration systems. The silica formed from the injection of silica precursors into

combustion systems is amorphous phase silica rather than crystalline silica (Owens and

Biswas 1996; Jang et al. 2006). Amorphous silica does not cause the toxic effects

observed from inhalation of crystalline silica (Reuzel et al. 1991).

In regards to welding fume control, this technology is superior to traditional shield

gas additives because the silica coating formed around the metallic aerosols insulates

all of the metals in the aerosols, as seen in Figure 1-2, rather than simply manipulating

oxidation state of chromium while failing to address the toxicity of manganese and

nickel. During welding, oxygen that is able to penetrate the shield gas is consumed

22

during the pyrolysis process, thereby reducing the concentration of reactive oxygen

species (O, O3) that would otherwise lead to the formation of Cr6+. The reactions in

Equations 1-1 and 1-2 are simplified versions of the stoichiometric pyrolysis processes

that lead to the formation of SiO2 using two common silica precursors,

tetraethyloxysilane (TEOS) and tetramethylsilane (TMS), respectively. In reality, there

are many possible intermediate reactions leading to the end products (Phadungsukanan

et al. 2009; Herzler et al. 1997).

Si(OC2H5)4 + Oxidant SiO2 + 8CO2 + 10H2O (1-1)

Si(CH3)4 + Oxidant SiO2 + 4CO2 + 6H2O (1-2)

Objective of Study

The overall goal of this study was to investigate the use of silica precursors as

shield gas additives to limit exposure to hazardous metals in fumes from GTAW and

GMAW. The study was conducted in two phases. In phase I, TEOS was tested as an

additive to GTAW shield gas to lower Cr6+ concentration in fumes. Cr6+, nitrate and total

Cr formation in fumes were measured and the ratio of Cr6+/Cr was compared for

baseline welding and welding with TEOS. In phase II, TMS was tested as an additive to

GMAW shield gas because this welding process is more commonly used and

responsible for far more emissions than GTAW. The total shield gas flow rate was

varied to investigate its effect on the efficacy of Cr6+ control. The flow rate of TMS

carrier gas used also was varied in each of the total shield gas flow rates tested to

determine the optimal feed rate. Welding voltage was altered to study whether the

precursors were effective under different operating conditions. Success in this phase of

the experiment was defined as a 90% reduction in the formation of Cr6+ in welding fume

particles.

23

Figure 1-1. Welding torches. A) GTAW torch and welding rod. B) GMAW torch.

Figure 1-2. Formation of welding particles with and without TMS addition.

24

CHAPTER 2 EXPERIMENTAL METHODS

Overview of Experimental Methods

Two types of welding were used in this study. The first method discussed in this

chapter is gas tungsten arc welding (GTAW). The methodology used for gas metal arc

welding (GMAW) is presented afterwards.

GTAW fumes were generated by a welder operated by a user with welding

experience who used a specific mass of welding rod in each sample. The user

maintained similar speeds when consuming welding rod between samples. However,

there was some variation between welding times in different samples. During GTAW

sampling, the welding was performed outdoors with a hood placed directly above the

welding. Ion chromatography and atomic emission spectrometry (ICP-AES) were used

for characterizing how precursors impacted the amount of Cr6+ emitted from GTAW.

Tetraethyloxysilane (TEOS, Acros Organics, 98%, CAS # 78-10-4) was tested during

this phase of the experiment.

GMAW fumes were generated automatically by a torch placed inside a chamber.

In GMAW samples, welding time was held constant as well as the wire feed rate to

consume the same mass of wire between samples. During GMAW sampling, the

welding was performed indoors inside a closed chamber. The closed chamber used in

the GMAW experiments allowed for Cr6+ mass to be compared directly between

samples via ion chromatography, eliminating the need for measurement of total

chromium. Tetramethylsilane (TMS, Acros Organics, 99.9%, CAS # 75-76-3) was tested

during this phase of the experiment.

25

GTAW Fume Generation and Sampling

Figure 2-1 shows the schematic of the welding and sampling system. A Miller

Maxstar 150 STL gas tungsten arc welder was used for producing the welding aerosols.

ER 308L stainless steel welding rods were used, which have an average chromium

content of 19.5%-22.0%. This rod also contained 9.0-11.0% nickel, 1.0-2.5%

manganese, as well as 0.35-0.65% silica added to stabilize the iron in the weld. The

remainder of the welding rod was iron. Forty-five grams of welding rod were consumed

per sample. Mild steel base metal was used for sampling in order to avoid interference

from chromium emitted from the base sheet metal.

The welding hoses were modified to allow injection of TEOS, shown in Figure 2-2.

A Y-fitting was added to the welding torch where the hose joined the welding gun. Wires

were connected to the end of the welder gas hose and the base of the welding torch to

maintain the electric current after the Y-fitting was added. The gas hose leaving the

shield gas cylinder was connected to a Y-fitting to split the gas flow into two parts. The

main gas flow was passed through the welder as in normal operation while the

remainder of the shield gas was passed over a reservoir of liquid TEOS.

The TEOS feed rate was controlled by varying the flow rate of carrier gas that was

passed over the liquid TEOS reservoir, maintained at 100 degrees Celsius (C), with a

mass flow controller. The recommended minimum gas flow rate designated for this

model of welder is 7.1 liters per minute (Lpm), which was adopted in this study. A 6x12

rectangular hood was placed approximately 18 directly above the mild steel base metal

on which the welding occurred. A 90 millimeter (mm) diameter glass fiber filter (Millipore

APFA 090) was placed at the top of the hood and connected to a high volume (Hi-Vol,

44 cfm) pump for collecting the aerosols.

26

Ultra high purity argon (Airgas, 99.999%) was used as the welding shield gas. This

is the shield gas most commonly used in GTAW to maintain arc stability during welding

operations. It acted as the needed carrier gas without requiring a secondary carrier gas.

