Size distribution analysis of airborne wear particles released by subway brake system

Wear 372-373 (2017) 169–176

Contents lists available at ScienceDirect

Wear journal homepage: www.elsevier.com/locate/wear

Size distribution analysis of airborne wear particles released by subway brake system Hyeong-Gyu Namgung a, Jong Bum Kim b, Min-Soo Kim c, Minhae Kim a,d, Sechan Park a,d, Sang-Hee Woo b, Gwi-Nam Bae b, Duckshin Park a, Soon-Bark Kwon a,d,n a

Transportation Environmental Research Team, Korea Railroad Research Institute (KRRI), Uiwang-si 16105 Republic of Korea Center for Environment, Health and Welfare Research, Korea Institute of Science and Technology (KIST), Seoul 02792, Republic of Korea c Metropolitan Transportation Research Center, Korea Railroad Research Institute (KRRI), Uiwang-si 16105, Republic of Korea d Railway System Engineering, University of Science and Technology (UST), Uiwang-si 16105, Republic of Korea b

art ic l e i nf o

a b s t r a c t

Article history: Received 31 August 2016 Received in revised form 8 December 2016 Accepted 17 December 2016 Available online 21 December 2016

Contributions of exhaust and non-exhaust sources to traffic-related particulate matter (PM) pollution in the atmosphere are almost identical and the most important non-exhaust source is known to be brake wear particles. In order to understand the properties and harmful effects of wear particles on people, accurate information on size distribution of brake wear particles is needed. Our previous study investigated the measured changes in size distribution of nanoparticles of 500 nm or smaller to understand the origin of nanoparticles due to temperature increases on the friction surface. The present study was intended to investigate the characteristics of size distribution (5.6 nm–32 μm) of PM released under different braking conditions by using different instruments. The measurement results under 9 braking conditions using 3 different instruments showed that the size distribution characteristics of particles can be divided into two main types according to braking energy. The first type is of PM up to 10 μm in size and with a peak number concentration at 0.2–0.75 μm regardless of braking energy, while the second type is of PM around 10 nm in size generated only when braking energy increased and particles that increased up to 100 nm. In addition, we found that the size distributions measured by the optical particle counter (OPC) and the aerodynamic particle sizer (APS) were consistent by assuming a mean diameter ratio of two instruments. & 2016 Elsevier B.V. All rights reserved.

Keywords: Brake wear particles Nanoparticles Particle sizing instrument Particle size distribution Braking energy

1. Introduction The disk - brake pad brake system applied in vehicles and trains converts kinetic energy of those vehicles into thermal energy and inevitably releases wear particles from friction during the braking process. The characteristics of such generation of wear particles differ according to a variety of factors such as brake type, brake and pad material and elements, initial brake velocity, brakingcontact force, vehicle operating conditions and testing method [1– 6]. Brake disc and pad materials are the same as those used with automotive vehicles and subway brakes. Although some released wear particles may be deposited on the surface of the brakes and the road, 35–50% of them remain suspended in the air as airborne particles [1,3]. Although the technology to reduce vehicle emission pollutants from combustion engines continues to improve as a n Corresponding author at: Transportation Environmental Research Team, Korea Railroad Research Institute (KRRI), Uiwang-si 16105, Republic of Korea. E-mail address: [email protected] (S.-B. Kwon).

http://dx.doi.org/10.1016/j.wear.2016.12.026 0043-1648/& 2016 Elsevier B.V. All rights reserved.

way of coping with constantly-tightening regulations on vehicle exhaust gas and particles, the development, or even study, of technology to reduce non-exhaust sources of particles such as brake wear, tire wear and re-suspension of existing road dust is relatively limited, while the contribution of non-exhaust sources to pollution of the urban atmosphere is expected to gradually increase [7]. In a review paper on brake wear particles, Grigoratos and Martini [5] summarized that the contributions of exhaust and non-exhaust sources to traffic-related PM10 were almost identical and that the most important non-exhaust source was brake wear. There have been many studies on size distribution and elements of brake wear particles [4,8–11]. There are also reports on the evaluation of the harm brake wear particles pose to the human body [12–15], their contribution to atmospheric pollution [16–19], and calculation of emission factors [1,20–22]. Findings and limitations from these studies, as they relate to distribution of number concentration of airborne wear particles released by brake wear, are summarized below. Sanders et al. [3] observed a stiff increase in number

