Powder Technology 211 (2011) 38–45

Contents lists available at ScienceDirect

Powder Technology

j ourna l homepage: www.e lsev ie r.com/ locate /powtec

Investigation of the recyclability of powder coatings

Jing Fu a,⁎, Matthew Krantz a, Hui Zhang a, Jesse Zhu a, Harry Kuo b, Yar Ming Wang b, Karen Lis b

a Department of Chemical and Biochemical Engineering, University of Western Ontario, London, Ontario, Canada, N6A 5B9b General Motors, GM R&D, Warren, MI, 48090-9055, United States

⁎ Corresponding author.E-mail address: jzhu@uwo.ca (J. Fu).

0032-5910/$ – see front matter © 2011 Elsevier B.V. Adoi:10.1016/j.powtec.2011.03.016

a b s t r a c t

a r t i c l e i n f o

Article history:Received 27 September 2010Received in revised form 10 March 2011Accepted 20 March 2011Available online 12 April 2011

Keywords:Powder coatingPowder recyclabilityElectrostatic spraying

100% recyclability is one of the major advantages of powder coating. However, it can never be achieved inreality. Coating powders, especially finer powders with particle size below 30 μm, were found to have muchworse flow performances after recycling from electrostatic spraying so as to decrease the recyclability.Therefore, this study was designed to investigate recycled coating powders to determine the underlying causeof decreased flow performance. The investigations were based upon three major factors that make thedifferences between original powder and its recycled powder: particle size, humidity exposure and flowadditive concentration. By adjusting the three factors independently, the influences to powder flow propertieswere analyzed. Results showed that the decreased particle size of the recycled powder had the mostsignificant effect on the flow properties. Additive concentration on the powder particles did not change withrespect to the particle specific surface area after electrostatic spraying. Humidity had only a minor effect onthe flow properties of powder coatings.

ll rights reserved.

© 2011 Elsevier B.V. All rights reserved.

1. Introduction

Powder coating with no emission of volatile organic compounds(VOCs) is a clean process. It has received increasing attention as aresult of stricter environmental regulations and its superior perfor-mance inmany aspectswhen compared to liquid coatings [1,2]. One ofthe main advantages of the powder coating process is the potential tocollect all over-sprayed material that did not adhere to the work pieceand recycle it as reclaimed powder [3]. The amount of reclaimedpowder produced in a powder coating operation depends upon manyfactors such as: the size and shape of the object, the powder coatingmaterial, the coating equipment, the equipment settings andapplication environments. It is not uncommon for more than 50% ofthe original powder to be recycled as reclaim. Although near 100%material utilization can be reached by reclaiming over-sprayedpowders, the flow properties of the reclaimed powder always tendto deteriorate, which makes it more difficult to handle and apply thereclaimed powder [4,5].

The flow properties of a powder refer to the ability of free-flowingin normal handling processes. In order to improve the flow propertiesof reclaimed coating powders, they are often blended with virginpowder; however, this method is only effective if the reclaimedcoating powders are blended at a low concentration. As a result, manypowder coating applicators have an accumulation of reclaimedpowder because its flow properties are too poor to apply alone and

it cannot be blended at sufficient concentration to consume all of thereclaimed powder that is produced.

There are many causes that can lead to the decreased flowproperties seen in reclaimed powders. In this work, three potentialfactors were investigated to determine their influence on the flowproperties of reclaimed powders.

1. Reclaimed coating powders have a smaller particle size whencompared to their virgin counterparts. This occurs because, duringelectrostatic spraying, the larger particles have a higher momen-tum than the smaller particles and are, therefore, less likely to beblown away from the work piece by the carrying gas or downwarddraft in the spray booth. The larger particles also hold a highercharge during electrostatic spraying, making them more likely todeposit on the work piece than smaller particles [6].

2. Reclaimed powders are exposed to moisture in the spray booth.This can affect the flow properties by introducing capillaryinterparticle forces and by changing the material properties ofthe coating through the absorption of water.

3. Nanoparticle additives, which are added to coating powders asflow aides, may change in concentration during the fluidization ofthe powder or during electrostatic spraying. A change in concen-tration of nanoparticle additives may explain the differences inflow properties between reclaimed and virgin powders.