The vapor pressure of TEOS at room temperature is only 2 millimeters mercury

(mmHg), which was too low to meet the theoretical amount of Si needed to produce a

mono-atomic layer on the welding aerosols, determined in Appendix A. A water bath

(100 C) was used to heat the reservoir and raise the vapor pressure of the liquid TEOS

to 86 mmHg (Stull 1947) in order to generate sufficient vapor with a reasonable shield

gas flow rate passing over the liquid.

GTAW Experimental Conditions

Baseline samples using only argon as a shield gas were run to determine

emissions of Cr6+ relative to total chromium fume formation during normal welding

operation. This scenario represented the standard welding method currently used in

industrial environments. The baseline samples ensured that the sampling time was long

enough to provide concentrations needed for the analytical methods. For the

experimental set, the TEOS carrier gas feed rate was set at 26.3% of the total 7.1 Lpm

shield gas flow rate, resulting in a TEOS molar flow rate of 6.9610-3 gram-mol per

minute (gmol/min). Three TEOS samples and six baseline samples were collected.

GTAW Sample Analysis

Inductively coupled plasma with atomic emission spectroscopy (ICP-AES, Perkin

Elmer Plasma 3200) was used to measure total chromium. This analytical method does

not differentiate between trivalent and hexavalent chromium. Therefore, IC (Dionex ICS

1500, IonPac AS9-HC analytical column, DS6 conductivity detector) was used to

measure the soluble hexavalent chromium species, chromate (CrO42-). IC was also

27

used concurrently for nitrate analysis. Using these methods in conjunction provided the

needed data to determine the reduction of Cr6+ species using the TEOS additive.

Prior to analysis, samples were extracted from filter media using techniques

specific to the analytical method used. Sample extraction for IC followed a modified

NIOSH Method 7604 (NIOSH 1994). Soluble Cr6+ species were extracted using a

9 millimolar (mM) sodium carbonate solution and were placed in a water bath at 100 C

for 1 hour. For ICP-AES analysis, acid digestion was used, based on previously verified

techniques (Karanasiou et al. 2005). Aerosols and silica coatings, along with the filter

media, were dissolved using 9 milliliters (mL) of 65% nitric acid and 1 mL of 48%

hydrofluoric acid. The acidic extraction was placed in a heat block for five hours at

150 C. Afterwards, 1 mL of 10% boric acid was added to neutralize the free fluorine in

the solution.

Additionally, transmission electron microscopy (TEM, Model 2010F, JEOL) was

used to provide images that depicted the SiO2 coating formed on fume particles.

Specialty grids designed for TEM (Pelco, Lacey Carbon Type-A, 300 mesh) were held

directly in the fumes to collect particles for the TEM analysis.

GMAW Fume Generation and Sampling

Figure 2-3 shows the schematic setup of the welding fume generation and

sampling system used during baseline and TMS welding. This system followed a

modified AWS fume hood design (AWS 1999). A conical chamber 36 in diameter at the

base, 8 in diameter at the pump, and 36 in height was constructed. A Lincoln

PowerMIG 140C welder was used for producing welding fumes. ER 308L stainless steel

welding wire was used, which has an average chromium content of 19.5%-22.0%. This

wire also contains 9.0-11.0% Ni, 1.0-2.5% Mn, and 0.35-0.65% Si. Mild steel base metal

28

was used for sampling in order to avoid interference from chromium emitted by the base

sheet metal. The welding torch hose was modified to allow injection of TMS. A Y-fitting

was inserted to connect the torch and the gas hose. A longer wire guide was installed to

ensure it was long enough to reach the contact tip. The trigger was removed and the

wires inside the trigger were extended outside the chamber to allow remote operation of

the welding gun. The hose from the shield gas cylinder was connected to a T-fitting to

allow the gas flow to be separated into two parts. The main shield gas flow was passed

through the welder as in normal operation while the remainder of the gas was used as

TMS carrier gas. The recommended minimum shield gas flow rate designated for this

model of welder is 15 Lpm. An initial shield gas flow rate of 30 Lpm was chosen to

produce quality welds after preliminary testing. The silica precursor feed rate was

controlled by varying the flow rate of carrier gas that was passed over the liquid TMS

reservoir, maintained at 0 C, with a mass flow controller. The carrier gas saturated with

TMS vapor was then rejoined to the main shield gas flow prior to reaching the welding

torch. The base metal was placed on a rotating turntable (MK Products Aircrafter T-25)

to maintain a constant weld speed. The turntable was placed in a hood in which the

fumes were generated and collected. A 90 mm diameter glass fiber filter (Millipore

APFA 090) was placed at the top of the hood and connected to a high volume (Hi-Vol,

45 cubic feet per minute (cfm)) pump for collecting the aerosols.

A mixture of 75% argon and 25% CO2 was used as the welding shield gas. The

shield gas acted as the needed carrier gas for TMS without requiring a secondary

carrier gas. TMS is a highly volatile compound with a vapor pressure at room

temperature of 598 mmHg (Aston et al. 1941). Appendix B contains information used to

29

calculate vapor pressure. The TMS reservoir was placed in an ice bath to lower the

vapor pressure to about 270 mmHg. This was done to slow down evaporation of the

TMS while still producing enough TMS vapor to exceed the minimum theoretical TMS

concentration needed to coat the aerosols in a mono-atomic layer of SiO2. See

Appendix A for minimum TMS concentration calculation and Appendix C for calculations

and rationale behind the actual operating TMS concentration produced in the carrier

gas.

GMAW Experimental Conditions

The experimental conditions in this study are summarized in Table 2-1. Baseline

samples using 75% Ar/25% CO2 as a shield gas were run to determine emissions of

Cr6+ during normal welding operation using short-circuit metal transfer. This scenario

represented the standard welding method currently used in industrial environments. The

baseline samples ensured that the sampling time was long enough to produce enough

Cr6+ to demonstrate 90% reduction without being limited by the detection limits of the

analytical methods.