170

H.-G. Namgung et al. / Wear 372-373 (2017) 169–176

concentration of 0.3 μm or smaller sized particles under severe braking conditions by using cascade impactors. Mosleh et al. [8] found the common peak in the range of 0.3–0.4 μm and the second peak in the range of 1–10 μm according to braking conditions in a pin-on-disc test. Iijima et al. [20] reported their observation of unimodal size distribution showing a constant peak at 0.8 μm under different braking conditions using an aerodynamic particle sizer (APS, TSI Inc.). Wahlstrӧm et al. [11] observed the bimodal peak of 0.28 μm and 0.35 μm by using the aerosol spectrometer (Grimm 1.109). Kukutschova et al. [9] measured the brake wear particles using the scanning mobility particles sizer (SMPS; TSI Inc.) and ASP but they did not describe the 0.45–0.5 μm particles that were not included in the measurement range of either SMPS or APS. Mathissen et al. [23] observed the generation of nucleation mode particles at around 10–60 nm during conditions of full braking using an engine exhaust particle sizer (EEPS; TSI Inc.). Abbasi et al. [24] also observed peak in the ultrafine region of around 100 nm using the aerosol spectrometer (Grimm 1.109) and SMPS. A review of these studies on size distribution of wear particles released by braking shows that there were limitations in size analysis method according to the measurement limitations (size measurement limitations) of the measuring equipment or the principle of measurement. There are few comprehensive studies that analyze size distribution of all sizes of wear particles that can range from several nanometers to tens of micrometers. Our previous study investigated the measured change in size distribution of nanosized particles of 500 nm or smaller every second under various braking conditions to investigate the evaporation-condensation-coagulation phenomenon of brake pad elements due to temperature increase of the friction surface [25]. This study was intended to analyze the results of size measurements from particle measuring devices applying different measuring principles (lightscattering method and time-of-flight method) that are different from the test in our preceding study, as well as to investigate the characteristics of size distribution in the range of 5.6 nm–32 μm of particulate matter released under different braking conditions. It will also present a method of interpreting the measurement results of the aerosol spectrometer and APS, which have similar measurement ranges.

2. Method 2.1. Braking test To measure brake wear dust, a brake disk and a non-asbestos organic (NAO) brake pad used in subway trains were mounted, and the particles released under various braking conditions analyzed. NAO brake pads contain nonferrous metals, inorganic and organic fibers, abrasives, lubricants and property modifiers such as glass, rubber, synthetic fiber and carbon but the exact composition rate of the pads is unknown. Braking tests were conducted using the real-scale brake dynamometer which met UIC code 541-3, international standards [26]. The brake dynamometer has 400 km/h of the maximum applicable speed and 60kN of the maximum brake force. Details of the test were presented in our previous study [25]. The test system was configured to measure the airborne wear particles released under various braking conditions using three particle sizing instruments through the flow splitter (TSI, 3708): the fast mobility particle sizer (FMPS; TSI 3091), aerodynamic particle sizer (APS; TSI 3321), and the optical particle counter (OPC; Grimm 1.109) simultaneously (Fig. 1). Braking conditions were set to 9 cases at three braking-contact forces (12 kN, 20 kN, 30 kN) at each of three initial braking velocities of 25 km/h, 50 km/ h, and 100 km/h.

Fig. 1. Experimental setup of full-scale brake dynamometer for airborne wear particle sizing.

2.2. Measurement of size distribution Table 1 shows the specifications of particle sizing instruments used to measure the airborne particles (aerosols) released during braking. FMPS controls the charging state of each particle in the size range of 5.6–560 nm to measure the size according to the difference of electrical mobility and provides size distribution data every second through 32 channels. The APS measures the time-offlight of particles passing through an accelerated nozzle to calculate aerodynamic diameter of unit-density spherical particles in the 0.5–20 μm range, and also provides size distribution data through 32 channels every second. In the case of APS, the data in the first channel (bin 1), which measures particles up to 0.523 μm, were excluded from analysis [27]. OPC measures the light scattering intensity of single particles to calculate the size and provides size distribution through 31 channels every 6 s. Unlike FMPS, which measures nanoparticles of 560 nm or smaller, APS and OPC have similar size measurement ranges. However, care needs to be exercised when interpreting physical size of actual particles released during braking and the results of size measurement from each instrument since FMPS measures size based on electrical mobility of particles, APS measures time-of-flight to calculate aerodynamic diameters of particles of 1 g/cm3 in density, while OPC measures size based on light-scattering intensity of particles. 2.3. Data analysis The size distributions of particles were measured using three instruments during the braking time (tB; time from beginning of braking to actual stopping), simultaneously. FMPS and APS provided size distribution data every second while OPC provided them every 6 seconds. Since the braking time under each braking condition differs according to initial velocity (Vi) and brakingcontact force (FB) as shown in Table 2, the mean concentration of each size throughout the entire braking time under each braking condition was analyzed. The concentration of background particles, which existed when braking is not applied, was measured before and after the testing and excluded from analysis. Braking energy (EB) is expressed by multiplication of the specific brakingcontact force at the initial braking velocity to the distance traveled until the braking is complete. Since the velocity decreases in a linear manner after braking is applied, braking distance can be expressed as 1/2 the multiplication of the initial velocity by the braking time as Eq. (1):

H.-G. Namgung et al. / Wear 372-373 (2017) 169–176

171

Table 1 Specifications of instruments for aerosol size distribution measurement. Particle sizing instrument

Manufacturer model no.