The goal of this work was to individually investigate each of theaforementioned factors and determine which factor(s) have the mostinfluence on the flow properties seen in reclaimed powders. Byunderstanding the root causes of this flow property reduction issue,

39J. Fu et al. / Powder Technology 211 (2011) 38–45

new technologies and handling procedures can be developed toimprove the recycleability of reclaimed coating powders.

2. Materials and methods

Powder samples used for the investigation are polyurethane andpolyester-epoxy fine powderswithmediumparticle sizes of between25 μm and 28 μm. The particle size is closely related to the flowproperties of the powder. The small particle size of the fine powdercan cause significant reduction in flow properties, because of theincreased interparticle forces, primarily the van der Waals force [7].To overcome this challenge, nano flow additives are introduced [8].These additives are coated on the surfaces of the powder particles toincrease the interparticle distances, so as to reduce the van derWaalsforce [9].

To address each of the possible factors leading to decreased flowproperties in reclaimed powders, separate experiments were per-formed. The following techniques and equipment setups were used inthis work:

2.1. Particle size distribution

The particle size distribution was obtained by laser diffractionmeasurement (Mastersizer 2000, Malvern Instruments, Worcester-shire, UK) following standard test procedures. The results werereported by giving the values of D10, D50, D90, and the specific surfacearea. D10, D50, and D90 stand for powder particle distribution. Forexample, D10 is defined as a diameter where 10 vol.% of the particles ofthe powder is less or is equal to the diameter. D10, D90 are the popularparameters used to determine the amount of fine and coarse particles,and D50 represents the medium particle size of the powder samples.The specific surface area (SSA) is another output from the resultdefined as surface area per unit mass (m2/kg). Powders collected fromthree different manufacturers were analyzed in this experiment.

2.2. Flow properties

Flow properties were evaluated using a combination of powdercharacterization methods, such as: rotational bed expansion ratio(RBER), angle of repose (AOR), and avalanche angle (AVA) [10–12].

The fluidized bed expansion ratio was measured using a rotatingdrum (Revolution Powder Analyzer, Mercury Scientific Inc., SandyHook, CT,USA). In this test, 120 mLof powder (tapped volume) is placedinto a transparent drum (diameter of 10.9 cm, length of 3.5 cm) and thedrum is rotated at increasing speeds. As thedrumrotates faster,more airbecomes entrained into the powder and volume of the powderincreases. To gauge how well a powder fluidizes, the volumetricexpansion ratio (current powder volume divided by the initial powdervolume) wasmeasured via a video camera and computer software. Themaximum rotating speed of the drum for the measurement was set at70 rpm, because the powder could start to rotate with the drum at ahigher speed due to the centrifugal force. A high rotating bed expansionratio indicates that a powder is easily fluidized. A rotating fluidized bedwas used in this work instead of a traditional fluidized bed because therotating drum allowed the powder sample to be sealed, without thepossibility for loss of powder and loss/gain of moisture during testing.The results are reported as RBER.

AOR for each powder was measured using a powder characteristictester (PT-N Powder Characteristic Tester, Hosokawa Micron PowderSystems Co., Summit, NJ, USA). Tomeasure the AOR, the powder samplewas slowly dispensed onto a flat surface to form a pile. The AOR wasthen taken as the angle between the surface of the pile and the flatsurface. Six AOR measurements were completed for each sample andthe average is provided.

The AVAwasmeasured using the same Revolution Powder Analyzeras used for RBER. This instrument works by placing a known amount of

powder into a transparent drum andmonitoring how the powder flowsas the drum is rotated. The AVA was determined by rotating 120 mL ofpowder (tappedvolume) at0.6 rpmandmeasuring themaximumanglethat the powder achieves before it avalanches (collapses) to the bottomof the drum. This testwas performed until 200 avalanches occur and theaverage was calculated.

Among these three powder characterization methods, the RBER isa representation of the dynamic powder flow property, and the othertwo methods are the representations of more static powder flowproperties. The dynamic flow property indicates howwell the powderflow during fluidization or pneumatic transportation. On the otherhand, the more static flow property indicates how easy the powdergets agglomerated. Overall, a larger RBER value, a smaller AOR valueand a smaller AVA value imply better flow properties of a powder.