Experimental set A was performed to determine the effect of TMS carrier gas flow

rate on Cr6+ formation. In experimental set B, high voltage was tested to account for

different emission rates of pollutants under globular metal transfer mode. In

experimental sets C and D, total shield gas flow rate was decreased to 25 and 20 Lpm,

respectively, to test the effect of shield gas flow rate on the efficacy of this technology.

The molar flow rates of TMS that correspond to the TMS carrier gas flow rates used in

experimental sets A, B, C, and D were calculated by multiplying the flow rate of carrier

30

gas by the concentration of TMS in saturated carrier gas, found in Appendix C. The

values for molar flow rates tested during GMAW sampling are presented in Table 2-2.

GMAW Sample Analysis

IC (Dionex ICS 1500, CS5A analytical column, DS6 conductivity detector) was

used to measure the soluble hexavalent chromium species, chromate (CrO42-). Prior to

analysis, samples were extracted from filter media using techniques specific to the

analytical method used. Sample extraction for IC followed a modified NIOSH Method

7604 (NIOSH 1994). Soluble Cr6+ species were extracted using a 5 mM sodium

carbonate/1 mM sodium bicarbonate solution and heated in a water bath to 100 C for

one hour. Additionally, TEM and energy-dispersive x-ray spectroscopy (EDS) (Model

2010F, JEOL) were used to provide images that depicted the SiO2 coating formed on

fume particles along with elemental analysis of those particles. Specialty grids designed

for TEM (Pelco, Lacey Carbon Type-A, 300 mesh) were loaded with fume particles for

analysis.

A scanning mobility particle sizer (SMPS, TSI Model 3081 Long DMA) was used to

obtain aerosol size distribution data between 10 nanometers (nm) and 515 nm. Welding

was performed for 10 seconds at which point the SMPS pump was turned on. After the

SMPS completed its 135 second sampling run, the hi-vol pump was turned on to clear

the chamber of particles.

A cascade impactor (U of W Mark III Source Test Cascade Impactor) was used to

obtain particle size distribution data between 0.1 m and 10 m for 30 Lpm shield gas

flow rate with 4.2% TMS. Welding was performed in 10 second increments. After each

10 second increment, the cascade impactor pump was turned on for 135 seconds after

which the hi-vol pump was used to clear the chamber of particles. This process was

31

repeated 15 times. A sampling flow rate of 20 Lpm through the impactor was used to

produce desired cut sizes, calculated in Appendix D. Impaction plates were coated with

Apiezon grease to minimize particle bounce. The grease was mixed with toluene to

facilitate spreading on impaction plates. Impaction plates were painted with the mixture

of grease and toluene and they were baked at 100 C for one hour followed by

desiccation for 24 hours to remove toluene. Plates were measured gravimetrically

(Sartorius MC 210 S, +/- 10 g) before and after loading to measure fume mass at each

cut size. Fume loading on filters used for IC analysis was also measured gravimetrically

for 30 Lpm shield gas flow rate samples. The glass fiber filter used for fume collection

was weighed before and after loading to determine total fume mass.

Figure 2-1. GTAW fume generation and sampling system.

32

Figure 2-2. GTAW torch modified to allow TEOS addition.

Wire

Y-fitting Gas Hose

33

Figure 2-3. GMAW fume generation and sampling system. A) Without TMS. B) With TMS.

34

Table 2-1. GMAW experimental conditions.

Experimental condition

Precursor carrier gas flow rate (% of total shield gas flow rate)

Total shield gas flow rate (Lpm)

Voltage Silica precursor

Baseline 0 20, 25, 30 Low None

A 1.4-4.2 30 Low TMS

B 0, 4.2 35 High TMS

C 5.0 25 Low TMS

D 2.1-6.3

20 Low TMS

Table 2-2. TMS molar flow rates used during GMAW sampling.

Experimental condition

Precursor carrier gas flow rate (% of total shield gas flow rate)

Total shield gas flow rate (Lpm)

TMS molar flow rate (mol/min)

A 1.4, 2.8, 3.5, 4.2 30 0.007, 0.013, 0.017, 0.020

B 4.2 30 0.020

C 5.0 25 0.020

D 2.1, 4.2, 6.3

20 0.007, 0.013, 0.020

35

CHAPTER 3 RESULTS AND DISCUSSION GAS TUNGSTEN ARC WELDING (GTAW)

Results

Data pertinent to chromium, nitrogen oxides, and ozone formation are presented.

As shown in Figure 3-1, the average Cr6+/Cr ratio decreased by 45% from the average

baseline Cr6+/Cr ratio (from 0.170 to 0.094) when TEOS was added. Figure 3-1 also

shows the nitrate concentrations at baseline and TEOS sampling conditions. Nitrate is

the end product of oxidation of atmospheric nitrogen, which is oxidized by the same

reactive oxygen species as chromium. Therefore, it can be used as an indicator of the

formation of reactive oxygen species and oxidation potential within welding fumes

during the welding process (Dennis et al. 2002). Upon the addition of TEOS, nitrate

concentration decreased from an average baseline value of 83 parts per million (ppm)

to 39 ppm. This 53% decrease from baseline conditions is in line with the Cr6+

reduction. This result demonstrates an oxidation potential of chromium within the

welding fume. Previous studies used other reducing agents as shield gas additives. For

example, Dennis et al. (2002) used 3% ethene and 3% nitrogen oxide and obtained

42% and 45% reduction in Cr6+ formation, respectively, which was slightly lower than

what was achieved in this study.

There was large variation in the values shown in Figure 3-1. GTAW is a manual

process in which inconsistencies when operating the welder can impact emission rates.

This can contribute to errors in measuring welding fume emissions that are present

even in automatic systems (Serageldin and Reeves 2009).

Figure 3-2 shows TEM images of fume aerosols collected. The metal vapors

formed during welding quickly condense and form primary particles a few nanometers in

36

diameter. These nanometer primary particles at very high number concentration

undergo rapid coagulation to form aggregates (Fig. 3-2a). Fig. 3-2b shows an aggregate

particle with TEOS addition. In this image, the metallic aerosol is coated in a thin SiO2

film. Metal aerosols appear very dark on the images because the electrons passed

through the sample to generate the image are less able to penetrate dense material like

chromium. Silica is less dense than the metal aerosols and produces a lighter colored

film around the dark aerosols. Figure 3-2c is a closer look at higher magnification. A

primary particle with distinct layering of different metals coated with SiO2 is clearly seen.

