INDUSTRIAL MANUFACTURING BLOG YOU SHOULD BE READING

Effects of highly absorbing pigments on near infrared cured polyester/melamine coil coatings

07 Apr.,2023

ELSEVIER

Ian Mabbett3'*,Jon Elvinsb, Cathren Gowenlocka, Paul Jones b, David Worsleya

a SPECIFIC, College of Engineering, Swansea University, Baglan Bay Innovation Centre, Port Talbot SA12 7AX, United Kingdom b Tata Steel, SPECIFIC, Baglan Bay Innovation Centre, Port Talbot SA12 7AX, United Kingdom

ARTICLE INFO ABSTRACT

In order to expand available colour range for an industrial coil coating line a range of 25 |m polyester melamine coatings were applied to galvanised steel substrate and rapidly cured using near infrared (NIR) radiative curing. The purpose was to improve understanding of this relatively new curing technology and identify any problems associated with differing absorption of a range of coloured coatings. It has been suggested that in order to increase efficiency of NIR cure, NIR absorbers should be added to the coating formulation. UV/Vis/NIR spectroscopy was used to deduce the parts that coating and substrate absorption play in topcoat cure and lab scale trials were run on coatings throughout the colour range with their heating profiles and surface finish being recorded and assessed. The results showed that in this particular application having a coating that absorbs too strongly in the NIR region can actually result in solvent boil defects due to cross linking and film formation occurring prior to solvent removal.

Contents lists available at SciVerse ScienceDirect

Progress in Organic Coatings

journal homepage www.elsevier.com/locate/porgcoat

Article history:

Received 18 January 2013

Received in revised form 21 March 2013

Accepted 27 March 2013

Available online 28 April 2013

Keywords: NIR

Near-infrared

Coil-coating

Fast-cure

Polyester/melamine

1. Introduction

Industrial coil coating can be defined as a continuous, highspeed process for depositing a coating onto a metallic substrate, frequently galvanised steel. Traditionally these coatings have been cured with large gas fired convection ovens at around 100-120 m min-1. The higher the line speed the more material can be produced annually. Fast radiative cure techniques allow higher line speeds up to 150-180 mmin-1 [1] which can allow direct integration with a galvanising line. Other advantages include a smaller oven footprint and a more energy efficient cure per square metre or material produced resulting in lower costs and a reduced carbon footprint [2].

In the construction sector, 20-25 |im polyester coatings on galvanised steel are commonly used in roofing and cladding applications [3]. Polyester (PE) resins are thermosetting resins commonly cross-linked with hexamethoxymethylmelamine (HMMM) [1,2]. The 30-45 s dwell time in a gas fired convection oven allows time for solvents to be removed as sample heats up to a peak metal temperature (PMT) of 216-232 °C so that the solvent is fully removed before acid catalysts are unblocked and promote cross linking to give full cure. 45s at 120mmin-1 would result in an oven length of 90 m and any increase in line speed would require

* Corresponding author. Tel.: +44 01792606601. E-mail address: i.mabbett@swansea.ac.uk (I. Mabbett).

an even larger oven and a considerable increase in gas consumption [4,5].

Near infrared (NIR) can be used to cure polyesters in <10 s under the lamps with reportedly little modification to convection formulations [6,7]. NIR is between the visible and infrared regions of the electromagnetic spectrum. NIR is a radiative curing technique but unlike UV cure it is essentially a fast thermal method of cure. Radiative curing techniques offer energy efficiency advantages because the heating is produced by direct absorption by the material itself meaning there is no need to preheat an entire oven and changes in settings can feed back almost instantly on panel temperature. Over the range of 250-2500 nm the NIR lamps used in this investigation give an output of 250 kWm-2 with the majority of the energy focussed between 800-1200 nm. It is important to note however that the tungsten halogen technology does have a broad emission curve as predicted by black body emission theories and as such there will be emission throughout the visible range as well as the NIR region, where the emission peaks. Theory suggests [8] that NIR radiation is able to penetrate deeper into an organic coating than mid range infrared and heats the coating uniformly without wasting energy heating the substrate [9-11]. In fact, in earlier research, the authors suggest that heating the substrate should be avoided all together and to focus on heating the topcoat itself to improve energy efficiency. However, more recent research [12-14] suggests that substrate absorption can be significant and is one of the reasons that at fast line speeds it is often easier to cure coatings with low absorption than those with higher absorption of the incident radiation. At slower line speeds or for different coating systems there are

Table 1

Range of substrate with absorption investigated.