2.3. Moisture exposure

Moisture absorption is one difference in between virgin powdersand reclaimed powders. It is generally believed that low humidity isrequired to maintain the flow properties of coating powders andenormous effort is given to keeping virgin powders dry. On coatingpowder production and application lines, care is taken to store virginpowders in sealed packages and in a well air-conditioned environ-ment. When used, virgin powders are fluidized and pneumaticallytransported using air with a dew point below −100 °F (Lis, 2008).However, when virgin powders are applied in the spray booth, theyare exposed to an environment with a relative humidity typicallybetween 50 and 70%. This level of humidity is maintained duringelectrostatic spraying to help encourage charge relaxation andprevent back ionization, but it is uncertain how such exposure tomoisture affects the flow properties of reclaimed powder. Throughoutthe spraying process, the reclaimed powder is periodically removedfrom the spray booth and stored in an air conditioned room until it isused.

To investigate the effects of moisture exposure, coating powderswere exposed to a humidified environment for duration up to 580 h.Over 500 g powder sample was placed in a beaker (diameter of12.5 cm, depth of 18.3 cm) and placed in a sealed barrel (diameter of28.5 cm, depth of 43.8 cm), which contained about 1.5 L of water. Themeasured relative humidity inside the barrel was 99%. Since thepowders were exposed to moisture while in a settled state (notfluidized), no material loss would occur during the process. Moistureabsorption was measured by weight gain and the flow propertieswere evaluated by RBER, AOR and AVA.

2.4. Additive concentration

The concentration of the nanometer sized additive was deter-mined using ASTM D5630-06 Standard Test Method for Ash Content inPlastics [13] as a guideline. Under this test method, powder coatingsamples were heated at 550 °C in a ceramic crucible to remove allcombustible material from the powder coating, leaving behind an ashof non-combustible material (including the nanometer sized flowadditives). For this test work, an acrylic clear coat powder (DuPontPerformance Coatings: Acrylpulverklarlack) was used because it doesnot contain inorganic fillers or pigments thatwould also remain in ash.By using acrylic clear coat powders, the non-combustible componentof the powder is predominantly the additive, and any changes in ashweight can be directly correlated to additive concentration in thepowder sample. These acrylic powders were ground using an airclassifying mill (ACM) and loaded with 0.8 wt.% nano additives.

During electrostatic spraying, three types of powders collected foradditive concentration analysis were denoted: virgin, transferred andreclaimed powders. The virgin powder was the powder prior to thespraying; the transferred powder was the powder attached to the

40 J. Fu et al. / Powder Technology 211 (2011) 38–45

substrate after spraying and the reclaimed powder was the powderremaining on spray booth walls.

2.4.1. Equipment setupsTo understand the effects of fluidization and electrostatic spraying

on additive concentration, a small fluidization column and alaboratory scale spray booth were constructed. For fluidization testwork, a small fluidized bed, 5.08 cm in diameter and 30 cm in height,was used to fluidize sample powders at a superficial gas velocity of0.5 cm/s. This gas velocity was selected to match the fluidizationof powder coatings in a typical fluidized hopper. Throughout thefluidization process, powder was sampled from the fluidized bed andanalyzed for additive concentration and particle size. For electrostaticspraying test work, a small spray booth was constructed as shown inFig. 1. This spray booth was 60.96 cm in all dimensions and attachedto a vacuum (Nilfisk-GM80) to provide air flow. An ITWGemamanualspray gun (ITW Gema EasySelect) was used to spray powder at a30 cm diameter target from a distance of 20 cm. A voltage of 35 kVwas applied and powder flow rate of 5.5 g/min was used. After 5 g ofpowder was sprayed, all powder that attached to the target wasremoved and segregated before more powder was sprayed (repeating5 g intervals). Powder that did not attach to the target (reclaimed)was mainly collected by the filter on the back wall, with someattached to other inner surfaces of the spray booth and recovered atthe end of the test. After spraying a total of 200 g, samples of the virginpowder, coated powder, and reclaimed powder were analyzed foradditive concentration and particle size.

The filter at the back of the spray booth consisted of a mediamaterial from a vacuum bag (NILFISK Ref No 82095000) and a framethat held the media. The bag had two plies, whose first ply was toblock coarse particles, and the second ply was to block fine particles.In the experiment, the NILFISK vacuum bagwas not used as is, instead,the second ply was placed in front of the first ply, so that the fineparticles and coarse particles were stopped at the first stage. To avoidextra loss of powder through the clean filter, about 200 g of powderwas pre-sprayed to saturate the medium, before the experiment.