Method of measurement

Particle size range Min.

Fast mobility particle sizer (FMPS) Optical particle counter (OPC) Aerodynamic particle sizer (APS)

TSI 3091 Grimm 1.109

Electrical mobility of 5.6 nm particles Scattered light by particles 0.25 μm

TSI 3321

Time-of-flight of particles

0.5 μm

Flow rate (L/min)

Time resolution (s)

Max. particle conc.

560 nm

10

1

–

32 μm

1.2

6

2000 #/cm3

20 μm

5

1

10,000 #/cm3

Max.

Table 2 Total concentration and generation rate of wear particles measured by each instrument under each condition. Initial velocity, Vi (km/h)

Braking-contact force, FB (kN)

Braking time, tB Braking energy, EB (s) (kN-m)

Total particle concentration (#/cm3)

Generation rate of particles (#/cm3/s)

FMPS

OPC

APS

FMPS

OPC

APS

25

12 20 30

15 10 8

219 243 292

6.7E þ 4 5.1E þ4 1.2.Eþ 5

5.6E þ 3 9.2E þ 3 6.5E þ 3

1.3E þ 4 1.7E þ 4 2.0E þ4

4.5Eþ 3 5.1E þ3 1.5E þ4

3.7Eþ 2 9.2Eþ 2 8.2Eþ 2

8.9E þ2 1.7E þ3 2.5E þ3

50

12 20 30

30 19 14

875 924 1021

1.2.Eþ 6 8.8E þ 5 8.9E þ 5

4.0E þ 4 2.5E þ 4 1.7E þ4

6.9E þ4 4.3E þ4 3.7E þ4

4.0Eþ 4 4.6Eþ 4 6.4Eþ 4

1.3E þ3 1.3E þ3 1.2E þ3

2.3E þ3 2.2E þ3 2.6E þ3

100

12 20 30

50 35 24

2917 3403 3500

1.6E þ7 6.0E þ 7 9.0E þ 7

5.5E þ 4 5.0E þ 4 4.0E þ 4

1.1E þ5 8.4E þ4 5.8E þ4

3.2Eþ 5 1.7E þ6 3.8Eþ 6

1.1E þ 3 1.4E þ3 1.7E þ3

2.3E þ3 2.4E þ3 2.4E þ3

⎛ V ×t ⎞ EB=μ × FB× S = μ × FB× ⎜ i B ⎟ ⎝ 2 ⎠

(1)

where, EB is braking energy, μ is the friction coefficient, FB is braking-contact force, S is stopping distance, Vi is initial velocity, and tB is braking time.

3. Results 3.1. Particle concentration and generation rate Table 2 shows initial velocity, braking-contact force, braking time and braking energy and the total number concentration of particles measured by each instrument under each condition. The friction coefficient, μ, was found to be 0.35 throughout the entire experiment. The total number concentration of particles at the lowest braking energy condition (25 km/h, 12kN, 219kN-m) was 6.7E þ 4 cm  3 by FMPS in the size range of 5.6–560 nm, 1.3Eþ4 cm  3 by APS in the size range of 0.5–20 μm and 5.6E þ 3 cm  3 by OPC in the size range of 0.25–32 μm. Under the highest braking energy condition (100 km/h, 30 kN, 3,500 kN-m), the concentration increased by 1343 times, 4.3 times and 7.1 times, respectively, of the lowest braking energy condition for each instrument. In terms of the particle generation rate (particles per unit time of braking), 4500 particles were generated up to the 560 nm particle size range at the lowest braking energy condition, and the value increased by 844 times to 2.8 million particles under the highest braking energy condition. On the other hand, particles sized 0.25 μm or larger (measured by both OPC and APS) showed a linear increase in generation rate as braking energy increased. Fig. 2 shows the average of total number concentration and particle generation rate measured by each sizing instrument under the low, medium and high braking energy cases. The low braking energy is the mean of three braking conditions at initial velocity of 25 km/h while the medium braking and high braking energy are

Fig. 2. Total particle number concentration and particle generation rate measured by (a) FMPS, (b) OPC and (c) APS.

172

H.-G. Namgung et al. / Wear 372-373 (2017) 169–176

the mean values of three braking conditions at initial velocity of 50 km/h and 100 km/h, respectively. The FMPS instrument, which measures the number concentration of particles of 56–560 nm, shows that the total concentration and generation rate increase explosively as braking energy increases. The y axis in the graph is the log scale. On the other hand, the APS instrument, which measures particles of 0.5–20 μm, shows that total number concentration of generated particles increases in a linear manner as braking energy increases, but the generation rate does not differ much between medium braking energy and high braking energy cases. APS measurement in the first channel (0.523 μm or smaller) is not included in total number concentration and generation rate. The OPC instrument, which measures particles of 0.25–32 μm, also shows that the total number concentration of generated particles increases in a linear manner as braking energy increases, but the particle generation rate decreases in the medium and high braking energy cases. Although the APS and OPC measurements show similar particle generation characteristics as the braking energy increases, the total number concentration and particle generation rate of the OPC measurement in terms of absolute concentration. 3.2. Size distribution measured by each instrument Fig. 3 shows the number distribution of wear particles according to braking energy. Results of the FMPS measurement (Fig. 3(a)) show somewhat complex peaks of particles less than about 80 nm in size. However, the size distribution of larger particles show a unimodal shape with a peak at around 190–200 nm.

