Characterisation of airborne ultrafine particles generated ... · welding processes such as metal-active gas (MAG) (4, 15), aims to assess the emission of ultrafine particles from

Characterisation of airborne ultrafine particles generatedfrom welding of steel plates

The present study focuses on the characterization of ultrafine particlesemitted in steel welding using Ar+CO2 mixtures, and determines the mainemission-influencing process parameters. A clear dependence was foundbetween the amount of ultrafine particles emitted (measured by particlenumber and alveolar deposited surface area) and, firstly, the distance tothe welding front and, secondly, the main welding parameters, namely thecurrent intensity and heat input in the welding process. The emission ofairborne ultrafine particles, like fume formation, seems to increase in linewith current intensity. A comparison of the tested gas mixtures showshigher emissions for more oxidant mixtures, i.e., mixtures with higher CO2content, which result in higher arc stability. The latter mixtures producehigher concentrations of ultrafine particles (as measured by number ofparticles per cm3 of air) and higher values of alveolar deposited surfacearea of particles, making them more hazardous to exposed workers.

By J.F. GOMES. Instituto de Biotecnologia e Bioengenharia / Instituto Superior Técnico – Universidade Técnica de Lisboa, Av. Rovisco Pais,1049-001 Lisboa, Portugal 2ISEL – Instituto Superior de Engenharia de Lisboa, Área Departamental de Engenharia Química, R. ConselheiroEmídio Navarro, 1959-007 Lisbon, Portugal. R. M. MIRANDA. Departamento de Engenharia Mecânica e Industrial, Faculdade de Ciências eTecnologia, FCT, Universidade Nova de Lisbon, 2829-516 Caparica, Portugal.

Arc welding is extensively used in metallic construction worldwide. However, it can produce dangerous fumes that may behazardous to the welder’s health (1) and it is estimated that, presently, 1-2% of workers from different professionalbackgrounds (accounting for more than 3 million people) are subjected to welding fumes and gas action (2). These authorsalso showed a correlation between processing parameters in metal active gas (MAG), that is, metal transfer modes, and thequantity of fumes formed, expressed as fume formation rate. Additionally, the gas mixtures have also an influence on fumes(quantity and composition), since the higher the oxygen content of the gas, the higher the fume formation observed. Withthe advent of new types of welding procedures and consumables, the number of welders exposed to welding fumes isgrowing constantly, in spite of the mechanization and automation of the processes (3). Simultaneously, the number ofpublications on epidemiologic studies (4) and welders’ protection devices are also increasing. Apart from that, theinfluence of ultrafine particulates, lying in the nano range, on human health has been pinpointed as an area of greatconcern (5-7) as airborne ultrafine particles can also result from common macroscopic industrial processes such as welding(8-9). The detrimental health effects of inhaling ultrafine aerosols were recognised long ago and various attempts havebeen made to minimise exposure, such as specific emission regulations and air-quality objectives in workingmicroenvironments.

Year 34 nº135 Third Quarter 2014

file:///Users/raulcastro/Documents/Mapfre/2013/Revistas/SMA/n135-en/docs/Seguridad-y-Medio-Ambiente-135-en.pdf

When considering human exposure to airborne pollutants, the exposure to airborne particles, and specifically to its finerfractions, such as sub micrometer particles, is of particular concern. Current workplace exposure limits, which have beenestablished long ago, are based on particle mass, but this criteria is not really applicable to nano-sized particles since thesematerials are, in fact, characterised by very large surface areas (at the same volume, nano sized particles have largersurface areas than micro sized particles, for instance); this has been flagged as the distinctive characteristic that couldeven turn an inert substance into a toxic one, having the same chemical composition but exhibiting very differentinteractions with biological fluids and cells (10-11). As a result, assessing workplace conditions and personal exposure basedon the measurement of particle surface area is of increasing interest. It is well known that lung deposition is the mostefficient way for airborne particles to enter the body and potentially cause adverse health effects. If ultrafine particles candeposit in the lung and remain there, and have an active surface chemistry to work with, then, there is exposure anddosing potential. Oberdörster (12) showed that surface area plays an important role in the toxicity of nanoparticles and isthe measurement metric that best correlates with particle-induced adverse health effects. Therefore, in order to be ableto assess exposure, it is important to work from an estimate of the surface area of emitted ultrafine particles, as they arepotentially able to deposit in the lower parts of the lung, such as the alveoli and clog them; they might even be transferredto the blood circulation system with resulting distribution in several end organs (13).