These images visually demonstrate that primary and aggregate metal particles formed

during TIG welding were covered in a film of amorphous SiO2.

Discussion

These results support previous studies that showed vapor phase sorbent

precursors increase effective particle size, prevent formation of the ultrafine mode of the

aerosols, and aid in collection efficiency of traditional control technologies (Owens and

Biswas 1996; Biswas and Zachariah 1997). Therefore, the use of less obtrusive

respirators or dust masks may become a more appealing option for workers exposed to

welding fumes. The SiO2 coating prevents or slows down the dissolution of chromium

as nanometer SiO2 takes weeks or more to dissolve in lung fluids (Reuzel et al. 1991).

Since the soluble portions of welding fumes are the most toxic, decreased solubility

decreases the effective toxicity of the aerosols (Antonini et al. 1999). This silica layer

also makes silica precursor compounds more effective reducing agents than other

gaseous species previously studied because oxidation of the metals is prevented long

after formation of the aerosols due to the silica layer.

37

Cost Analysis

The increase in operating costs when using TEOS as a shield gas additive is

calculated in Appendix D. Using lab quality TEOS as an additive at the feed rate tested

in this study creates an additional cost of about $3.80 per hour of welding time. Based

on an estimate of consumption rate used in this study (3.6 grams of consumable used

per minute), the cost of TEOS addition would be roughly $17.60 per kg of welding rod

used. The cost can further be reduced using industrial quality chemical.

Figure 3-1. GTAW Cr6+ and nitrate data before and after TEOS addition.

38

Figure 3-2. TEM images of GTAW fume particles. A) Baseline agglomerate particle without SiO2 coating. B) Agglomerate particle with SiO2 coating when welding with TEOS. C) Primary particle with SiO2 coating when welding with TEOS.

39

CHAPTER 4 RESULTS AND DISCUSSION GAS METAL ARC WELDING (GMAW)

TMS Cr6+ and Particle Size Distribution Results

30 Lpm Shield Gas Flow Rate

A total shield gas flow rate of 30 Lpm was tested to determine the effect of TMS

addition on Cr6+ formation in GMAW fumes. The results indicate adding TMS to shield

gas reduced formation of Cr6+, as seen in Figure 4-1. When 4.2% of the shield gas was

used as TMS carrier gas, the mass of Cr6+ was reduced to below the IC detection limit

of roughly 4.5 g for all replicate samples. Since it was impossible to determine the

exact mass for samples that were below detection, the mass for all non-detectable

samples was set equal to the detection limit. This was a reduction in Cr6+ mass of at

least 92.9% compared to baseline conditions.

The reduction in Cr6+ mass achieved when using TMS as a shield gas additive

exceeded the results seen in previous studies that incorporated reducing agents into

shield gas or welding wire (Dennis et al. 1996; Dennis et al. 2002). The differences in

Cr6+ mass between TMS feed rates were statistically significant (p = 6.8 10-11, ANOVA

single factor, = 0.05).

SMPS particle size data using a total shield gas flow rate of 30 Lpm demonstrated

that adding silica precursors to shield gas increased the particle size of fume particles,

as seen in Figure 4-2. When welding without TMS, the peak in the nanometer size

range at about 20 nm was very large relative to the peak seen at 200-300 nm. The

baseline mass median diameter (MMD) calculated from SMPS data was 380 nm which

agreed very well with previous research that studied this characteristic (Jenkins et al.

2005).

40

As increasing amounts of TMS were fed into the system, the metal nanoparticles

were scavenged by SiO2 agglomerates and the particle size distribution shifted towards

larger particle sizes. The count geometric mean diameter increased from 60 nm out of

the nanometer range to 180 nm as TMS feed rate was increased (Appendix F). The

increase in particle size observed when TMS was added agreed with previous studies

that utilized silica precursors to control lead emissions from combustion systems. In

those studies, adding silica precursors increased particle size out of the nanometer

range (Owens and Biswas 1996a; Owens and Biswas 1996b).

The mass size distribution of welding fume particles with 4.2% TMS additive

collected using a cascade impactor is displayed in Figure 4-3. There was evidence of

overloading on the lower stages of the impactor. This might lead to re-entrainment of

some fume particles that would otherwise not have made it to the final filter.

Nevertheless, the cascade impactor data show the right tail of the particle size

distribution that was too large to be measured using SMPS. The presence of particles

between 1-10 m in diameter further indicated that SiO2 particles were assisting in

coagulation and increasing fume particle size because baseline welding fume particles

formed through nucleation and coagulation in this size range were absent in previous

research (Jenkins et al. 2005).

TEM images were obtained to determine whether a SiO2 coating was present on

fume particles when TMS was used as an additive. The images demonstrate that in

some cases a coating of SiO2 was formed but there were many particles without a

distinct coating, such as those seen in Figure 4-4. The image showed that some surface

enrichment with lighter elements did occur which mirrors results of previous studies of

41

welding fume particle structure that showed aerosol surface enrichment with silicon

chlorine and fluorine (Maynard et al. 2004). Figure 4-5 shows an agglomerate particle

that had a coating of SiO2 that encapsulated the entire agglomerate rather than surface

enrichment of individual primary particles. Figure 4-6 displays individual primary

particles with thick SiO2 coatings. This coating in Figures 4-5 and 4-6 was more

significant than the surface enrichment seen on individual primary particles under

baseline conditions in previous research (Maynard et al. 2004; Konarski et al. 2003).

Elemental analysis of the particles in Figure 4-6 was performed using EDS, shown in

Table 4-1. About 17.5% of the particle mass was silicon, which was much higher than

surface enrichment with silicon (about 5%) in baseline welding fumes measured in

previous research (Minni et al. 1984).