Fig. 3. The number distribution of wear particles according to braking energy (EB) measured by (a) FMPS, (b) OPC and (c) APS.

Only the concentration increases while the shape does not, even when braking energy increases. For particles of 0.08 μm or less, the overall shapes of size distribution are the same for low-EB and mid-EB while the number concentration of mid-EB is around 11 times higher. However, the number concentration of 10 nm particles increases under mid-EB. Under the high-EB condition, the peak at 10 nm makes up the maximum number concentration in all sizes, and the concentration also increases in particles larger than 10 nm (the y-axis is in log scale). Such characteristics of size distribution according to braking energy are clearly confirmed by the graph in Fig. 3(a) presented in normalized concentration form. The brake disc-pad components are nucleated into new particles (around 10 nm in size) through evaporation and condensation as the braking condition becomes more severe, and the size distribution widens since particles of various sizes are generated during the process of coagulation (agglomeration) by collision between the newly generated and existing particles as the number concentration rapidly increases [1,4,9,23,25,28,29]. The OPC measurement result shown in Fig. 3(b) shows a trimodal size distribution with the peak diameter formed at 0.35– 0.4 μm and the second and third peaks at 0.54 μm and 0.75 μm, respectively. However, the shape of size distribution does not change with the braking energy (small box in the figure). This OPC measurement result is very similar to the size distribution (major peak at around 0.3–0.4 μm) in previous studies on brake wear particles using the same measuring instrument [11,24,29–31]. Wahlstrom et al. [11] reported similar peaks of 0.1, 0.28, 0.35 and 0.55 μm on four different brake pad materials measured by OPC in a pin-on-disc braking test. While the number concentration does not change much as braking energy increases in the first peak of 0.35–0.4 μm, it increased 4.3 times with mid-EB and 9.8 times at high-EB over the concentration at low-EB for 0.54 μm, and 9 times and 21 times, respectively, in the case of the third peak of 0.75 μm. Although the particles generated increase in the APS measurement as they do in the OPC measurement, the characteristics of size distribution do not change (Fig. 3(c)). As the number in the small box shows, there is no significant change in the shape of size distribution. Compared to the result of OPC measurement, which measures a similar size range, APS shows a unimodal shape with a peak size of 1–1.2 μm. Previous studies involving measurement of brake wear particles reported similar characteristics of size distribution [9,20,32]. Size measurement using a cascade impactor (MOUDI) measuring the aerodynamic particle diameter also showed a major peak of around 1–2 μm [3]. Table 3 shows the mean diameter calculated by each device for each braking-contact force at low, medium and high braking energy. FMPS calculates the mean diameter separately by judging that there is a difference in generation of large and small particles on the basis of 80 nm as shown in Fig. 3(a). At 80 nm or less, there was almost no difference according to braking-contact force, with low-EB and mid-EB showing mean diameters of 30.4 and 16.7 nm, respectively. Although this similarity existed in mean diameter according to braking-contact force at 80 nm or less, the mean particle size was affected greatly by braking energy and resulted in particles being smaller by around 46%. The mean diameter increased to 13.3 nm, 16.5 nm and 22.3 nm at the braking-contact force of 12, 20 and 30 kN, respectively, under high-EB conditions. This is attributed to the fact that the mean diameter increased as very small particles agglomerated as the total number concentration of generated particles increased 3.7 times and 5.6 times at 20 and 30kN, respectively, compared to 12kN under high-EB conditions as shown in Table 2 previously. The mean diameters of particles 80 nm or larger in FMPS measurements were 209.1 nm and 201.1 nm without much difference according to braking-contact force under low-EB and mid-EB conditions: mean diameter decreased by around 4% as braking energy increased. Under

H.-G. Namgung et al. / Wear 372-373 (2017) 169–176

173

Table 3 Comparison of mean diameter measured by FMPS, OPC and APS. Braking energy (Initial velocity)

Braking-contact force (kN)

Mean diameter

APS/OPC conversion factor (C.F)

FMPS (nm) o 80 nm

480 nm

OPC (μm)

APS (μm)

Mean diameter ratio

Mode diameter ratio

Low-EB (25 km/h)

12 20 30 Avg.

29.8 33.4 28.0 30.4

208.8 208.6 210.1 209.1

0.41 0.43 0.42 0.42

1.32 1.36 1.35 1.34

3.22 3.16 3.21 3.20 7 0.03

3.43 2.57 3.43 3.14 7 0.40

Mid-EB (50 km/h)

12 20 30 Avg.

