Abstract
Curing conditions influence the properties of wood modified with melamine resin. Beech wood (Fagus sylvatica L.) was impregnated with melamine resin and cured in a dry and a hot steam process to investigate the influence of the relative humidity while curing. The topochemistry of the modified wood was assessed by light microscopy (LM) and scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM-EDX) to visualise structural and chemical changes on cell wall level. LM in combination with staining did not show differences between the curing processes but allowed to distinguish untreated and treated wood. SEM micrographs revealed a more severe impact of dry curing conditions on the structural integrity of the material than steam curing by the formation of micro cracks. Dry-cured specimens showed a higher resin concentration in the cell wall than steam-cured specimens with the same overall distribution pattern as shown by EDX line scans.
Keywords
Introduction
Wood in outdoor applications is exposed to moisture conditions that led to dimensional changes and fungal attack. Improvements in the mechanical and chemical properties through wood modification would not only expand the field of application but also create new markets for native wood as a renewable and sustainable resource (Hill 2006). Modification with thermosetting resins such as methylated melamine formaldehyde (MMF) resin is one of the established wood modification systems besides thermal modification, acetylation and furfurylation. MMF resins have a wide range of applications in the wood industry (Kohlmayr et al. 2014). They are used as adhesives, binder material, for finishing surfaces and as impregnation agents. If MMF was applied as a wood modification agent, decking and cladding made of modified wood would be located in a high-price market sector. Therefore, control of the modification process and the modified material is essential when homogenous products are demanded.
Process conditions such as time and temperature affect the properties of modified wood. High-temperature curing resulted in a more complete curing of the resin (Scheepers et al. 1993). Klüppel and Mai (2013) further discussed this matter and found that dry curing conditions lead to more complete curing than wet conditions (specimens wrapped in PET-foil to retain moisture). Wet conditions led to resin precipitation in the cell lumens and consequently higher resin content compared to the cell walls.
Microscopy techniques have been used by several authors to detect the changes in wooden materials after gluing or wood modification. Kielmann et al. (2014) used light microscopy (LM) and UV microspectrophotometry (UMSP) to visualise MMF deposits in cell lumens. Biziks et al. (2015) visualised the penetration depth of different molecular weight phenol formaldehyde (PF) resins through the inability of safranin to stain the cross sections of modified beech wood. This method would be very useful, if applicable to investigate the influence of curing processes on the penetration and stainability of beech modified with melamine resin.
Sernek et al. (1999) used brilliant sulphoflavine (BSF) and safranin staining to detect the urea-formaldehyde resin (UF) bondline in beech plywood. Mahrdt et al. (2015) detected the UF bondline and UF penetration by combined dyeing and fluorescence microscopy imaging. This technique was first established by Leemann and Ruch (1972). BSF was here used to quantify proteins in plant cells.
Numerous authors have used electronic imagery to verify the effects of modifications on the wood material. The most widely used methods were UMSP (Gindl et al. 2003; Mahnert et al. 2013), electron energy loss spectroscopy (EELS) (Rapp et al. 1999), and scanning electron microscopy with energy-disperse X-ray spectroscopy (SEM-EDX) (Rapp 1999).
UMSP and EELS require demanding sample preparation, whereas SEM-EDX only requires small smoothly cut wooden blocks. The SEM-EDX technique is particularly emphasised as the specimen preparation for SEM analysis is rather simple and the EDX verification of nitrogen is a reliable method to locate melamine resin in cell walls.
The aim of this study was to analyse the influence of curing conditions on melamine-treated wood through light and electron microscopy. In particular, there were three topics: the impact of the curing conditions on the staining of specimens in LM imaging, the micro structural changes of the melamine modified cell wall matrix through SEM and the resin distribution across the wooden matrix and the cell wall layers as analysed by EDX.
Material and methods
Material
Specimen sizes for LM and SEM-EDX.
Methods
Impregnation and curing
Oven-dry specimens were impregnated (vacuum 0.01 MPa for 1 h; diffusion phase at atmospheric pressure for 2.5 h) with MMF resin (solid content 19%) and 1% Triethanolamine as buffer. The curing took place in a laboratory oven having the ability to control the temperature as well as the humidity (XVC305 UNOX S.p.A., Padova Italy). Two curing processes were set up, each lasted 24 h at 120°C and differed in humidity: 0 and 100% steam (0 and 52% RH; saturated steam atmosphere) were used. The specimens were allowed to dry at room temperature for 24 h prior to curing.
Solution uptake (SU; 1) and weight per cent gain (WPG; 2) were recorded to ensure a uniform treatment. The WPG of the specimens for leaching was calculated theoretically in order to prevent post curing when recording the oven dry weight: Separate specimens were used for the determination of the WPG of the leaching group. The ratio of SU and WPG (RWPG/SU; 3) was used to calculate the WPG where ist could not be measured directly.