In 1996, the International Commission of Radiological Protection (ICRP) developed a comprehensive lung deposition modelfor radioactive aerosols. Several parameters are required to construct the model including breathing rate, lung volume,activity, nose/mouth breathing, etc.; the model-derived deposition curves (for tracheobronchial and alveolar deposition)might vary according to these parameters. For industrial hygiene applications, ACGIH (14) developed a definition of areference worker, in order to derive the corresponding deposition curves for tracheobronchial and alveolar lung deposition,based on the ICRP model: the tracheobronchial deposition curve represents the fraction of aerosol that deposits in thetracheobronchial region of the lung, while the alveolar deposition curve represents the fraction of the aerosol that depositsin the alveolar region of the lung. For exposure assessment applications it is common to sample aerosols in terms of theirdeposition in a specific region of the human lung, this depending on the aerosol being sampled. As for ultrafine particles,due to their very fine dimensions, the health effects would be related to deposition deep in the alveolar regions of thelung, so the respirable fraction of the aerosol is the metric of interest, in terms of assessing the potential surface of thealveoli that might be clogged by these ultrafine particles.

This study, following on from preliminary work from these authors that demonstrated the presence of ultrafine particles inwelding processes such as metal-active gas (MAG) (4, 15), aims to assess the emission of ultrafine particles from steelwelding and to correlate these with operational parameters and ipso facto with molten metal transfer modes. Similarstudies have also been carried out for welding fumes generated in other processes, such as manual metal arc welding(MMA), metal inert gas welding (MIG) and tungsten inert gas welding (TIG) for stainless steel electrodes (16). None of theprevious studies, however, has assessed the deposited surface area of emitted particles.

Materials and methods

MAG welding tests were performed in a laboratory using an experimental set-up consisting of an automatic welding machineKemppi, model ProMig 501, controlled to ensure a stable electric arc, a constant welding speed and, therefore,repeatability of the welding process.

Bead-on-plate welds were performed on 3-mm gauge mild carbon steel plates S235 JR with the chemical composition shownin table 1. Tests were also performed on austenitic 10-mm gauge stainless steel plates AISI 304 with the chemicalcomposition also shown in table 1. Welding consumables used were a solid wire AWS 5.18 ER79S-6, with a diameter of 1 mmfor carbon steel and a solid wire AWS ER316 LSi, with a diameter of 0.8 mm, for stainless steel. Both solid wirecompositions are shown in table 2. The operating parameters used are depicted in table 3 for carbon steel and in table 4for stainless steel. Different conditions were tested for welding, varying the gas protection mixture and current intensityand voltage in order to produce short-circuits, globular and spray metal transfer modes. The welding voltage was adjustedautomatically by the welding machine in accordance with the wire feeding speed. Each test condition was replicated twice.Between each test, time intervals were provided to allow for some dissipation of aerosol in the closed atmosphere of thesampling location.

Gas protection was selected from the most frequently used industrial mixtures, generating high quantities of welding fumes(17) and hence high quantities of carbon dioxide and oxygen, to improve penetration and bead width. Molten metal transfer

modes were also chosen from the commonest welding procedures as those that produce high quantities of welding fumes(17). All welding tests were performed in the flat position. The gas mixtures were supplied by Air Liquide.