Substrate Galvanised Surface Composition

Galvalloy/Galfan 95.2wt% Zn, 4.8 wt% Al

Magizinc 96.8 wt% Zn, 1.6 wt% Mg, 1.6 wt% Al

HDG 99.8 wt% Zn

Zalutite/Galvalume 55 wt% Al, 43.5 wt% Zn, 1.5 wt% Si

Iron N/A (just bare iron)

indeed merits for increasing topcoat absorption to increase energy efficiency but this is not always true in all systems. In order to cure a cross-linking coating in a vastly reduced time there needs to be a modified cure profile allowing for a two stage cure, one oven zone heating the material to a temperature at which solvents evaporate, then a small dwell time to allow for their complete removal, followed by a second heating zone to initiate cross linking.

In this study the absorption characteristics of 25 |im polyester coated galvanised steel was investigated to establish what effect the coating absorption has on cure using NIR radiation. On a combined galvanising and paint line in industry it has been noticed that colours that absorb less incident radiation are easier to cure than the more absorbing coatings which is contrary to the theory in the literature at the time of commissioning the line. Whilst colour itself is not an indicator of NIR absorption or reflectance the operators of these production lines may be making these observations based on two principles; the absorption of the visible light emitted at high intensity from the NIR lamps and also as they are dealing with organic coated steel for exterior construction applications there is an appetite against changing to higher cost NIR transparent tinting pigments which may not be proven in weathering tests in these formulations so they are likely to be using carbon black to tint. Carbon black is a broad wavelength range absorber and certainly absorbs strongly in the wavelengths emitted by tungsten halogen NIR lamps. Through spectroscopic studies and by examining heat profiles and surface finish of lab samples cured with an AdPhos coil lab unit it is proposed that the substrate absorption plays a more important part than initially expected and is advantageous in the cure of less absorbing samples. Moreover the addition of NIR absorbing pigments has previously been proposed [8] as a method to increase energy efficiency of NIR cure but this work suggests that this could increase risk of solvent boil and poor surface finish. However, better understanding of how the NIR radiation interacts with coating and substrates enables relevant modifications for formulations and cure profiles to improve this situation for a colour range.

2. Experimental

2.1. Substrate

Steel can be protected from corrosion by galvanising with zinc rich coating alloys. The two types of galvanised coating of interest to this study were hot dip galvanised steel (HDG) and Magizinc, since these are the commercial choices for exterior polyester products for roofing and cladding. However zinc has a pronounced absorption peak in the NIR range so other substrates were evaluated to show how this zinc peak is altered with each type of galvanised surface. The composition of each of these surfaces is given in Table 1.

2.2. Paint formulation

Polyester melamine paints were formulated at the BASF coatings labs in Deeside. A range of colours were produced using near commercial formulations modified slightly for NIR fast cure applications. Each of the colours has a different NIR absorption. All

pigment were milled to below 15 | m measured using a Hegman gauge. Paints were specified to be between 40 and 50% gloss at 60° and viscosity checked to take 65-70 s to drain a DINN 4 flow cup at 21 °C. White and silver metallic paints are particularly important commercially Table 2 shows the formulation of these paints, but due to their proprietary nature detail on solvent blends and actual pigments used must be withheld.

2.3. Spectroscopy

UV/Vis/NIR total reflectance spectra were recorded for a range of metallic substrate using a Perkin Elmer Lambda 750S high performance spectrophotometer. Total reflectance combines specular and diffuse components of sample reflectance. Since metallic substrate is totally opaque what is not collected in total reflectance can be assumed to be absorbed. Clean metallic surfaces tend to have low emissivities meaning the radiation that does get absorbed is retained strongly heating the substrate efficiently under irradiance with NIR curing lamps.