SU = Solution uptake [%] WPG = Weight per cent gain [%] M0 = Oven dry mass before impregnation [g] Mi = Mass after impregnation [g] M1 = Oven dry mass before impregnation [g] M2 = Oven dry mass after curing [g] RWPG/SU = Ratio of WPG divided by SU
Light microscopy
The specimens for LM were prepared from small specimens as mentioned above. Sections of 20 µm thickness were prepared using a sliding microtome with disposable blades. One series of sections of all treatment groups was stained in a safranin solution (0.5%) for 10 min. Another series of sections was stained with BSF, washed out for two weeks in demineralised water, rinsed in 50 and 96% ethanol/demineralised water mixture (1 h each) and finally embedded in Euparal, dried at 60°C overnight and fixed on microscopic slides.
Scanning electron microscopy and energy-dispersive X-ray spectroscopy
The samples for SEM-EDX were prepared using fine hand tools and a sliding microtome with disposable blades. The specimens were placed on sample holders and carbon sputter-coated. SEM micrographs were taken with an EVO LS 15 (Carl Zeiss Microscopy GmbH Jena Germany), 8.5 mm working distance, 10 kV acceleration voltage and 430 pA spot size. The EDX images were recorded using an X-MAX 50 mm² detector (Oxford Instruments GmbH, Wiesbaden Germany) in combination with the AzTecEnergy program, a recording time of 300 s, and a scan size of 1024 px. Line scans were placed to span the distance across two cell walls. The recording time was 300 s. Nitrogen and carbon data were recorded and used for the analysis of the nitrogen distribution. Comparison and thus quantification of elemental distribution on wood is challenging as the surface is rough. In order to improve and compare the data of several line scans and specimens, the nitrogen data were normalised using the carbon data. Under the supposition that the carbon content is uniform throughout the cell wall layers (Blazej 1979), the recorded carbon data (counts per second; cps) of an even, horizontal area were defined as the normalisation constant. Then the nitrogen cps of every spot of the line were normalised over the mean carbon cps of that defined area (Equation (4)).
Ncps
n
= Normalised nitrogen count per second Ncps
x
= Nitrogen count per second Ccps
x
= Carbon count per second Ccps
N
= Normalisation constant; carbon count per second
The simple moving average (20 SMA; Microsoft Excel 2016) was used to facilitate the line scan graphs of the nitrogen distribution across the cell wall.
Results and discussion
Impregnation and curing
Specimens with uniform WPGs (±10% of group mean value) were chosen for the analysis. The average WPG varied between 13.4 and 17.2% with an overall average of 15.4%.
Light microscopy
Transverse sections of the specimens treated with MMF resin and stained with safranin and BSF are shown in Figure 1.
Top row: light micrographs (100×) of 0.5% safranin stained beech sections, (a) untreated beech (b) dry cured melamine-treated beech (c) steam cured melamine-treated beech. Bottom row: light micrographs (100×) of BSF stained sections. (d) untreated beech (e) dry cured melamine-treated beech (f) steam cured melamine-treated beech. Scale bar 100 µm.
The untreated references showed a saturated coloration by the safranin, whereas the melamine-treated specimens were barely stained. Biziks et al. (2015) visualised the difference in penetration depth of different molecular weight PF resins through the inability of safranin to stain the cross sections of modified beech wood compared to untreated beech. In this study, there was only one resin used and therefore no drastically different penetrations depths were to be seen. The effect of the different curing conditions became invisible in thin sections after staining. BSF staining led to brightly coloured melamine-treated sections. The untreated sections remained unstained. There were no apparent differences between the dry and steam cured sections. Different curing regimes can result in different resin distributions as demonstrated by Klüppel and Mai (2013). Therefore we assume that differences in cell wall penetration in the present study were not pronounced enough to distinguish between the processes by staining and LM. It can be concluded, that the staining methods for UF resin (Sernek et al. 1999, Mahrdt et al. 2015) proved to be highly efficient for the general detection of melamine resin in wood but not for the differentiation between the dry and high relative humidity process.
Scanning electron microscopy and energy-disperse X-ray spectroscopy
Scanning electron microscopy
SEM images of transverse surfaces of small blocks of MMF-treated beech specimens cut by microtome are shown in Figure 2(a,b).
SEM images 2500 × of (a) dry cured melamine-treated beech and (b) steam cured melamine-treated beech and (c)/(d) the respective EDX mapping of the nitrogen distribution. The arrows indicate micro cracks in the cell walls. 1: cell wall rupture across a single cell wall from the lumen to the middle lamellae. 2: internal cell wall rupture parallel to the cell wall located in the S2.