Figure 4-7 shows an amorphous agglomerate particle composed primarily of SiO2.

The TEM imageries obtained show that coating by SiO2 was not uniform for all fume

particles. Some particles were not coated, some primary particles were thickly coated

with SiO2, some agglomerate particles were covered in a thin layer of SiO2, and some

separate amorphous agglomerates composed mostly of SiO2 were all present. Figures

4-5 and 4-6 indicate that silica precursor additives may be capable of reducing toxicity

of all of the metals in welding fume particles through the formation of a SiO2 film on

metal particles. This phenomenon is similar to that observed in previous research where

metal nanoparticles formed in combustion systems can be trapped in a web of

amorphous SiO2 (Owens and Biswas 1996). Metal nanoparticles that are coated in SiO2

are less likely to exhibit toxic effects because the SiO2 coating could take weeks to

dissolve in the respiratory system (Roelofs and Vogelsberger 2004). This would provide

42

the respiratory systems removal mechanisms more time to remove the particles before

exposure to the toxic metals occurs. In addition, there are agglomerates of metal fume

particles with SiO2 particles formed by intercoagulation, as shown in Figure 4-8.

Total fume mass was measured gravimetrically to evaluate the impact of TMS

addition on GMAW fume generation rate, shown in Table 4-2. Although OSHA does not

currently regulate total welding fume concentration, it is important to minimize total

particulate concentration in occupational environments. Meanwhile, NIOSH does have a

recommended exposure limit of the lowest feasible concentration of total welding fume.

A technology that leads to a drastic increase in fume mass would present problems

when trying to produce the lowest feasible concentration of fume particles during

GMAW.

30 Lpm Shield Gas Flow Rate Using High Voltage

Under high voltage, the welding process undergoes globular metal transfer mode.

This process creates violent explosions of metal droplets leaving the tip of the welding

wire. The baseline Cr6+ formation under these conditions was 17.4% higher than using

short circuit metal transfer mode. The addition of TMS did not have as dramatic an

effect on Cr6+ formation as under short circuit conditions, as seen in Figure 4-9. This

may be due to the violence associated with globular metal transfer. Some molten

spatter droplets are ejected far from the arc and may leave the area where TMS

scavenges reactive oxygen species and collected during sampling, leading to an

increase in measured Cr6+ mass from fume formation. Although 90% reduction in Cr6+

mass was not achieved during this type of welding, the 47% reduction was significant

(p-value

43

spatter and fumes produced during this operating condition. Spatter creates additional

labor costs as the spatter particles that deposit on the base metal must be removed with

a grinder. Therefore, the success of TMS addition using short-circuit metal transfer was

more important for reducing Cr6+ emissions from welding sources than during globular

transfer.


Total shield gas flow rate was decreased to 25 Lpm to reduce the consumption

rate of shield gas and associated operating costs. There was a decrease in baseline

Cr6+ mass when the shield gas flow rate was decreased from 30 to 25 Lpm. Figure 4-10

shows that when 5.0% of the shield gas flow rate was used as TMS carrier gas, there

was a significant decrease in Cr6+ mass of about 40% with a p-value far below 0.05

(Students T-test, 2 tails, unequal variance). There was some white powder observed

inside the head of the welding torch after sampling. The reduction in Cr6+ indicated that

some of the TMS was surviving until reaching the area where reactive oxygen species

were present. However, the decreased efficiency of Cr6+ reduction coupled with the

white powder inside the head of the welding torch indicated that some TMS was likely

prematurely decomposing. The welding torch was heating during sampling and the TMS

residence time in the hot torch was long enough for it to thermally decompose prior to

entering the area where fume particle formation was occurring. Compared with 30 Lpm

shield gas flow rate data, this set of experiments shows that total shield gas flow rate

impacted effectiveness of TMS and the goal of 90% reduction in Cr6+ mass was not

reached under 25 Lpm shield gas flow rate.

The SMPS particle size data for 25 Lpm shield gas flow rate followed the same

pattern as that for 30 Lpm, as displayed in Figure 4-11. When welding without TMS, a

44

large peak was seen at about 20 nm and a much smaller peak was present at

200-300 nm. As the TMS feed rate was increased, the peak in the nanometer range

became smaller relative to the peak larger than 0.1 m. The number concentration for

the nanometer range peak at 0.84% TMS carrier gas was higher than the number

concentration for this peak under baseline conditions. However, the height of this peak

relative to the peak at larger particle sizes still follows the same pattern as the 30 Lpm

shield gas flow rate data. As TMS feed rate was increased, the peak in the nanometer

size range shrank relative to the peak above 0.1 m. The aberration in absolute number

concentrations was most likely due to variable fume emission rates encountered during

welding. The shape of particle size distributions followed the same pattern of change

despite the variation in number concentration.


The impact of TMS addition on Cr6+ generation during GMAW using a total shield

gas flow rate of 20 Lpm is shown in Figure 4-12. A large amount of white powder was

observed inside the head of the welding torch after sampling. It likely resulted from

thermal decomposition of TMS inside the head of the welding torch. When this

happened, TMS did not function as a reducing agent or coat fume particles in a SiO2

film because the TMS had already reacted before reaching the area where fume particle

formation was occurring. There was no statistically significant difference between the

amount of Cr6+ mass in the fumes regardless of the amount of TMS that was added

(p = 0.229, ANOVA single factor, = 0.05).

The addition of TMS led to an increase in fume particle size just as it had using

higher shield gas flow rates, as shown in Figure 4-13. During baseline welding, SMPS

measurements showed a large number of particles with an average size of about

45

20 nm. As increasing amounts of TMS were added, this peak decreased and eventually

vanished. The particle size distribution shifted towards larger particles with the

nanoparticles being replaced with particles larger than 0.1 m in diameter. These

results and TEM imagery (Figure 4-8) are consistent with previous modeling research

that found coarse particles such as the large amorphous SiO2 particles formed here are

capable of scavenging nanoparticles through intercoagulation (Lee and Wu 2005). It is

also possible that some metal vapor is scavenged by SiO2 particles, a result

demonstrated in previous research studying the impact of sorbents on metals in

combustion systems. When SiO2 is formed in the area where metals are evaporating,

metal vapor in the combustion system is scavenged, which prevents nucleation.