Table 1.Chemical composition of the base materials used (% w/w)

Type of steel C Mn P S Si Ni Cr N

S235 JR 0.017 1.40 0.035 0.035 - none none -

AISI 304 0.08 2.0 0.045 0.030 0.75 8.0-10.5 18-20.0 0.10

Table 2.Chemical composition of the solid wire used (% w/w)

Type of wire C Mn P S Si Ni Cr Cu Mo

AWS 5.18 ER70S-6 0.06-0.15 - 0.35 0.8-1.15 0.15 0.15 0.50 0.15

AWS A5.9 ER316 LSi 0.03 1.0-2.5 0.03 0.03 0.65-1.00 11-14 18-20 0.75 2.0-3.0

Table 3. Experimental test conditions for carbon steel welding

Gas mixture ARCAL 21 (90% Ar + 10% CO2)

Test 1 2 3

Wire feed velocity (m/min) 4.0 6.3 11.2

Transfer mode Short circuit Globular Spray

Gas mixture ATAL (82% Ar + 18% CO2)

Test 1 2 3

Wire feed velocity (m/min) 4.0 6.3 No logrado

Modo de transferencia Short circuit Globular Spray

Gas mixture 100% CO2

Test 1 2 3

Wire feed velocity (m/min) 5.0 7.5 No logrado


Table 4.Experimental test conditions for stainless steel welding

Gas mixture ARCAL 12 (95% Ar + 5% CO2)

Test 1 2 3



Gas mixture ARCAL 121 (81% Ar + 18% He + 1% CO2)

Test 1 2 3


Modo de transferencia Short circuit Globular Spray

Gas mixture ARCAL 129 (91% Ar + 5% He + 2% CO2 + 2% N2)

Test 1 2 3

Velocidad de alimentación del alambre (m/min) 5.0 7.5 No logrado


In order to assess the emissions of ultrafine particles on the nearby welding environment, sampling (using NSAM) was doneas depicted in figure 1. During sampling, the exhaust system was turned off, and doors and windows to the workshop werekept closed. For measuring ultrafine particle exposure a Nanoparticle Surface Area Monitor (NSAM), TSI, Model 3550, wasused. This equipment estimates the human alveolar-deposited surface area (ADSA) of particles expressed as squaremicrometre per cubic centimetre of air (µm2/cm3), corresponding to the alveolar region of the lung. The operation of thisequipment is based on diffusion charging of sampled particles (regardless of its size, shape and agglomeration, as discussedby Gomes (4)), followed by detection of the charged aerosol using an electrometer, as described elsewhere (15). Itestimates the surface area of ultrafine particles capable of being deposited in the alveolar region of the lung, using thepreviously mentioned ACGIH deposition model (14). Due to the inexistence of a specific occupational exposure limit valuefor ultrafine particles, a baseline value, for comparison purposes, was obtained for each measurement task.

Figure 1. Location of the sampling point

Particles were also collected using a Nanometer Aerosol Sampler (NAS), TSI, Model 3089, on 3mm diameter copper gridspolymer coated for further observation by transmission electron microscopy (TEM), Hitachi, model H-8100 II, equipped withan energy dispersive X-ray spectroscopy (EDS) probe.

Results and discussion

The measured data for each welding tests are shown in the next figures as follows: figures 2 to 4 referring to carbon steel,and figures 5 to 7 to stainless steel. In each one, the top figure represents the measured ADSA, expressed as squaremicrometres per cubic centimetre of air (µm2/cm3), corresponding to the alveolar region of the lung versus welding time,for each welding transference mode (short-circuit, globular and spray). The bottom figure represents ADSA versus theapplied welding current intensity. It should be noted that these graphs show, not the instantaneous ADSA, but theintegrated curves for the whole welding period, which means that the graphs show the accumulated ADSA for the wholewelding test. The welding tests lasted from 30 to 84 seconds.

Figure 2. Results for welding tests with ARCAL 210

Figure 3. Results for welding tests with ATAL

Figure 4. Results for welding tests with 100% CO2

It will readily be seen that both the applied transfer modes and also the gas mixtures used have a marked influence onnanoparticle emission. Table 5 shows the average ADSA values obtained for the welding tests with carbon steel. Using gasmixture ARCAL 21, spray transfer mode is the one that results in the highest ADSA values, and the same occurs for the othergas mixtures: with increasing values of welding parameters, ADSA values are also higher. A priori the highest ADSA valuesmight be expected for 100% CO2: however, this behaviour was not obtained for the latter gas mixture but rather ATAL, withonly 18% CO2. Using ATAL the highest ADSA values were obtained in the short-circuit transfer mode, instead of spray mode,which corresponds to higher current intensities.