The powder from the target and the spray booth as well as thebooth filter was collected with a self-made device shown in Fig. 2. Itconsisted of a canister and a cylindrical filter (Whatman CatalogueNo. 2810432). The canister exhaust was connected to a vacuum line toprovide suction. The powder on the sprayed target was sucked intothe canister and then trapped by the filter, when the suction hose was

Fig. 1. Laboratory scale spray

moved and placed very close to the target. After that, the filter wasdetached from the canister and the collected particles were recoveredfrom the filter and stored in a plastic bag for further analysis. Thanksto this device, the powder collecting process was much easier andmore accurate then brushing or scraping. It was estimated from themass balance that the collecting and transporting processes led toa powderweight loss of about 6%. Because the canister filter wasmuchfiner than the spray booth filter, pre-spraying process was notrequired.

2.5. Blending of virgin and reclaimed coating powders

In this experiment, three series of powder mixtures were used forthe analyses: a) mixture of virgin and real reclaimed powders (V–R);b) mixture of virgin and ground virgin powders (V–GV) and c)mixture of virgin and sieved virgin (V–SV). The virgin powder wasthe grey virgin powder from Manufacturer A. The real reclaimedpowder was the recycled powder from the same virgin powdercollected on the actual powder coating production line. On the otherhand, ground virgin and sieved virgin were used to simulate thereclaimed powders. The ground virgin powder was made by using agrinder to grind the virgin powder to similar size as the realreclaimed powder. The sieved powder was prepared by using anultrasonic-vibrating sieve to remove the large particles to obtain apowder having a similar size as the real reclaimed powder. Theblending ratios of virgin powder to reclaimed powder were 100%,75%, 50%, 25% and 0% in weight fraction (wt.%) for the three series ofpowder mixtures.

The virgin–reclaimed mixture series (V–R) was used to determinehow the flow properties of reclaimed powder were affected. Theresults was also compared to the virgin–simulated reclaimed mixtureseries (V–GV and V–SV) to determinate if the changes in flowproperties between virgin and reclaimed powders were primarily aresult of particle size. Since the simulated reclaimed powders have thesame size as the reclaimed powder, their use will eliminate the effectof size shifting during analyses. If characterization results (RBER, AORand AVA) obtained from the virgin–simulated powder mixture seriesare the same as the virgin–reclaimed powder, then the only factorcausing reduction in flow properties is the particle size. Otherwise,other factors such as humidity and additive concentration may also beresponsible for the flow properties deterioration.

booth and its side view.

Fig. 2. Method of transporting powder from a target.

41J. Fu et al. / Powder Technology 211 (2011) 38–45

3. Results and discussion

3.1. Particle size distribution

Decreasing the particle size of a powder has been shown to resultin decreased flow properties of Group C powders. This trend is welldocumented in the powder technology literature [7,14,15]. In thiswork, virgin polyurethane coating powder from Manufacturer A,polyester-epoxy coating powders from Manufacturers B and C, andtheir corresponding reclaimed powders collected from the actualcoating line were analyzed. These Group C coating powders were pre-loaded with nano additives. However, the amount of the nanoadditives was small (in the range from 0.2% to 0.8%), thus, it did notchange the Geldart's group that the powders belong to. The averageparticle size distributions for each manufacturer are provided inTable 1. It can be seen that coating powders from all manufacturersshow a decrease in particle size when they are reclaimed. Thisdecrease in particle size is typically about 2 μm for the D10 and D50,this decrease in particle size can result in changes to flow properties.According to Geldart's powder classification system, there is atransition between Group A powders (aeratable and easily fluidized)and Group C powders (cohesive) when the D50 particle size decreasesbelow 25–27 μm for powders of roughly 1.4 specific gravity, which isof most coating powders [14,15]. Because the particle size of virginpowders already borders the transition between Geldart Groups A andC, a small decrease in particle size can result in a significant reductionin flow properties [16]. This is expected to be a major contributor tothe differences seen between virgin and reclaimed powders; this mayalso explain reports of inconsistent performance found in variousvirgin powders on the market.

Additionally, the influence of van der Waal's interparticle cohesiveforce dramatically increases when the particle size decreases below10 μm [7,16]. This suggests that the amount of fine material in coatingpowders, that below 10 μm, will also play a governing role indetermining the flow properties. From Table 1, it is seen that there

Table 1Average particle size distributions for virgin and reclaimed powders produced byManufacturers A, B and C.