Nanomaterials can exhibit unique toxic effects based on their size (Biswas and Wu

2005; Jeng and Swanson 2006). Therefore, scavenging nanoparticles is a beneficial

outcome as it may reduce toxicity of welding fume particles by removing the

toxicological mechanisms of these particles that result from their size. The potential

reduction in toxicity based on elimination of nanoparticle pathways would occur

regardless of whether or not metal speciation in the welding fumes changed.

Effects of Shield Gas Flow Rate

The MMD was between 0.35-0.4 m for baseline samples at all shield gas flow

rates tested. The total baseline number concentration decreased and the geometric

mean particle size increased as shield gas flow rate increased, as seen in Table 4-3,

which agrees with observations reported in previous studies (Dennis et al. 1997). It is

possible that particle size increased with increasing shield gas flow rate because

greater dispersion of metal vapor lowered the saturation ratio of metal vapor in the

46

system. This could have caused the rate of nucleation to decrease relative to the rate of

condensation.

As total shield gas flow rate decreased, the reduction of Cr6+ in welding fumes

observed after TMS addition decreased. However, the shift of particle size out of the

nanometer range was observed in all shield gas flow rates tested, as seen in

Figure 4-14. This points to intercoagulation as the dominant mechanism in increasing

particle size rather than SiO2 coating on metal agglomerates. If SiO2 coating was

responsible for increased particle size, Cr6+ concentration would have decreased at all

shield gas flow rates tested and TEM imagery would have shown consistent SiO2

coating of fume particles. This means that Cr6+ reduction was due primarily to TMS

scavenging reactive oxygen species during pyrolysis rather than formation of SiO2

coating on particles containing Cr6+. Intercoagulation between nanosized fume particles

and larger SiO2 particles occurred whether TMS decomposed in the area where fume

particle formation and oxidation occurred or whether TMS decomposed before reaching

the area of fume particle formation. However, the change in particle size due to

intercoagulation would not help meet OSHA standards for Cr6+ concentration. In order

to achieve Cr6+ reduction using TMS, residence time of the precursor vapor in the head

of the welding torch must be short enough to ensure TMS does not heat enough to

decompose before exiting the torch.

TMS Cost Analysis

The operating cost associated with adding TMS through delivery as a vapor from a

liquid reservoir is calculated in Appendix D. The cost of TMS addition at the most

effective flow rate tested was $78.96 per kg of welding electrode consumed. The cost

for TMS addition was high due to the expense of purchasing high purity laboratory

47

grade chemical. The added cost can be reduced by purchasing lower purity industrial

grade chemical.

Regulatory Compliance

The baseline welding samples produced at 30 Lpm shield gas flow rate created an

average Cr6+ concentration of 35.1 g/m3. This concentration was over seven times

higher than the OSHA permissible exposure limit of 5 g/m3 for Cr6+ (OSHA 2006).

When 4.2% of the 30 Lpm shield gas flow rate was used to deliver TMS vapor, the

resultant maximum Cr6+ concentration was 2.5 g/m3. The use of TMS brought the

GMAW process into compliance with OSHA limits for Cr6+ without the use of any local

ventilation.

Figure 4-1. Average Cr6+ mass as a function of TMS carrier gas flow rate at 30 Lpm total shield gas flow rate.

48

Figure 4-2. GMAW fume particle size distributions at 30 Lpm shield gas flow rate.

49

Figure 4-3. GMAW fume particle size distribution under 4.2% TMS flow rate measured with cascade impactor.

50

Figure 4-4. GMAW fume agglomerate without SiO2 coating.

51

Figure 4-5. GMAW fume agglomerate with SiO2 coating indicated by arrows.

SiO2 coating

52

Figure 4-6. GMAW primary particles coated with SiO2.

53

Figure 4-7. Amorphous SiO2 agglomerate fume particle.

54

Figure 4-8. Intercoagulation between metal particles and SiO2 agglomerate.

Table 4-1. Qualitative elemental analysis of Figure 4-6 fume particles.

Element Weight %

Si K 17.66 Cr K 28.06 Mn K 27.93 Fe K 26.35

SiO2

Metal Fume Particles

55

Table 4-2. Fume generation rate using various TMS feed rates.

TMS carrier gas flow rate (% of 30 Lpm shield gas flow rate)

Average mass of fume particles standard deviation (mg)

0 29.5153.612

1.4 27.7280.735

2.8 26.8201.551

3.5 25.4063.288

4.2 22.1895.307

Figure 4-9. Average Cr6+ mass in GMAW fumes using globular metal transfer mode.

56


57


58


59


Table 4-3. Baseline particle size characteristics at different shield gas flow rates.

Total shield gas flow rate (Lpm) Total concentration (#/cm3)

Count geometric mean diameter (nm)

20 8.14E+05 40.7

25 7.10E+05 52.1

30 5.13E+05 60.9

60

Figure 4-14. Particle size trends at different shield gas and TMS carrier gas flow rates.

61

CHAPTER 5 CONCLUSIONS

The GTAW study conducted as a part of this research showed that vapor phase

sorbents can reduce Cr6+ emissions from this type of welding. Nitrogen oxide

concentrations were also reduced during GTAW. Some particles were coated in a film

of amorphous SiO2. This part of the study demonstrated the feasibility of using silica

precursor vapor to produce in-situ silica sorbents for control of Cr6+ in welding fumes.

Mechanical properties of the weld after TMS addition is an important piece of

information in evaluating the effectiveness of any welding fume control technology that

must be addressed prior to any application of vapor phase sorbents to GTAW shield

gas.