Table 6 shows the average ADSA values obtained for stainless steel welding: again, ADSA values increase with higher currentintensity values, except for ARCAL 121, which was to be expected, considering that the spray mode results in higher valuesthan those with globular transfer mode. This could be due to the welding parameters used: possibly the current intensity(199 A) used represents a zone where the fume emission rate is somewhat low, as shown in figure 8, resulting in a lowervalue when compared to the other transfer modes. In fact, ARCAL 121 is the gas mixture resulting in the highestnanoparticle emissions, as measured by ADSA. Another possible cause for this is related to the high He content in the gasmixture (18%): He exhibits a very high ionization energy value of 24.58 eV, so this will result in an electric arc having highertemperatures capable of producing a higher volatilization of elements both from the base material and the consumable.

Table 5. Average ADSA values obtained during welding tests for carbon steel

Gas mixture and transfer mode used Average ADSA (µm2/cm3.s)

ARCAL 21 ATAL 100% CO2

Short circuit 8.325 22.266 12.899

Globular 13.306 42.896 18.292

Spray 17.574 - -

A comparison of tables 5 and 6 also shows that higher ADSA values are obtained for welding stainless steel than carbonsteel. Figures 9 to 11 show the collected nanoparticles using the NAS equipment: figure 9 refers to welding carbon steelwith ARCAL 21 and figures 10 and 11 refer to welding stainless steel with ARCAL 121 and ARCAL 129, respectively.Nanoparticles shown in figures 9 are spherical, amorphous, aggregated together and with dimensions ranging from 10 to 90nm. Also regarding stainless steel welding, nanoparticles are spherical and amorphous, agglomerated, with dimensionsranging from 10 to 40 nm.

Table 6. Average ADSA values obtained during welding tests for stainless steel

Gas mixture and transfer mode used Average ADSA (µm2/cm3.s)

ARCAL 12 ARCAL 121 ARCAL 129

Short circuit 23.637 75.390 33.644

Globular 37.054 94.136 78.361

Spray 39.376 65.829 80.861

Figure 8. Relation between electric arc strength and current for different welding processes(17)

The particles were subjected to chemical analysis, by EDS, and the obtained spectra are shown in figures 12 and 13, forcarbon steel (ARCAL 21) and for stainless steel (ARCAL 129), respectively. For carbon steel, the elements detected wereiron, silicon and manganese (copper is also shown in the spectra but it is due to the grids used for collecting nanoparticles).For stainless steel, the detected elements were iron, chromium and nickel; this confirms the origin both from the basematerial and also the consumable.

Figure 9. TEM image for nanoparticles collected for welding with ARCAL 21

Figure 12. Chemical composition of nanoparticles collected for welding with ARCAL 21

Figure 13. Chemical composition of nanoparticles collected for welding with ARCAL 129

Conclusions

From this work the following conclusions can be derived:

nanoparticle emission, as measured by ADSA, increases directly with welding parameters such as current intensity andtension;

as regards the studied transfer modes, the spray mode is the one resulting in the highest nanoparticle emission;

the short-circuit mode resulted in lower average ADSA values for all gas mixtures tested. It should be noted that thistransfer mode exhibits the lowest current intensity and tension, resulting in lower temperatures of the electric arcthus volatilizing lower quantities of elements from the base materials and consumables;

globular transfer mode, for the majority of tested conditions, results in ADSA values between short-circuit and spraymodes. This transfer mode is known to create instability in the electric arc; it would therefore seem that electric arcinstability is not the main cause of nanoparticle emission but rather the electric arc temperature;

as for carbon steel welding, the gas mixture composed of 82% Ar+18% CO2 resulted in high ADSA values, which wassomewhat unexpected, as previous fume-emission studies showed that fume formation increased directly with CO2 inthe gas mixture;

as regards stainless steel, the gas mixture composed of 81% Ar+18% He+1% CO2 resulted in high ADSA values, whichseems to result from the high quantity of He in the mixture. As He has a high ionization energy, the resulting electricarc has higher temperatures which produces a higher volatilization of elements both from base material andconsumable;

a comparison of the base materials used shows that higher ADSA values were obtained for stainless steel than forcarbon steel. Also, the emitted nanoparticles for stainless steel showed the presence of nickel and chromium whichare potentially carcinogenic elements;

the collected particles, both from carbon steel and from stainless steel have nanometric dimensions, the majoritybetween 10 and 100 nm.