Manufacturer Virgin powder Reclaimed powder

D10(μm)

D50(μm)

D90(μm)

D10(μm)

D50(μm)

D90(μm)

A 13.1 27.5 52.2 11.8 26.0 51.8B 12.7 24.7 45.3 10.5 22.3 43.0C 16.0 25.8 40.9 14.3 24.4 40.5

is an increase in the amount of fines between virgin and reclaimedpowders, which is represented by a decrease in D10.

3.2. Exposure to moisture

In this experiment, moisture absorption in coating powders wasinvestigated by exposing coating powders produced by ManufacturerA and Manufacturer B to an environment of 99% relative humidity atatmospheric pressure and a temperature of 23 °C. Fig. 3 shows themoisture absorption (byweight gain) for a virgin graypowder producedby Manufacturer A and a virgin gray powder from Manufacturer B. Itis seen that moisture absorption occurs fastest within the first 50 hof exposure and reaches saturation after roughly 200 h exposure.

Each powder was characterized for flow properties at exposures of0, 96, and 580 h, with the results summarized in Table 2. It can be seenthat there was little or no change in AOR (angle of repose) and AVA(avalanche angle) after an exposure of 96 h. A more significantdecrease in flow occurred after an exposure of 580h, when the AORand AVA values were significantly higher. This suggests that the flowproperties of these coating powders will not be significantly affectedby a brief exposure to humid environments. On the other hand,decreases in RBER (rotational bed expansion ratio) were found forboth powder samples. Fig. 4 indicates that the gray powder fromManufacturer B had a larger slope of decrease, and could not furtherperform the test after around 143h due to the poor flowability. Thedata from the rotating fluidized bed suggests humidity can signifi-cantly reduce the powder performance after a long exposure time.

Fig. 3. Moisture absorption for coating powders produced by Manufacturers A and B.

Table 2Flow properties of virgin powders exposed to humidity.

Sample AOR (°) AVA (°) RBER

Exposure time (hr) 0 96 580 0 96 580 0 96 580

Manf. A Gray 40.1 41.2 43.4 42.4 42.5 43.5 1.426 1.426 1.417Manf. B Gray 44.2 46.1 49.7 44.7 44.6 46.2 1.264 1.215 n/a

Table 3Effect of fluidization on the particle size distribution and the concentration of additive.

Time(h)

Δ Mass Particle size distribution(μm)

SSA(m2/kg)

Additiveconcentration(wt.%)

D10 D50 D90

0 n/a 8.65 25.16 73.64 317.29 0.8002:00 −1.1% 8.75 26.77 78.60 304.08 0.7656:15 −4.1% 8.90 27.65 79.03 286.74 0.739

42 J. Fu et al. / Powder Technology 211 (2011) 38–45

3.3. Changes in additive concentration

Nanometer sizedparticles are commonly added to coating powdersas a method of improving flow properties. However, it is not wellunderstood what happens to these additives when the powders areused in the powder coating process. For example, are the additivessusceptible to becoming entrained in fluidization gas and removedfrom the powder? Do the additives segregate and concentrate in thereclaimed powder or the powder that adheres to the work piece? Thissection discusses how the additive concentration in powder coatingschanges when fluidized and electrostatic sprayed.

3.3.1. FluidizationTo test the influence of fluidization on the concentration of

nanoparticle additives, special samples of acrylic clear coat powderscontaining a known initial concentration of nano additive, wereprepared for fluidization tests. During this experiment the change inadditive concentration was monitored.

As the powder was fluidized, it was observed that some of thepowder became entrained in the fluidizing gas and left the fluidizedbed. This was monitored by measuring the weight of the samplewithdrawn from the fluidized bed with respect to fluidization timeand the particle size is shown in Table 3. It was also observed that theparticle size distribution of the powder remaining in the fluidized bedincreased with fluidization time, indicating that more fine particleswere leaving the fluidized bed than coarse particles. This correspondsto a decreasing specific surface area (SSA) given that smaller particleshave larger surface area for a given mass. Finally, it was observed thatthe additive concentration also decreased from an initial concentra-tion of 0.80% (wt.%) to a final concentration of 0.74%.