16.0 16.2 17.8 16.7

201.9 203.1 198.4 201.1

0.38 0.38 0.38 0.38

1.25 1.37 1.37 1.33

3.29 3.61 3.61 3.50 7 0.15

3.43 3.19 3.43 3.35 7 0.11

High-EB (100 km/h)

12 20 30 Avg.

13.3 16.5 22.3 17.4

196.1 195.8 181.3 191.1

0.37 0.36 0.36 0.36

1.41 1.45 1.54 1.47

3.8 4.0 4.0 4.04 7 0.19

3.43 2.97 3.96 3.45 7 0.41

21.5

200.4

0.39

1.38

3.58 7 0.37

3.32 7 0.36

Total Avg.

high-EB conditions, mean diameters rapidly decreased from 196.1 nm to 195.8 and 181.3 nm as the braking-contact force increased from 12 to 20 and 30kN. The decrease in the mean diameter of particles 80 nm or larger is attributed to the increased total number concentration of particles around 80 nm as shown in Fig. 3(a). The FMPS measurement showed the mean diameter of particles 80 nm or larger decreased as braking energy increased. The OPC measurement showed almost no difference in mean diameter according to braking-contact force at each braking energy condition. The mean diameter decreased from 0.42 mm to 0.38 and 0.36 mm as braking energy increased from low to medium to high-EB, respectively. On the other hand, the APS measurement showed similar mean diameters under low-EB and mid-EB conditions (1.34 and 1.33 mm respectively), while mean diameter increased to around 1.47 mm under high-EB conditions. The total average of mean diameters calculated by each instrument was 21.5 nm and 200.4 nm with FMPS, and the average mean diameters were 0.39 and 1.38 μm with OPC and APS, respectively. Although APS and OPC were expected to show similar results since the measured size ranges were 0.5–20 mm and 0.25–30 mm, respectively, actual results showed a significant difference. The studies by Sanders et al. [3] and Wahlstrom et al. [11] confirmed that the aerodynamic diameters and optical diameters can be different and can be converted to each other at a fixed ratio. This study also presented the mean diameter ratio and mode diameter ratio using the sizes measured by APS and OPC. The average mean diameter ratio was calculated to be 3.20, 3.50 and 4.04 under low, medium and high-EB conditions, respectively, as shown in Table 3. 3.3. Comparison of APS and OPC measurement results Fig. 4(a) shows the average values of measurements under each braking energy condition of OPC and APS for comparison. Although the two instruments showed similar dN/dlogDp concentration values (around 42,000 cm  3) at the peak, the mean diameter differed by around 3.58 times as the OPC and APS calculated the mean diameters to be 1.38 μm and 0.39 μm, respectively. If the two instruments measured the same wear dusts and the result differed only because of the difference in measurement method, the diameters measured by APS can decrease by 3.58 times (mean diameter based) as shown in Fig. 4(b). Or the diameters can decrease by 3.32 times (mode diameter based) based

Fig. 4. Comparison of wear particle number concentration measured by OPC and APS: (a) raw data, (b) corrected by APS/OPC diameter ratio, (c) corrected by particle density.

174

H.-G. Namgung et al. / Wear 372-373 (2017) 169–176

on peak diameter. In both cases, the APS measurement includes all OPC measurements and shows similar size distribution. Peters et al. [33] measured polystyrene latex particles (PSL) and Arizona dust particles using OPC and APS simultaneously and reported a systematic difference in number concentration according to diameter. They estimated OPC measurement to be closer to actual concentration than APS measurement for particles in the 0.3– 0.7 μm diameter range since the optical method is more accurate for measuring the number and diameter of particles. Fig. 4(c) shows the size distribution measured by APS and corrected with estimated particle density and OPC measurements. The diameters measured by APS can be larger than the actual physical diameters since the density of wear dust is higher than unit density (1 g/cm3). The case assuming ρp ¼ 5 g/cm3[3] and the case where wear dust is iron (Fe) particles (ρp ¼ 8 g/cm3) are presented together. If the density increases to 5 or 8 times unit density, the diameter as measured by APS decreases by 5 , 8 times, respectively [34], and thus the size distribution shifts to the left. OPC and APS size distributions are based on the assumption that wear dusts are all spherical particles. Fig. 5 presents images of the airborne wear particles sampled at the same time as size distribution measurement by using the TEM (F20 G2, Tecnai) and FE-SEM (S5000H, Hitachi) analysis. As shown in Fig. 5(a), the wear dusts generated during braking are non-spherical particles of various shapes and sizes (mostly 1 μm or smaller). The SEM images shown in Fig. 5(b) and (c) show particles that were torn and agglomerated, respectively, during braking. Since the wear dust generated during braking has a non-spherical or agglomerate shape, the aerodynamic shape factor (χ) of particles is larger than 1. Since the time-of-flight of individual particles shortens if their aerodynamic shape factor is greater than 1 when measuring the diameters with APS, the estimated sizes are smaller than the actual volume equivalent diameter [33]. On the other hand, OPC, which measures the refractive index of light scattering of individual particles, tends to overestimate the diameters since the agglomerate particles have a higher refractive index than spherical particles of equivalent volume [35,36]. A study reported that, in the measurement of Arizona dust (density ¼2.65 g/cm3, χ ¼1.5, refractive index ¼1.5), which is generally used as the standard test particle, using APS and OPC simultaneously, OPC showed a mean diameter of 0.51 70.01 μm while APS showed a mean diameter of