Silica precursors led to reduction in Cr6+ in GMAW fumes of over 90%. The use of

TMS as a shield gas additive brought Cr6+ concentration down from a level that

exceeded the OSHA permissible exposure limit by over seven times to within regulatory

limits. This result may help industry to keep pace with tightening OSHA limits for this

hazardous air pollutant. The increase in particle size observed when TMS was added

provides another benefit for using these chemicals. It may become more feasible to use

less obtrusive filtration systems or personal protective equipment such as dust masks

instead of full face respirators to protect workers from welding fume particles after their

size has been increased using silica precursors. New technologies to scavenge

nanoparticles from effluent gas streams will be required as nanomaterial manufacturing

spreads. Silica precursors may provide a tool for dealing with this challenging problem.

TEM imagery demonstrated that some particles were coated in a film of amorphous

62

SiO2. All of the metals contained in the particles that were coated in SiO2 were insulated

regardless of speciation.

More research is needed to determine how effectively silica precursors mask the

toxic effects of nickel and manganese through encapsulation in amorphous silica.

Research using less expensive silica precursors may lead to a more cost effective

process that is more appealing to industry. Further research is necessary to quantify the

effect of silica precursors as shield gas additives on the mechanical properties of the

materials being welded.

63

APPENDIX A DETERMINATION OF THEORETICAL MINIMUM PRECURSOR CONCENTRATION IN

FUME PARTICLE FORMATION AREA

In order to estimate how much TMS must be added to coat the aerosols produced

during GMAW, it was first necessary to estimate the volume and radius of a single SiO2

molecule using Equation A-1. The molecular mass of SiO2 divided by the density of

SiO2 was multiplied by Avagadros number to calculate v1. The diameter of an SiO2

molecule was calculated using v1, assuming molecules were spherical. It was then

necessary to calculate the total volume of aerosols produced during GMAW. A typical

particle size distribution was obtained from the literature (Zimmer and Biswas 2001).

The next step was to calculate the change in volume that occurred upon adding a

mono-atomic layer of SiO2 by adding two diameters of SiO2 molecules to the aerosol

diameter, assuming spherical particles. Finally, the concentration required to produce

the above change in volume was calculated using Equation A-2.

(A-1)

(A-2)

C = Change in concentration required to produce change in volume N = Number concentration (#/cm3) Nav = 6.022E23 dp = dp0 + 2dSiO2 v1 = 4.05E-20 cm

3 dSiO2 = 4.43E-8 cm

The concentration calculated using this technique relies on the assumption that

coating of uniform thickness occurs on all particles although in reality the coating is

64

almost certainly not homogeneous. The values used for the calculation of the minimum

precursor concentration are presented in Table A-1. The total concentration derived in

Table A-1 was 103-104 times lower than the concentration of TMS saturated carrier gas

used during experiments. The rationale behind the actual operating TMS carrier gas

feed rates can be found in Appendix C.

Table A-1. Particle size data used to calculate minimum precursor concentration, based on particle size distribution from literature (Zimmer and Biswas 2001).

N (#/cm3) dpo (cm) [(N)/(NAVv1)] [(dp)3-(dpo)

3] C (mol/cm3) 1.00E+05 1.80E-06 1.29E+01 9.03E-19 1.16E-17 4.48E+05 1.90E-06 5.77E+01 1.00E-18 5.80E-17 6.44E+06 2.00E-06 8.29E+02 1.11E-18 9.20E-16 1.72E+07 2.20E-06 2.22E+03 1.34E-18 2.96E-15 6.05E+07 2.40E-06 7.79E+03 1.59E-18 1.24E-14 2.55E+08 2.60E-06 3.28E+04 1.86E-18 6.09E-14 1.25E+09 2.80E-06 1.61E+05 2.15E-18 3.46E-13 8.59E+09 3.20E-06 1.11E+06 2.80E-18 3.09E-12 2.31E+06 3.40E-06 2.98E+02 3.15E-18 9.38E-16 1.39E+07 3.60E-06 1.79E+03 3.53E-18 6.30E-15 1.68E+07 3.75E-06 2.16E+03 3.82E-18 8.26E-15 2.41E+07 3.95E-06 3.11E+03 4.24E-18 1.32E-14 4.70E+07 4.30E-06 6.06E+03 5.01E-18 3.04E-14 6.40E+07 4.60E-06 8.24E+03 5.73E-18 4.72E-14 9.88E+07 5.00E-06 1.27E+04 6.76E-18 8.59E-14 1.68E+08 5.40E-06 2.17E+04 7.87E-18 1.71E-13 3.09E+08 5.70E-06 3.98E+04 8.76E-18 3.49E-13 5.36E+08 6.00E-06 6.90E+04 9.70E-18 6.69E-13 7.29E+08 6.45E-06 9.39E+04 1.12E-17 1.05E-12 8.88E+08 6.80E-06 1.14E+05 1.24E-17 1.42E-12 1.54E+09 7.40E-06 1.99E+05 1.47E-17 2.93E-12 1.84E+09 7.90E-06 2.37E+05 1.68E-17 3.97E-12 4.10E+09 8.30E-06 5.27E+05 1.85E-17 9.75E-12 9.47E+09 9.00E-06 1.22E+06 2.17E-17 2.65E-11 1.08E+10 9.50E-06 1.39E+06 2.42E-17 3.36E-11 1.98E+10 1.00E-05 2.55E+06 2.68E-17 6.82E-11 2.22E+10 1.10E-05 2.85E+06 3.24E-17 9.24E-11 3.81E+10 1.20E-05 4.90E+06 3.85E-17 1.89E-10 4.67E+10 1.30E-05 6.01E+06 4.52E-17 2.71E-10 9.89E+10 1.40E-05 1.27E+07 5.24E-17 6.67E-10 1.51E+11 1.55E-05 1.95E+07 6.42E-17 1.25E-09 2.62E+11 1.60E-05 3.38E+07 6.84E-17 2.31E-09

65

Table A-1. Continued.