Table 4 compares the change in powder mass to the change in SSAand additive concentration. It is observed that the change in additiveconcentration is similar to the change in SSA and the ratios of additiveconcentration-to-SSA are almost the same for the three samples. Thisresult supports the concept that nanoparticle additives uniformly coaton the surface of the powder coating particles and remain wellattached throughout the fluidization process. Because the additiveconcentration changes at a similar magnitude to the change in SSA ofthe powder and the additive did not change with respect to the SSA, itis believed that the decreasing additive concentration is a result of fine

Fig. 4. Rotational bed expansion ratio (RBER) of virgin powders exposed to humidity.

particles being preferentially entrained in the fluidization gas, and nota result of the additives separating from the coating powder and beinglost exclusively. In other words, it is expected that more of theadditives are attached to the finer particles because they have moresurface area, so when this size fraction is lost to the fluidization gas,the percentage of additive concentration decreases proportionally.

3.3.2. Electrostatic sprayingUnlike in fluidization, powder during electrostatic spraying was

not only affected by air flow but other factors including electrostaticforce. Following the experiments that simulated an actual electro-static spraying in a miniature spray booth, the particle size andadditive concentration were compared between virgin, transferredand reclaimed powders. Table 5 provides the particle size distribution,SSA and additive concentration of the powders. Compared to thevirgin powder, it was found that transferred powder had an increasedparticle size and the reclaimed powder had a decreased particle size.This result was expected and corresponds well with the industry datain Table 1. Table 5 also compares the change in SSA and concentrationof additive among the virgin, transferred and reclaimed powders.

In Table 6, it was found that the change in additive concentrationduring electrostatic spraying almost mirrored the change in SSA,which is similar to the fluidization tests. More importantly, the ratiosof additive concentration-to-SSA indicated that additive concentra-tions of virgin, transferred and reclaimed powder with respect to SSAstayed almost constant.

The findings from Table 6 support the hypothesis that thenanoparticles added to powder coatings remain attached to the powdercoating particles during powder coating processes. In other words, whenthe particle size distribution of powder coatings becomes segregated, theconcentration of the additives will change in proportion to the change inSSA.

3.4. Blending of virgin and reclaimed powder

Recalling from Section 2.5, three series of powder blends were usedfor the analysis in this experiment: a) mixtures of virgin and realreclaimed powders (V–R); b) mixtures of virgin and ground virginpowders (V–GV) and c) mixtures of virgin and sieved virgin (V–SV).Parameters such particle size distribution, RBER (rotational bedexpansion ratio), AOR (angle of repose) and AVA (angle of repose) areused to characterize the powder blends. First of all, the virgin powderwasblendedwith real reclaimedpowder at different ratios todetermineif the flow properties of reclaimed powder can be improved by dilutingit with virgin powder. After that, the two types of simulated reclaimedpowders (GV and SV) were blended with the virgin powder and

Table 4Comparison of change in mass, SSA, additive concentration of additive and additiveconcentration-to-SSA.

Time(h)

Δ Mass Δ SSA Δ Additiveconcentration

Additive concentration-to-SSA

0 n/a n/a n/a 0.002522:00 −1.1% −4.3% −4.2% 0.002526:15 −4.1% −7.7% −9.6% 0.00258

Table 5Effect of electrostatic spraying on the particle size distribution, SSA and additiveconcentration.

Sample Particle size distribution(μm)

SSA(m2/kg)

Additiveconcentration(wt.%)

D10 D50 D90

Virgin 7.17 28.14 68.36 146.01 0.876Transferred 9.19 32.99 75.62 115.89 0.664Reclaimed 5.54 23.84 61.59 185.96 1.147

0

20

40

60

80

0 20 40 60 80 100

Par

ticl

e si

ze (

µm)

Virgin powder (wt.%)

D90 V-GV mixture D50 V-GV mixture D10 V-GV mixture

D90 V-R mixture D50 V-R mixture D10 V-R mixture

D90 V-SV mixture D50 V-SV mixture D10 V-SV mixture

Fig. 5. Particle size distributions of mixtures of virgin powder with (a) finely groundvirgin powder (solid lines); (b) real reclaimed powders (dotted lines); (c) sieved virginpowders (dashed lines).

46

47

48

49

50

0 25 50 75 100

AV

A (

deg

.)

37

39

41

43

0 25 50 75 100

AO

R (

deg

.)

1.3

1.4

1.5

RB

ER

43J. Fu et al. / Powder Technology 211 (2011) 38–45

subjected to the same characterization tests. Characterization resultswere then compared between the three powder mixture series.