0.867 0.09 μm, or around 1.7 times larger [33]. Therefore, the physical properties (density, refractive index and aerodynamic shape factor) of airborne wear particles must be accurately understood, and the fact that such physical properties can change according to the size must also be considered.

4. Discussion Size distribution characteristics according to braking condition were measured simultaneously with three instruments using different measurement principles. The diameter measurement results from 5.6 nm to 30 μm provided by the instruments are presented in Fig. 6. Since the APS instrument measures with the assumption of unit density as described above, the APS-measured diameters were corrected with the APS/OPC mean diameter ratio calculated for braking energy shown in Table 3 in consideration of the fact that the density of wear particles is actually greater (APS overestimates diameter). Moreover, a case where particle density was 5 g/cm3 was also presented for APS, reflecting the findings of a previous study that the average density of wear particles generated during braking was 5 g/cm3[3]. Fig. 6(a) shows the mean diameter at three different brakingcontact forces under low braking energy. For APS, the mean diameter ratio to OPC (3.2) was applied to the measured values. FMPS had a peak value at around 0.1–0.2 μm while OPC and APS had similar peak values at around 0.4 μm, with a wide size distribution to around 8 μm. However, the number concentration as measured by OPC was reversed at around 1.3 μm from the measurement by APS. When particle density was assumed to be 5 g/cm3, the APS measurement had higher measured values and wider size distribution than the OPC measurement. Although the OPC and APS size distribution of braking particles in mid-EB conditions was not much different from that in low-EB conditions, they differed in the FMPS measurements. While the existing peak diameter of 0.1–0.2 μm was maintained, the number concentration increased by around 10 times, and the generation of new nanoparticles was observed. The increase in number concentration was particularly notable in the 10 nm region (Fig. 6(b)). Such a phenomenon was even clearer as braking energy increased. The existing peak at 0.1– 0.2 μm was maintained while the number concentration increased as shown in Fig. 6(c). The number concentration of particles of

Fig. 5. Image analysis of airborne wear particle by TEM image (a), and FE-SEM images (b, c).

H.-G. Namgung et al. / Wear 372-373 (2017) 169–176

2.

3.

4.

5. Fig. 6. Whole wear particle number distributions measured by three instruments under different braking energies: (a) low-EB, (b) mid-EB and (c) high-EB.

0.1 μm or less increased by more than 100 times compared to those at mid-EB conditions. The difference between the APS measurements when applying a conversion factor of 3.9 and the OPC measurement increased greatly under high-EB conditions. The difference between the APS and OPC results increased as diameter increased when particle size was 0.7 μm or greater. On the other hand, the APS and OPC measurements were very similar when a density of 5 g/cm3 was assumed. Although it may be difficult to observe the physical properties of different diameters provided by different measuring instruments with a graph as described in the above paragraph, analysis indicates that the size distribution characteristics of particles generated during braking can be divided into mainly two types according to braking energy condition: particles around 0.2–10 μm that have similar distribution characteristics regardless of braking energy, and the nanoparticles of 60–70 nm or less generated only when braking energy increases. While the OPC instrument, which measures size using light-scattering properties of particles, and the APS instrument, which measures aerodynamic diameters, have a similar measurement range, the measured values must be corrected in consideration of the physical properties of the particles. Applying a simple correction method using the average mean diameter ratio of OPC and APS, as done in this study, is expected to help with understanding the overall size distribution characteristics of wear particles generated during braking. 5. Conclusions