Acknowledgements

The authors acknowledge the support of Fundación MAPFRE, Madrid, Spain through an I. Hernando de Larramendi researchgrant. The authors would also like to thank Ms. Catarina Pereira and Mr. António Campos, Miguel Bento and Tiago Pereira,who assisted in performing the welding tests. Their gratitude also goes to Dr. Telmo Santos, from FCT-UNL, who helped withthe data analysis and Dr. Patrícia Carvalho, from Microlab-IST, who performed the electronic microscopy analysis.

References

1. Gomes, J. Higiene e segurança da soldadura, 1983, ISQ: Oeiras.

2. Pires, I., Quintino, L., Miranda, R., Gomes, J. Fume emissions during gas metal arc welding, Toxicology &Environmental Chemistry, 2006, 88, 385-394.

3. Ascenço, C., Gomes, J., Cosme, N., Miranda, R. Analysis of welding fumes, Toxicology & Environmental Chemistry,2005, 87, 345-349.

4. Gomes, J., Albuquerque, P., Miranda, R., Vieira, M. Determination of airborne nanoparticles from welding operations,Journal of Toxicology and Environmental Health, Part A, 2012, 75, 747-755.

5. Jenkins, N., Eager, T. Chemical analysis of welding fume particles, Welding Journal, 2005, Supp., 87-93.

6. Dasch, J., D’Arcy, J. Physical and chemical characterization of airborne particles from welding operations inautomotive plants, Journal of Occupational and Environmental Hygiene, 2008, 5, 444-454.

7. Buonanno, G., Morawska, L., Stabile, L. Exposure to welding particle in automotive plants, Journal of Aerosol Science,2011, 42, 295- 304.

8. Moroni, B., Viti, C. Grain size, chemistry and structure of fine and ultrafine particles in stainless steel welding fumes,Aerosol Science, 2011, 40, 938-949.

9. Elihn, K., Berg, P., Liden, G. Correlation between airborne particle concentrations in seven industrial plants and

estimated respiratory tract deposition by number, mass and elemental composition, Journal of Aerosol Science, 2011,42, 127-141.

10. Oberdörster, G., Gelein, R., Ferin, J., Weiss, B. Association of particulate air pollution and acute mortality:involvement of ultrafine particles, Inhalation Toxicology, 1995, 7, 111-124.

11. Rickerby, D., Morrison, M. Nanotechnology and the environment: A European perspective, Science and Technology ofAdvanced Materials, 2007, 8, 19-24.

12. Oberdörster, G. Pulmonary effects of inhaled ultrafine particles, International Archives of Occupational andEnvironmental Health, 2001, 74, 1-8.

13. Kreyling, W., Semmler, M., Erbe, F., Mayer, P., Takenaka, S., Schulz, H., Oberdörster, G., Ziesenis, A. Translocation ofultrafine insoluble iridium particles from lung epithelium to extrapulmonary organs is size dependent but very low,Journal of Toxicology and Environmental Health – A, 2002, 65, 511-535.

14. Phalen, R. Particle size-selective sampling for particulate air contaminants, Ed. J.H. Vincent, 1999, ACGIH: Cincinnati.

15. Gomes, J., Albuquerque, P., Miranda, R., Santos, T., Vieira, M. Comparison of deposited surface area of airborneultrafine particles generated from two welding processes, Inhalation Toxicology, 2012, 24, 774-781.

16. Berlinger, B., Benker, N., Weinbruch, S., L’Vov, B., Ebert, M., Koch, W., Ellingsen, D., Thomassen, Y. Physicochemicalcharacterization of different welding aerosols, Analytical and Bioanalytical Chemistry, 2009, 399, 1773-1780.

17. Pires, I., Quintino, L., Miranda, R. Analysis of the influence of shielding gas mixtures on the gas metal arc weldingmetal transfer modes and fume formation rate, Materials and Design, 2007, 28, 1623-1631.