N (#/cm3) dpo (cm) [(N)/(NAVv1)] [(dp)3-(dpo)

3] C (mol/cm3) 1.78E+11 1.70E-05 2.29E+07 7.71E-17 1.77E-09 4.67E+10 1.80E-05 6.01E+06 8.65E-17 5.19E-10 2.77E+10 2.00E-05 3.56E+06 1.07E-16 3.80E-10 1.56E+10 2.10E-05 2.01E+06 1.18E-16 2.37E-10 1.08E+10 2.30E-05 1.39E+06 1.41E-16 1.96E-10 8.30E+09 2.40E-05 1.07E+06 1.54E-16 1.64E-10 2.57E+09 3.00E-05 3.30E+05 2.40E-16 7.92E-11 2.57E+09 3.10E-05 3.30E+05 2.56E-16 8.45E-11 2.57E+09 3.20E-05 3.30E+05 2.73E-16 9.01E-11 1.84E+09 3.30E-05 2.37E+05 2.90E-16 6.86E-11 1.54E+09 3.40E-05 1.99E+05 3.08E-16 6.12E-11 1.29E+09 4.00E-05 1.66E+05 4.26E-16 7.08E-11 3.87E+08 4.20E-05 4.99E+04 4.69E-16 2.34E-11 1.91E+08 4.15E-05 2.46E+04 4.58E-16 1.13E-11 4.70E+07 4.40E-05 6.06E+03 5.15E-16 3.12E-12 4.83E+06 4.60E-05 6.21E+02 5.63E-16 3.50E-13

Total C

8.68627E-09 mol/cm3

66

APPENDIX B DETERMINATION OF TMS AND TEOS VAPOR PRESSURE

The vapor pressures for TMS and TEOS were calculated using the Antoine

equation (Equation B-1). The Antoine coefficients used in the calculations are presented

in Table B-1 (Aston et al. 1941; Stull 1947).

Table B-1. Antoine coefficients for TMS and TEOS.

A B C

TEOS 4.17312 1561.277 -67.572 TMS 3.97703 1047.242 -36.057

(B-1)

P = vapor pressure (bar) T = temperature (K), 273 K for TMS, 373 K for TEOS

67

APPENDIX C DETERMINATION OF PRECURSOR CONCENTRATION IN SATURATED CARRIER

GAS

The following calculations show how the actual TMS feed rates used during this

research were calculated. Assuming the carrier gas leaving the liquid reservoir was

saturated with TMS vapor, the ideal gas law (Equation C-1) was used to calculate the

concentration of TMS vapor in the saturated carrier gas.

(C-1)

(vapor pressure of TMS vapor at 273 K)

(gas constant)

The TMS molar feed rates were calculated by multiplying the above concentration by

the carrier gas flow rates passed over the liquid reservoir, found in Table 2-1. For

example, when 4.2% of the 30 Lpm shield gas flow rate was used as TMS carrier gas,

the calculation would be performed as in Equation C-2.

TMS Molar Feed Rate =30 L shield gas

min

0.042 L carrier gas

1 L shield gas

103 cm3 carrier gas

1 L carrier gas

1.58910-5mol TMS

cm3 carrier gas

=2.0010-2 mol TMS

min

(C-2)

The same calculation was performed for TEOS using appropriate values for partial

pressure and temperature. The calculation was done for a temperature of 373 degrees

Kelvin (K) where vapor pressure was 0.114 atmospheres (atm). The corresponding

68

concentration of TEOS in saturated carrier gas was 3.713 10-6 mol TEOS/cm3. The

actual experimental concentration of TMS in the area of fume formation should be

higher than the minimum theoretical concentration calculated in Appendix A because of

dilution upon rejoining the main shield gas flow. It further decreased upon leaving the

welder into the ambient air where fume particle formation occurred. These factors were

taken into account when choosing the carrier gas flow rates to be tested.

In previous research that used reducing agents to minimize oxidation of chromium,

the use of 5.64 10-4 gmol/min ethene or 5.64 10-4 gmol/min nitrogen oxide by

Dennis et al. (2002) obtained 42% and 45% reduction in Cr6+ formation, respectively.

Based on this result, the current study used higher TMS feed rates to meet the desired

goal of 90% reduction in Cr6+ formation. The TMS carrier gas flow rates chosen were

also influenced by preliminary research performed by Dr. Daniel P. Chang in which

1.6% of the total GMAW shield gas flow rate was used as TMS carrier gas. Cr6+ was

reduced by 64% in Dr. Changs work. The current study used more of the shield gas as

TMS carrier gas than Dr. Chang in an attempt to exceed the 64% reduction attained by

Dr. Chang. In experimental set A, the specific TMS carrier gas percentages varied from

1.4 4.2% for a 30 Lpm shield gas low rate, presented in Table 2-1. Experimental set C

tested 5.0% of the 25 Lpm shield gas flow rate to see if a lower shield gas flow rate

impacted successful delivery of TMS additive to the welding arc. The results showed

that there was some reduction of Cr6+ mass but not enough to reach 90%. The shield

gas flow rate was further decreased to 20 Lpm and the TMS was varied between

2.1-6.3% in experimental set D.

69

APPENDIX D COST ANALYSIS

GTAW Operating Cost of TEOS Addition

The additional operating cost incurred based on the cost of laboratory grade TEOS

is calculated below. The cost of 98% pure TEOS obtained from Acros Organics used in

the calculation is $41.09 U.S. dollars (USD) per liter or 0.437 USD per gram TEOS. The

TEOS carrier gas flow rate used in this study was 1.874 Lpm. The concentration of

TEOS in saturated carrier gas was 3.7137 10-3 mol TMS/L carrier gas and the

molecular weight of TEOS is 208.33 grams per mole (g/mol). The added operating cost

of using TEOS as a GTAW shield gas additive is $3.80 USD per hour of welding time,

as calculated using Equation D-1.

Additional Operating Cost =1.874 L TEOS carrier gas

min

3.713710-3 mol TEOS

1 L carrier gas

208.33 g TEOS

1 mol TEOS

0.0437 USD

1

Control of Hexavalent Chromium Emission

Documents

Transcript of Control of Hexavalent Chromium Emission