1. This study analyzed size distribution of airborne brake wear

175

particles generated under various braking conditions using a full-scale brake dynamometer and brake disks and NAO brake pads used in urban railway vehicles. Three types of aerosol instruments (FMPS, OPC and APS) that can measure particles 5.6 nm–32 μm in size were used simultaneously, and the particles were sampled for SEM and TEM image analysis. FMPS, which measures particles of 560 nm or smaller, showed a geometrically increasing, total number concentration of generated particles and particle generation rate as the braking energy increased. On the other hand, APS and OPC showed the total number concentration of generated particles increasing in a linear manner as braking energy increased, but the particle generation rate remained almost constant. Although OPC and APS measure a similar size range, the measurement results of mean diameter were significantly different. This is because APS measurement provides aerodynamic diameter. This study shifted APS size distribution using the mean diameter ratio of APS to OPC. Since the mean diameter of brake wear particles measured with OPC was 0.39 μm, or around 3.58 times the APS measurement of 1.38 μm, the diameters measured by APS were reduced by 3.58 times, confirming that the size distributions of the two instruments were similar. The OPC and APS size distributions are based on the assumption that the wear dusts are all spherical particles. SEM and TEM image analysis confirmed that the wear dusts generated during braking were non-spherical particles of various shapes and sizes. Therefore, the physical properties (density, refractive index and aerodynamic shape factor) of generated particles must be accurately understood, and the fact that such physical properties can change according to the size must be considered. The measurement of wear dusts generated during braking under each braking energy condition using the FMPS, OPC and APS instruments showed that the size distribution characteristics of particles generated during braking could be divided mainly into two types according to the braking energy condition. The first type was particles up to 10 μm in size and having a peak concentration at 0.2–0.75 μm regardless of braking energy, while the second type was particles around 10 nm in size generated only when the braking energy increased (nanoparticles generated through the evaporation-condensation process on the friction surface) and particles that increased in size to 100 nm through agglomeration. The number concentration of such nanoparticles geometrically increased as braking energy increased.

Acknowledgement This research was supported by ‘Development of Air Pollutant Removal Technology for Operation Train (16RTRP-B082501-03)' of the Railway Technology Research of the Ministry of Land, Infrastructure and Transport (MOLIT).

References [1] B.D. Garg, S.H. Cadle, P.A. Mulawa, P.J. Groblicki, C. Laroo, G.A. Parr, Brake wear particulate matter emissions, Environ. Sci. Technol. 34 (2000) 4463–4469. [2] P.J. Blau, H.M. Meyer III, Characteristics of wear particles produced during friction tests of conventional and unconventional disc brake materials, Wear 255 (2003) 1261–1269. [3] P.G. Sanders, N. Xu, T.M. Dalka, M.M. Maricq, Airborne brake wear debris: size distribution, composition, and a comparison of dynamometer and vehicle tests, Environ. Sci. Technol. 37 (2003) 4060–4069. [4] J. Sundh, U. Olofsson, Relating contact temperature and wear transitions in a wheel-rail contact, Wear 271 (2011) 78–85. [5] T. Grigoratos, G. Martini, Brake wear particle emissions: a review, Environ. Sci. Pollut. Res. 22 (2015) 2491–2504.

176

H.-G. Namgung et al. / Wear 372-373 (2017) 169–176

[6] P.C. Verma, L. Menapace, A. Bonfanti, R. Ciudin, S. Gialanella, G. Straffelini, Braking pad-disc system: wear mechanisms and formation of wear fragments, Wear 322–323 (2015) 251–258. [7] F. Amato, F.R. Cassee, H.A.C.D. van der Gon, R. Gehrig, M. Gustafsson, W. Hafner, R.M. Harrison, M. Jozwicka, F.J. Kelly, T. Moreno, A.S.H. Prevot, M. Schaap, J. Sunyer, X. Querol, Urban air quality: the challenge of traffic nonexhaust emissions, J. Hazard. Mater. 275 (2014) 31–36. [8] M. Mosleh, P.J. Blau, D. Dumitrescu, Characteristics and morphology of wear particles from laboratory testing of disk brake materials, Wear 256 (2004) 1128–1134. [9] J. Kukutschová, P. Moravec, V. Tomášek, V. Matějka, J. Smolík, J. Schwarz, J. Seidlerová, K. Šafářová, P. Filip, On airborne nano/micro-sized wear particles released from low-metallic automotive brakes, Environ. Pollut. 159 (2011) 998–1006. [10] J.K. Gietl, R. Lawrence, A.J. Thorpe, R.M. Harrison, Identification of brake wear particles and derivation of a quantitative tracer for brake dust at a major road, Atmos. Environ. 44 (2010) 141–146. [11] J. Wahlstrӧm, A. Sӧderberg, L. Olander, A. Jansson, U. Olofsson, A pin-on-disc simulation of airborne wear particles from disc brakes, Wear 268 (2010) 763–769. [12] O. von Uexkűll, S. Skerfving, R. Doyle, M. Braungart, Antimony in brake pads-a carcinogenic component? J. Clean. Prod. 13 (2005) 19–31. [13] H.L. Karlsson, L. Nilsson, L. Mӧller, Subway particles are more genotoxic than street particles and induce oxidative stress in cultured human lung cells, Chem. Res. Toxicol. 18 (2005) 19–23. [14] T. Kuwayama, C.R. Ruehl, M.J. Kleeman, Daily trends and source apportionment of ultrafine particulate mass (PM0.1) over on annual cycle in a typical California city, Environ. Sci. Technol. 47 (2013) 13957–13966. [15] M. Gasser, M. Riediker, L. Mueller, A. Perrenoud, F. Blank, P. Gehr, B. RothenRutishauser, Toxic effects of brake wear particles on epithelial lung cells in vitro, Part. Fibre Toxicol. 6 (2009) 1–13. [16] J. Sternbeck, Å. Sjӧdin, K. Andréasson, Metal emissions from road traffic and influence of resuspension-results from two tunnel studies, Atmos. Environ. 36 (2002) 4735–4744. [17] N. Bukowiecki, R. Gehrig, M. Hill, P. Lienemann, C.N. Zwicky, B. Buchmann, E. Weingartner, U. Baltensperger, Iron, manganese and copper emitted by cargo and passenger trains in Zurich(Switzerland): size-segregated mass concentrations in ambient air, Atmos. Environ. 41 (2007) 878–889. [18] D.S.T. Hjortenkrans, B.G. Bergbäck, A.V. Häggerud, Metal emissions from brake linings and tires: case studies of Stockholm, Sweden 1995/1998 and 2005, Environ. Sci. Technol. 41 (2007) 5224–5230. [19] P. Pant, R.M. Harrison, Estimation of the contribution of road traffic emissions to particulate matter concentrations from field measurement: a review, Atmos. Environ. 77 (2013) 78–97. [20] A. Iijima, K. Sato, K. Yano, M. Kato, K. Kozawa, N. Furuta, Emission factor for antimony in brake abrasion dusts as one of the major atmospheric antimony sources, Environ. Sci. Technol. 42 (2008) 2937–2942.