The paints were cured into 20 |im free films to allow UV/Vis/NIR spectra to be measured in both reflectance and transmission. Total reflectances of each of the paints cured on galvanised steel were also recorded. This allows for a more thorough examination of the major contributors to the absorption, heating and thus curing of the coatings for each of the colours in the full paint system.

The industry tends to describe cure in terms of resistance to solvents. In particular it is common to assess cure by the number of double rubs with 2-butanone (methyl ethyl ketone or MEK) the coating can endure before substrate becomes visible. MEK testing of this kind is somewhat subjective but ASTM D4752 describes controls that can be put in place such as the type of cloth used, etc. to make it less so. It is also possible to use Fourier transform infrared spectroscopy to assess cure of polyester melamine paint systems [15]. A Perkin Elmer spectrum 100 FTIR spectrophotometer was used to record FTIR spectra of each sample, utilising a diamond universal attenuated total reflectance accessory (UATR). The area of a reference peak associated to 'ester' carbonyls on the polyester backbone can be compared to a candidate peak from in-plane deformation vibration of the triazine ring of the HMMM cross linker. A more detailed look at the fine structure of all the spectra compiled for the polyesters in this study identified those peaks at around 1722cm-1 and 1545cm-1, respectively. Calibration curves were also drawn up comparing these ratios to samples cured in a conventional oven for a given period of time. Where samples are referred to as cured this means that the ratio of these peaks is equivalent to a conventionally cured sample having reached a peak metal temperature (PMT) of 230 ° C or greater and having survived 100 double rubs with 2-butanone.

2.4. Convection and NIR cure

To carry out cure studies using conventional heating or NIR, panels of substrate were bar coated with wire wound K-bars or Meyer bars as described in ASTM D4147. Panels of 200 mm x 100 mm x 0.49 mm with 5 |im clear chrome free pre-treatment primers were coated with 60 | m wet films to give dry film thicknesses of 20 | m of polyester topcoat. These panels had cure profiles measured with k-type thermocouples (spot welded to the centre of the back side of the sample) and an omega TC-08 data logger and the same set of panels were used for spectroscopic studies described above. Convection cure was carried out using a Mathis convection oven set to 450 °C. Dwell times of40-45 s achieve PMT's of 220-230 °C at this sample size. This is well within the polyester specification of 216-232°C for full cure at 20 |im.

Lab scale NIR curing was carried out using an AdPhos coil lab V2. The AdPhos equipment consists of a set of six tungsten halogen

Table 2

Polyester topcoats produced in a range of colours.

Colour

Polyester resin/wt%

Solvent/wt%

TiO2/wt%

Carbon black/wt%

Other pigments/wt%

Other additives/wt%

Silver

56.44 48.28

24.87 25.44 26.46 21.06 22.23 18.49 23.85

0.77 0.71 1.25

4.1 0.11

0.99 0.21

13.47 10.33 14.64

15.38 22.41 19.79 21.04 19.67 25.07 20.45

lamps in a module with an output equivalent of 250 kWm-2 and a sample holder to carry wet coated substrate under the lamp module. Fast cure of polyesters requires two distinct cure stages. First the sample must be heated to a temperature where solvents are removed. Once the volatile solvents have been removed the coating must be heated to a cross linking temperature. If cross linking occurs before the volatiles are fully removed, the film will form too early and the 'escaping' solvents will cause solvent boil blisters on the surface. In NIR cure there is less time for this to occur discreetly and on the coil lab there is only one lamp module. However the lab unit can achieve a two stage cure by passing the samples through the oven once to remove the solvent and achieve a cross linking PMT on the return journey.

3. Results and discussion

3.1. Spectroscopic studies

The UV/Vis/NIR total reflectance spectra of a range of metallic substrates (Fig. 1) showed that different galvanising alloys can result in differing absorption characteristics. Whilst all the samples appear highly reflective there is still significant absorption. Surface cleanliness and roughness will affect absorption and gauge and emissivity of the substrate will also influence how effectively it heats up under NIR irradiation. Of particular interest in the reflectance spectra is the peak associated with Zn at around 1000 nm, this absorption is in the region of the maximum emission from the NIR source and would suggest enhanced heating of galvanised coatings that are exposed to the lamps. For example, in the case of Magizinc vs. HDG (both 0.49 mm gauge) average total reflectance over the whole spectrum are 72% and 88%, respectively (but with reflectance at ^max of the output spectra of the Adphos emitters of 46% and 57%, respectively) and this resulted in the Magizinc reaching a PMT of 208°C vs. 195 °C reached by the HDG sample when irradiated at 12 m min-1 with 90% power on first pass through oven and 60% power on second pass.