The V–R lines in Fig. 5 presents the particle size distribution of themixtures of virgin and real reclaimed powders ranging from 100%virgin powder to 100% reclaimed powder. It is shown that the particlesize decreases as the portion of the reclaimed powder in the blendincreases. Fig. 6 provides the corresponding flow properties of theseblends. All of the characterization techniques employed show that theflow properties decrease as the portion of the reclaimed powderincreases. This is signified by a decreasing RBER, increasing AOR, andincreasing AVA with increasing percentage of the reclaimed powder.

To discern if the decrease in flow observed for the reclaimedpowders was a consequence of the smaller particle size or some otherfactors inherent to the reclaimed powders, the mixtures of virgin andGV (ground virgin) powders were analyzed. The particle sizedistributions of the mixtures are shown as the V–GV lines in Fig. 5.As experimentally designed, the V–R lines and V–GR lines coincided,confirms similar in particle size distribution between these twopowder series.

In order to compare the characterization results between themixtures of virgin and real reclaimed powders and the mixtures ofvirgin and ground virgin powders, the flow properties were plotted asV-R and V-GR lines accordingly in Fig. 7. The unit of horizontal axis inFig. 7 was changed to particle size instead of percentage of virginpowder in Figs. 5 and 6. It is because the particle sizes of differentpercentage of virgin powders are different in V–R, V–SV and V–GVmixtures, even though they are very close as shown in Fig. 5.

When comparing these mixtures, it was found that the mixture ofvirgin and ground virgin powders performed worse when tested usingeach characterization technique. Thismay seem to be a surprising resultgiven that this powder mixture had no exposure to a humidenvironment, which can reduce powder performances as discussed inchapter 3.2; however, the additive concentrations between the twomixtures were not equal. Because the reclaimed powder had a smallerparticle size than the virgin powder, it is expected that the additiveconcentration in the reclaimed powder was higher in proportion to theincrease in its SSA. In contrast, the additive concentration in the groundvirgin powder remained unchanged since the powder was ground fromthe original virgin powder in a closed system with no opportunity foradditive gain or loss. Table 7 provides the corresponding change inadditive concentration and also compares the SSA of the virgin,reclaimed, and the ground virgin powders.

Data from Table 7 indicated that the finely ground virgin powderhad increased SSA due to decreased particle size, but at the same time,the additive concentration was not increased. Therefore the ratio ofadditive concentration-to-SSA was decreased, which caused the flow

Table 6Comparison of SSA to the concentration of additive for electrostatic sprayed powder.

Sample Δ SSA Δ Additive concentration Additive concentration-to-SSA

Virgin n/a n/a 6.00E-5Transferred −20.63 −24.16% 5.73E-5Reclaimed 27.36 30.92% 6.17E-5

properties of the ground virgin powder to become less than thereclaimed powder of the same particle size distribution. In otherwords, the additive concentration relative to the SSA of the powder

1.1

1.2

0 25 50 75 100

Virgin powder (wt. %)

Fig. 6. Flow properties of powder blends composed of virgin and real reclaimedpowders as a function of percent virgin powder content.

Fig. 7. Flow properties of blends of virgin powders with reclaimed powder, finelyground virgin powder and sieved virgin powder based upon D50.

44 J. Fu et al. / Powder Technology 211 (2011) 38–45

has a significant influence on flow properties. Therefore, in order tomatch both particle size distribution and relative additive concentra-tion with respect to the SSA, the sieved virgin powder was used tosimulate the real reclaimed powder. By following the sievingprocedure, the virgin powder was again reduced to similar size asthe real reclaimed powder shown as V–SV lines in Fig. 5. This particlesize reduction method was expected to have an analogous effect onthe additive concentration, as fluidization or electrostatic sprayingwould have, which maintained the additive concentration-to-SSAconstant, see Table 7.

The flow properties of the sieved virgin powders are also shown inFig. 7 as V–SV lines, where it can be seen that the sieved virginpowders had similar flow properties to the reclaimed powders of thesame particle size when analyzed using AOR and AVA. This suggests

Table 7Comparison of SSA and assumed change in additive concentration for virgin, reclaimed,and finely ground virgin powders.