SEM images were used to evaluate the structure and condition of the cut surface of the specimens. Micro cracks were detected in both dry and steam-cured specimens with substantially more cracks in the dry-cured material. The micro cracks in the dry-cured material were found across single cell walls from lumens to the middle lamellae (arrow 1) and internal cell wall ruptures parallel to the cell wall located in the S2 (arrow 2). The influence of resin modification on the structural integrity of beech assessed by electron micrographs was reported by Bollmus (2011). The propagation of macro cracks was monitored but no micro cracks were reported. Mahnert et al. (2013) investigated the resin distribution in MMF-treated koto (Pterygota macrocarpa K. Schum.) and limba (Terminalia superba Engl. & Diels) with UMSP, but did not detect any micro cracks. To the best knowledge of the authors the effect of curing conditions on the micro structure of resin-treated wood had not been examined yet. It is known that drying conditions greatly influence the quality of dried wood. Fast drying with large drying rates leads to steep moisture gradients causing stress to the wooden matrix (Klüppel and Mai 2013). Data about the influence of high-temperature drying conditions (115°C) on the microstructure and the mechanical properties of Scots pine (Pinus sylvestris L.) suggested that high temperature drying caused micro cracks, but the mechanical properties were unaffected (Terziev and Daniel 2002). A similar temperature (120°C) was applied during resin curing in this study. A potential reason for the formation of the cracks might be the drying conditions rather than the resin modification. Bollmus (2011) also found that the curing conditions affected the structure of ray parenchyma cells, but there was no difference between dry and hot steam curing and resin or water impregnated specimens. Applying high temperatures between 120 and 130°C while curing led to the recorded damages (Bollmus 2011).
Energy disperse X-ray spectroscopy
EDX mapping of elemental nitrogen
The EDX mapping of nitrogen showed the distribution of the resin (Figure 2(c,d)). Both processes showed a rather even distribution across the scanned surface and across the cell walls. Li et al. (2012) also recorded an even distribution of the modification agents maleic anhydride and methacrylate across the cell wall matrix via their respective content. There was excess resin visible in the lumens, forming granules (‘bubbles’). More granules were visible in the steam-cured specimens than in the dry-cured specimens. Mahnert et al. (2013) reported the occurrence of granules in MMF modified wood, cured under wet conditions. Furuno et al. (2004) described that the granules form above a certain solid content/resin concentration (PF), below this concentration all the resin was located in the cell walls. Granules might indicate a saturated cell wall under the given curing conditions. Saturated cell walls might occur at high resin concentrations but also under high relative humidity curing conditions. The steam-cured specimens showed lower nitrogen concentrations in the cell walls in combination with more frequent granules in the lumens. This might be an indication of a difference in resin concentration between the cell wall and lumen of the steam-cured specimens and the dry-cured specimens. Klüppel and Mai (2013) reasoned that the higher diffusion potential of dry curing led to higher resin concentrations in dry-cured specimens.
EDX line scans
SEM images of two representative cross section areas of dry and steam cured beech with the respective EDX line scans and nitrogen cps (counts per second) can be seen in Figure 3.
SEM micrographs (5000×) of dry-cured beech (left) and steam-cured beech (right) with the position of the line scan (thick black line, below). Line scans with nitrogen counts per second (cps) across two cell walls of dry cured (left; n = 6) and steam cured (right; n = 4) melamine modified beech. Thin black lines: Moving average (20 SMA) of the normalised nitrogen counts per second (cps) of several line scans on different specimens.
The line scans revealed the resin distribution across the cell walls of MMF-treated beech with generally lower MMF concentrations towards the middle lamella (ML) than in the outer S2 and S3. Measurements from other authors revealed different concentration gradients from S2 to the ML. Higher MMF concentrations in beech towards the ML were reported by Kielmann et al. (2014) as analysed by UMSP. Mahnert et al. (2013) also reported higher resin concentrations in the ML than in the S2 of MMF-treated koto and limba via UMSP measurements. In contrast, Rapp (1999) used SEM-EDX and recorded a slightly higher nitrogen concentration in the ML than in the S2 and a steeply rising gradient in the S3 towards the lumens, similar to the results in this study. Rapp (1999) explained the findings with the higher accessibility of lignin-rich areas like the S3 and ML over the cellulose-rich S2 and a resin diffusion gradient from the lumen towards the ML. Furuno and Goto (1973) found lower resin concentrations in the S2 than in the ML. Gindl et al. (2003) recorded higher MMF concentrations in the S2 than the ML. The behaviour is explained by the higher affinity of the hydrophilic MMF resin towards the S2 as a less lignified cell wall region.
A generally lower nitrogen cps was attributed to the steam-cured specimens. The distribution of nitrogen across the cell wall itself did not seem to be affected by the curing conditions and showed a similar pattern.
Conclusions
Beech wood was treated with a MMF resin and cured under differently humid conditions to determine the influence of the curing process on the distribution of the chemical on a microscopic level. There were more micro cracks in the dry cured than in the steam cured material. Further on, the steam-cured specimens showed lower nitrogen concentrations in the cell walls in combination with more frequent granules in the lumens. LM in combination with two staining methods was a suitable method to distinguish between MMF-treated and untreated specimens. There was, however, no difference between the different curing processes detectable. SEM-EDX was a valuable tool to visualise the structural changes induced by the curing processes. The line scan function in SEM-EDX was suitable to detect the elemental nitrogen of the MMF resin and can be used to detect the resin distribution across cell walls. SEM-EDX could be very useful method for future analysis of the influence of curing processes on the material properties of resin-modified wood.