Hot dip galvanised steel (HDG) was coated with a range of colours of polyestertopcoat and the total reflectance was measured in the UV/Vis/NIR regions as shown in Fig. 2. These spectra show as we would expect that the white polyester reflects the most incident light and the black polyester absorbs the most in the visible region, but in the case of these near commercial paints this trend extends into the NIR as well, most likely due to differing amounts of carbon black as shown in Table 2. All the other colours are in between these two extremes. Closer examination of the spectra reveals an absorption peak in the white polyester spectrum that matches the effect of Zn on the HDG spectra in Fig. 1. This could suggest that some radiation makes it through the topcoat and interacts with the substrate. TiO2 pigment in the white polyester gives the coating its colour by reflecting well in the visible region 380-780 nm but becomes more transparent in the NIR region. The silver polyester shows a peak nearer 800 nm characteristic of the aluminium flake used to achieve its metallic finish.

Once it was established that substrate absorption and heating could be significant 20 | m polyester free films were formed for each colour by coating onto PTFE coated steel and curing in a convection oven. The transmission UV/Vis/NIR spectra of these coatings was then recorded in order to determine any transparency because if some of the light emitted by the NIR source makes it through the topcoat as suggested in Fig. 2 then substrate absorption play a significant part in the cure of the paint and should be examined further. The spectra in Fig. 3 do in fact show that the white polyester does allow some light through to the substrate, particularly in the NIR region. Although silver shows around 50% reflectance the other 50% is very efficiently absorbed by the coating with no light getting through to the substrate. Black polyesters also absorb all incident radiation in the coating itself and the noise in this spectrum shows how difficult it can be to attenuate against such a strong absorption.

Experience has shown that commercial black and silver polyester coatings on galvanised steel are more difficult to cure than white coatings. All the other colours tend to fall in between these two extremes in terms of ease of achieving a good cure with

Wavelength / nm

Fig. 1. UV/Vis/NIR total reflectance spectra of a range of metallic substrates.

Wavelength / nm

Fig. 2. UV/Vis/NIR reflectance curves of some polyesters of varying colour.

100! 90 80 -70 -60 -S 5040 -30 -20 -10 -

------

1400 Wavelength / nm

Fig. 3. Vis/NIR transmission through free films of clear, white, black and silver polyesters (400-2500 nm to avoid PE UV absorptions).

a satisfactory surface finish and a reasonable window of operating parameters to ensure consistent quality. Black and silver coatings are more prone to solvent boil, where the topcoat has fully cured prior to full removal of solvents, trapping them and resulting in blisters. The spectra show that the more difficult colours to cure tend to absorb all incident radiation in the topcoat itself whereas the more successfully cured coatings allow some radiation to pass through to substrate, interact there raising substrate temperature, then pass back through the topcoat. This suggests that some level of transparency of the topcoat to NIR radiation is a good thing, allowing the metal substrate to heat, driving solvents upwards and preventing the early film formation associated with total absorption occurring in the topcoat (potentially within the first few | m for black and silver coatings).

3.2. Cure profiles

Since there is a difference in absorption of NIR by these particular coatings of different colour the power settings must be altered to allow different colours to reach an appropriate peak metal temperature (PMT) to cure in a given time. PMT is actually not a very good indication of the temperature the coating reaches in fast cure operations because there is not as much time for coating and metal substrate to equilibrate which is very clear in the case of highly NIR absorbing samples which appear to cure at lower PMTs than expected and burn if they reach the target PMT. There are additional problems when measuring temperature with thermocouples because the thermocouples themselves may absorb NIR and heat

Fig. 5. Temperature profiles of a white and black polyester heated using the same NIR settings, showing the increase in PMT reached by the black, compared to the white coating.

up. Fig. 4 highlights the difference in power settings required for the white coating vs. the black coating to reach PMTs in a similar range (218 °C for white and 228 °C for black). Here again the two passes are used to replicate the commercial oven where the two zones are used to control firstly solvent loss and then in the second pass cross linking. Furthermore the black sample showed significant solvent boil in these conditions and the sample would therefore not pass a quality test (as can be seen in Fig. 6).