Sample SSA(m2/kg)

Additiveconcentration

Additive concentration-to-SSA

Virgin 187.92 Original OriginalReclaimed 226.13 Increased

(20.3% by SSA)Original

Finely ground virgin 219.07 Original Decreased (16.6% by SSA)

that the main factor causing the differences in flow propertiesbetween virgin and reclaimed powder is particle size distributionsince the additive concentration-to-SSA remained constant. It alsoconfirms that electrostatic spraying, humidity or other factors are notcontributing to the decrease in more static flow properties observedin the reclaimed powders. In contrast, RBER of the sieved virginpowder was improved when compared to the reclaimed powders. Itagrees with the conclusions from Section 3.2 that humidity exposureor other factors in the powder coating process may have greatereffects on more dynamic flow properties.

4. Conclusions

The purpose of this research was to investigate the differencesbetween virgin and reclaimed powders and to determine the mostinfluential factors that contribute to the decreased flow properties ofreclaimed powder coatings. Three potential factors were addressed,including:

1. Smaller particle size of reclaimed powders when compared to theirvirgin counterparts.

2. Exposure of reclaimed powders to moisture in the spray booth.3. Differences in additive concentration (nanoparticles) between

reclaimed and virgin powders.

Through this study, it was found that the particle size and additiveconcentration relative to the specific surface area of the powder had alarge effect on flow properties. Reducing particle sizewill dramaticallyincrease powder cohesion due to increased interparticle forces. Andreducing additive concentrations with respect to the specific surfacearea will weaken the improvement of powder flow performances dueto less additive adhering to the powder particle surfaces. On the otherhand, humidity and other factors of the powder coating process hadless influence. In general, the reduction in particle size that is commonbetween virgin and reclaimed powders was confirmed to be a majorcontributor to the decreased performance seen in reclaimed powders.It was also found that the concentration of the nanometer size flowadditives was increased in the reclaimed powder; however, the SSA ofthe reclaimed powder was increased due to smaller particle sizes. As aresult, the ratio of additive concentration-to-SSAwas kept the same asthe virgin powder. Therefore, the change in additive concentration ofreclaimed powders is not expected to have obvious effect on therecyclability. On the other hand, a process such as grinding, which willincrease the specific surface area without a corresponding increase ofthe additive concentration, can significantly affect powder flowproperties.

References

[1] J.C. Kenny, T. Ueno, K. Tsutsui, Analytical approach for high quality appearancepowder coatings, Journal of Coatings Technology 68 (1996) 855.

[2] Kittle, K.J. and Rushman, P. F., 1997, Powder coating compositions and their use,U.S.Patent No. 5,635,548.

[3] Gregory J. Bocchi, Join with PCI to improve powder coating community, ProductsFinishing 70 (6) (2006) 73–74.

[4] K. Lis, General Motors Corporation, company report, 2008 February 14, 2008.[5] K. Lis, General Motors Corporation, company report, 2009 February 20, 2009.[6] Xiangbo Meng, Zhang Hui, Zhu Jingxu, Characterization of particle size evolution

of the deposited layer during electrostatic powder coating process, PowderTechnology 195 (3) (Nov. 10 2009) 264–270.

[7] J. Visser, An invited review — Van der Waals and other cohesive forces affectingpowder fluidization, Powder Technology 58 (1989) 1–10.

[8] Zhu, J. and Zhang, H., 2004, Fluidization additives to fine powders, U.S. Patent6,833,185.

[9] Zhu, J. and Zhang, H., 2006, Powder blending methods for uniformly dispensingadditive particles to fine powders, U.S. Patent, filed.

[10] M.J. Krantz, H. Zhang, J. Zhu, Characterization of powder flow: static and dynamictesting, Powder Technology 194 (2009) 239–245.

[11] Qing Huang, Hui Zhang, Jesse Zhu, Flow properties of fine powders in powdercoating, Particuology 8 (1) (Feb. 2010) 19–27.

45J. Fu et al. / Powder Technology 211 (2011) 38–45

[12] Qing Huang, Hui Zhang, Jesse Zhu, Onset of an innovative gasless fluidized bed-comparative study on the fluidization of fine powders in a rotating drum and atraditional fluidized bed, Chemical Engineering Science 65 (3) (2010) 1261–1273.

[13] “D 5630–06 standard test method for ash content in plastics”, American Societyfor Testing and Materials (2006).

[14] D. Geldart, The effect of particle size and size distribution on the behavior of gasfluidized beds, Powder Technology 6 (1972) 201–209.

[15] D. Geldart, Types of gas fluidization, Powder Technology 7 (1973) 285–297.[16] Y. Chen, et al., Fluidization of coated group C particles, AIChE Journal 54 (1) (2008)

104–121.