[21] N. Bukowiecki, P. Lienemann, M. Hill, R. Figi, A. Richard, M. Furger, K. Rickers, G. Falkenberg, Y. Zhao, S.S. Cliff, A.S.H. Prevot, U. Baltensperger, B. Buchmann, R. Gehrig, Real-world emission factors for antimony and other brake wear related trace elements: size-segregated values for light and heavy duty vehicles, Environ. Sci. Technol. 43 (2009) 8072–8078. [22] H. Hagino, M. Oyama, S. Sasaki, Airborne brake wear particle emission due to braking and accelerating, Wear 334 335 (2015) 44–48. [23] M. Mathissen, V. Scheer, R. Vogt, T. Benter, Investigation on the potential generation of ultrafine particles from the tire-road interface, Atmos. Environ. 45 (2011) 6172–6179. [24] S. Abbasi, J. Wahlstrӧm, L. Olander, C. Larsson, U. Olofsson, U. Sellgren, A study of airborne wear particles generated from organic railway brake pads and brake discs, Wear 273 (2011) 93–99. [25] H.G. Namgung, J.B. Kim, S.H. Woo, S. Park, M. Kim, M.S. Kim, G.N. Bae, D. Park, S.B. Kwon, Generation of nanoparticles from friction between railway brake disks and pads, Environ. Sci. Technol. 50 (2016) 3453–3461. [26] UIC code 541-3, Brakes-Disk brakes and their application – General conditions for the approval of brake pads, International Union of Railways, 6th edition, November, 2006. [27] T.M. Peters, D. Leith, Concentration measurement and counting efficiency of the aerodynamic particle sizer 3321, J. Aerosol. Sci. 34 (2003) 627–634. [28] J. Kwak, S. Lee, S. Lee, On-road and laboratory investigations on non-exhaust ultrafine particles from the interaction between the tire and road pavement under braking conditions, Atmos. Environ. 97 (2014) 195–205. [29] U. Olofsson, A study of airborne wear particles generated from the train traffic-block braking simulation in a pin-on-disc machine, Wear 271 (2011) 86–91. [30] J. Wahlstrӧm, A. Sӧderberg, L. Olander, U. Olofsson, A disc brake test stand for measurement of airborne wear particles, Lubr. Sci. 21 (2009) 241–252. [31] J. Wahlstrӧm, U. Olofsson, A field study of airborne particle emissions from automotive disc brakes, J. Autom. Eng. 229 (2015) 747–757. [32] A. Iijima, K. Sato, K. Yano, H. Tago, M. Kato, H. Kimura, N. Furuta, Particle size and composition distribution analysis of automotive brake abrasion dusts for the evaluation of antimony sources of airborne particulate matter, Atmos. Environ. 41 (2007) 4908–4919. [33] T.M. Peters, D. Ott, P.T. O’Shaughnessy, Comparison of the Grimm 1.108 and 1.109 portable aerosol spectrometer to the TSI 3321 aerodynamic particle sizer for dry particles, Ann. Occup. Hyg. 50 (2006) 843–850. [34] W.C. Hinds, Aerosol Technology: Properties, Behavior, and Measurement of Airborne Particles, second ed., John Wiley & Sons, New Jersey, Hoboken, 2012. [35] M. Quinten, R. Friehmelt, K.F. Ebert, Sizing of aggregates of spheres by a whitelight optical particle counter with 90° scattering angle, J. Aerosol Sci. 32 (2000) 63–72. [36] C.H. Chien, A. Theodore, C.Y. Wu, Y.M. Hsu, B. Birky, Upon correlating diameters measured by optical particle counters and aerodynamic particle sizer, J. Aerosol Sci. 101 (2016) 77–85.