If the line speed and power of the NIR unit are kept constant but the NIR absorption of the topcoat is changed for a particular substrate then the samples will reach different PMTs. The near commercial paints used in this investigation have varying amounts of carbon black and other pigments that absorb strongly in the regions of intense emission from the AdPhos lamps and thus in this practical case coatings of different colour appear to reach different PMTs. The temperature profile of NIR cured white and black 20 |im polyester topcoats on 200 x 100 x 0.49 mm HDG substrate coated with 5 |im pre-treatment primer are given in Fig. 5. It is clear from this data the importance of the NIR absorbance of the coating and its heating rate.

In order to produce the full colour range, all have different absorption characteristics, and therefore the range of settings through which each can be successfully cured varies widely. The wider the range of parameters over which each coating can be cured the better as this should facilitate its introduction to full scale roll to roll production with less challenges Current

.* ■ + i,» • .

I Power on first pass Power on second pass

- ' - : • - ' • • •

\ gj. t . »

È. \ • ^

, '(¡jf - V.

*';•*' ' • y'.

-»v. - . : f •

White Polyester (PMT 218 °C) Black Polyester (PMT 228 °C)

Fig. 4. The difference in power settings required for white and black polyesters to reach equivalent PMTs.

Fig. 6. Solvent boil defects in black polyester coating (left hand side of image).

Table 3

How well each coating cures and its robustness to process variability, using a curability scale from 1 to 5 where 5 is good process window and 1 means the sample does not cure well at all.

Colour

Reflectance at 920 nm (gmax emission at 100% power on AdPhos lamp)/%

Range of line speeds with successful cure/m min-1

Range of power settings with successful cure/% per lamp run

Curability

Silver

69 18 16 7 4 44

9-15 9-15 9-15 9-15 12 12

5 2 2 1 0

commercial white polyesters have the largest window of opportunity in terms of altering parameters with no detrimental effects to the finish and black and silver polyesters were the least robust to altering settings. This is because the white has very little absorbance in the NIR. These blacks use carbon black as their pigment and this absorbs strongly in the NIR. Alternative black pigments with lower NIR absorption such as paliogen black for example have been tried during this investigation, but the visible component of the NIR lamps emission still heated them rapidly and the increase in cost and lack of corrosion and weathering data in this particular context means that their adoption in products guaranteed for exterior roofing and cladding material is unlikely. Likewise for silver metallics, the aluminium flake absorbs NIR strongly and its low emissivity means it retains that heat well. Alternatives with lower absorbances and higher emissivities such as mica just do not give the same appearance, which is noticeable in such a popular product. The other colours tested fell within these two extremes of NIR absorption; the actual orders given in Table 3; where 5 means the sample cures well with good process window, 4 means good cure achieved with smaller cure window, 3 suggests cure achieved but with little margin for error, small cure window meaning any setting changes are problematic. 2 means cure is only achieved with imperfections on the topcoat and 1 just does not cure well at all.

In the field of solar thermal heating the importance of absorbance (and therefore colour) is understood. In that context a solar reflective index (SRI) (ASTM 1980) is calculated which enables prediction of steady state temperatures reached due to insolation. In simple terms, ASTM 1980 explains that for a surface that is exposed to the sun, with a conductivity of zero, that surface will reach a steady state temperature that is described by Eq. (1).

al = sa(T4 - T4y) + hc(Ts - Ta)

where a, solar absorbance = 1 - total solar reflectance; l, solar flux (Wm-2); e, thermal emissivity; a, Stefan Boltzmann constant (5.66961 x 10-8 Wm-2 K-4); Ts, steady state surface temperature (K); Tsky, sky temperature (or ambient away from surface) (K); hc, convective coefficient (Wm-2 K-1); Ta, air temperature (or ambient near surface) (K).

So for a known emissivity and solar reflectance Eq. (1) can be solved iteratively at different values of convection coefficient to work out the steady state temperature.

In practice in the context of SRI calculations the alternative approach in Eq. (2) is used and the solar reflective index is defined by Eq. (3).

1066.07a-31.98e 890.94a2 + 2453.86ae Ts — 309.7 +--n i u---_ __-——- (2)

6.78s+ hc

(6.78 + hc )2

SRI = 100

Tb -Ts Tw _ Ts

where Tb and Tw are steady state temperatures reached by black and white surfaces, respectively. SRI is usually used under standard

solar and ambient conditions so that Eq. (3) can be reduced to Eq. (4).

SRI = 123.97 - 141.35/ + 9.6555x2

where x — ((a - 0.029e)(8.797 + hc))/(9.52053e + hc).

AdPhos supply a spread sheet that calculates filament temperature at different power settings for their lamps (Eq. (5)). If this is used in conjunction with Planck's law for spectral irradiance (Eq. (6)) and the emissivity of tungsten in the filament is applied then a theoretical black body emission curve can be produced for each power setting.

where T, colour temperature at operation voltage (K); U, operation emitter voltage = % power on control panel x 380 V; U0 = nominal emitter voltage = 400 V; T0 = colour temperature at nominal voltage =3200 K.

R(g) =iehc/kXT - 1

where wavelength and this can be solved for the range 300-2500 nm for this application; c, speed of light (3 x 108 ms-1); h = Planck's constant (1.38 x 10-23JK-1); k = Boltzmann constant (4.1 x 10-15 eVs-1); T = temperature of the emitter or for non perfect black bodies can be eT with e representing emissivity. In this case T can be taken to mean T from Eq. (5) multiplied by the emis-sivity of the tungsten emitter.

Alternatively emission at each power setting can be estimated using spectrometers such as the Ocean Optics HR2000+ coupled with an NIRQuest to give empirical measurement covering the required range of wavelengths.

These emission curves could be used to replace the AM1.5 solar weightings described in ASTM E903 for total solar reflectance (TSR) calculation, which is subsequently entered into the SRI calculations. 5 nm resolution of the spectra is used and essentially the lamp emission at each wavelength in Wm-2 | m-1 is multiplied by 5 to cover the 5 nm resolution and this figure is then multiplied by the reflectance of the surface at that wavelength, expressed as a decimal. This would then weight the total reflectance of the sample surface at each wavelength to the available irradiance from the emission of the AdPhos lamps. Replacing the Solar Absorbance (1-TSR) term in Eqs. (1)-(4) with this new value would allow the predicition of the surface temperature reached if a totally non-conductive material was left under NIR irradiation until a steady state environment was set up. Within the SRI calculations emissiv-ity of the sample is required. This can be measured using a 'Devices and Services AE1 emissometer'. There is also a term for solar flux or total insolation, in this case that would be related to the emission of AdPhos lamps and would either require application of the Stefan-Boltzmann law for radiant intensity (Eq. (7)) or the integration of

Fig. 7. Different ways that the NIR can interact with the paint system and its metallic substrate.

calculated or measured spectra over a known area of irradiance (Eq. (8)).

1(T) = oT4 for a black body or 1(T) = eoT4 for a real material

1(T) = or in this case

2500 nm

1(T) =

J 300 nm

There is also a convective coefficient hc in the SRI calculations, usually to estimate the effect of heat transfer by wind on a solar thermal heater, which could be replaced with a term to describe the effect of air flow within the lamp zone's extraction system. To work properly in this new context the SRI calculations need to be modified to allow for time dependency (differentiating Ts with respect to time) since typical times in NIR lamps will not allow equilibration to steady state conditions and the sample is also not totally insulating so substrate size and gauge and coating thickness needs to be taken into account since their thermal mass will affect the temperature reached. Finally there is a further complication created by the two passes through the lamp zone in lab unit cured polyesters. These complicated interactions of all of the factors discussed make temperature predictions difficult but work is on going to find ways of predicting PMT for given process conditions using just sample reflectance, emissivity and physical dimensions.

3.3. Absorption of NIR radiation by each part of the pre-painted galvanised steel system

Coatings that cure easily with NIR tend to have some transparency in the topcoat allowing the radiation to pass through to the substrate and heat from the bottom of the paint coating upwards. There are a number of ways that the incident radiation can interact with the system and in reality each mechanism of absorption or reflectance will occur to some extent. Fig. 7 shows all the ways that incident radiation can interact with a coating system. For a coating that behaves similarly to a white polyester all of these mechanisms will contribute to the cure.

In this case A refers to coating absorbing, A' is where coating absorbs but with better penetration, B involves radiation making its way through the coating and being absorbed by the substrate, C radiation makes it through coating and is reflected back by substrate and is absorbed in coating on way back through, D radiation makes it through coating and is reflected back by substrate and escapes through coating, E radiation is reflected at surface of paint. All the same interactions would be possible with pre-treatment/primer and any other layers. This example only deals with irradiation from the top surface.

Darker, more absorbing coatings tend to behave more like black polyesters. Looking in more depth at the possible interactions of the NIR radiation with the system it can be seen that the major

Black coating

Substrate

Fig. 8. How NIR interacts with a highly absorbing coating.

contributor in this case is absorption by the topcoat (Fig. 8), close to the surface, as suggested by spectroscopic study of each layer and the system as a whole. The question marks in Fig. 8 refer to there being a possibility at very high loadings of a highly absorbing pigment that there will be little significant reflectance from the surface and that the NIR may not penetrate past the top few microns of the surface. This is why moderate content of highly absorbing pigments such as carbon black or aluminium flake can in fact make it very difficult to successfully cure a 25 |im polyester coating on a galvanised steel substrate at a speed that is typical for roll to roll production on a coil coating line.

This increased understanding of how NIR interacts with coatings and substrate commonly used by coil coating industries explains many of the defects experienced on production lines when colour (and in practical terms this change often affects the NIR absorption) is varied. Using this knowledge temperature profiles in industrial processes can be tailored to the product required. Various ways of decreasing the NIR absorption of the topcoat and increasing the rapid heating of the substrate-coating interface have also been suggested and subsequently tested [16]. These include modified solvent blends and slower catalysts, use of NIR transparent pigments to replace carbon black for tinting or creating a metallic effect without using aluminium flake. Increasing absorption of substrate and modifying the cure profiles and settings have also been investigated. To avoid solvent boil the best approach is to ensure the topcoat has some transparency to the NIR and the substrate absorbs with some efficiency. NIR can then be used to heat the system in a two stage cure profile, first to a temperature where solvent is removed, then after a small time for the continuation of this evaporation a second heat zone is utilised to bring the coating to a temperature where blocked acid catalysts are unblocked so rapid cross-linking can occur. All of this must happen in less than 10s and leave a fully cured surface with no solvent boil or other defects.

4. Conclusions

NIR is becoming a popular option for rapid cure of coatings in the coil coating industry particularly where fast line speeds are required or where there is limited space for the large footprint of a convection oven. NIR technology has the potential to reduce the cure time of a 20 | m polyester coating on a galvanised steel substrate from 30 to 45 s via conventional heating methods down to <10 s under the lamps. This offers obvious advantages in terms of increased production and removing bottlenecks from a continuous coil coating process. It is suggested that there are significant energy efficiency benefits too since with all radiative curing techniques the absorption takes place in the material itself so there is no need to preheat large ovens and there is an instant feedback if any parameters need to be changed in the oven. The use of absorbing pigments to boost energy efficiency has been suggested by some groups for certain products, however this may not work in all cases since too much absorption in the top few microns of a cross-linking polymer may initiate film formation prior to full removal of any solvents.

Spectroscopic studies of coatings and substrate are the first port of call in understanding how each part of the system is likely to heat up when irradiated. Thermo gravimetric analysis (TGA) can be useful to understand what temperatures solvents are likely to be removed at and also if there is limited knowledge of the formulation then TGA can ascertain the percentage of solvents on the mix. Lab based cure trials collecting cure profiles and surface finish details can deduce the best routes for successful NIR cure in each case. When dealing with industrial coil coatings for exterior applications it is likely that they will be tinted with carbon black, which has a broad absorption and is a strong absorber in the NIR region. In these cases the colour of a coating will give a good hint as to its NIR absorption characteristics and therefore its cure performance but these are anecdotal and no replacement for full characterisation of a coating's absorption at the full range of wavelengths emitted by the curing lamps combined with lab scale cure trials prior to attempting to cure a particular formulation on a full scale production line. Whilst energy efficiency can be improved by adding highly absorbing pigments, a true understanding of the fundamentals of each system is required first because in some cases such as rapid cure of industrial coil coatings too much absorption can lead to solvent boil and unacceptable surface finish. The ideal situation in this case is to have a topcoat which is slightly transparent to NIR and an absorbing substrate to heat the coating from the bottom up in a two stage process which separates solvent removal from cross linking and film formation.

Acknowledgements

The authors would like to thank EPSRC and WEFO via the ESF for funding and Tata Steel Colours and BASF Coatings Ltd. for materials and expertise. Particular thanks To Rob Ireson of Tata and Dr Nick Brown, Paul Davies, Graham Swanston and Steve Corkish from BASF

References

[1] K.K.O. Bär, How to save energy costs with the NIR-technology in coil coating processes, Steel Grips 4 (2006) 225-228.

[2] Bär KKO, Clean technologies in coil coating - present and future advances, in: 45th ECCA Autumn Congress, 2011.

[3] K.J. Park, K.M. Lee, Kyung jae park, kang mo lee, and young ho lee: NIR drying technology at dongbu steel's continuous coil coating line, Polymer 5 (2007) 4-6.

[4] W. Zhang, R. Smith, C. Lowe, Confocal Raman microscopy study of the melamine distribution in polyester-melamine coil coating, J. Coat. Technol. Res. 6 (2008) 315-328.

[5] W.R. Zhang, T.T. Zhu, R. Smith, C. Lowe, An investigation on the melamine self-condensation in polyester/melamine organic coating, Prog. Org. Coat. 69 (2010) 376-383.

[6] K.K.O. Bär, J. Anderson, N. Frederiksen, A.S. Gmbh, B. Str, New opportunities for inline paint applications in galvanizing lines, in: AIST Conference, May, 2004.

[7] K.K.O. Bär, NIR technology - new opportunities for metal processing and metal coating lines, in: Metallurgical Conference Riva Del Garda, 2004.

[8] R. Knischka, U. Lehmann, U. Stadler, M. Mamak, J. Benkhoff, Novel approaches in NIR curing technology, Prog. Org. Coat. 64 (2009) 171-174.

[9] M. Stuhldreher, Near-infrared curing: what it is; what it is not, Powder Coat. 13 (2002) 25-29.

[10] K.K.O. Bar, NIR-technology - novel potential use in coil and coating lines, in: ESTAL Congress, 2003.

[11] K.K.O. Bär, J. Anderson, NIR-technology - an innovation in coil coating, in: NCCA Fall Meeting, 2003.

[12] T. Watson, I. Mabbett, H. Wang, L. Peter, D. Worsley, Ultrafast near infrared sintering of TiO2 layers on metal substrates for dye-sensitized solar cells, Library (2010).

[13] M. Cherrington, T.C. Claypole, D. Deganello, I. Mabbett, T.M. Watson, D. Worsley, Ultrafast near-infrared sintering ofa slot-die coated nano-silverconducting ink, J. Mater. Chem. (2011).

[14] A. Geitner, Bär KKO, Application potential of NIR-technology in coil processing lines, in: 3rd International Conference on Thermomechanical Processing of Steels, 2008.

[15] W.R. Zhang, C. Lowe, R. Smith, Depth profiling of coil coating using step-scan photoacoustic FTIR, Prog. Org. Coat. 65 (2009) 469-476.

[16] I. Mabbett, Applications of Near Infrared Heating of Interest to the Coil Coating Industry, Swansea University, 2011.

Previous: How Does a Water Well Drilling Rig Work?

Next: Fiber Laser Welding Explained: How's it Different?