Effect of Primary Spruce Lamella Aging on the Bending Characteristics of Glulam Beams

Abstract

Glued laminated (glulam) beams are used in the roofs, ceilings and walls of buildings as well as in bridges and towers. At present, with the limitation of tree harvesting, the production of glulam beams from recycled wood sources is implemented with the proviso that their mechanical properties and resistance to pests, fire and weathering will not be aggravated. This work deals with the primary effect of aging Norway spruce wood (Picea abies Karst. L.) lamellas on the moduli of rupture (MOR) and elasticity (MOE) in bending of three-layer glulam beams composed of sound and aged lamellas and polyurethane (PUR) glue. Three methods of lamella aging were used: (A) natural, lasting 60 years in the form of roof trusses with a greater or lesser degree of bio-attack by woodworm (Anobium punctatum De Geer); (B) artificial, caused by increased temperatures from 160 to 220 °C for 4 h; (C) artificial, caused by 2% water solutions of inorganic preservatives, namely, CuSO₄ × 5H₂O, ZnCl₂, H₃BO₃ or (NH₄)₂SO₄, for 28 days. The lowest MOR values were determined for glulam beams in which all three lamellas or two surface lamellas had a greater degree of bio-attack (60.5 MPa, a decrease of 25.9%) or were exposed to primary aging at 220 °C (62.6 MPa, a decrease of 23.3%). On the contrary, the exposure of lamellas to 160 or 180 °C did not significantly influence the MOR of beams (76.0–82.7 MPa, an average decrease of 1.6%). The MOE of glulam beams ranged from 7540 to 10,432 MPa without an obvious influence of the method of lamella aging or their location in the beams. Linear correlations between the MOR or MOE of glulam beams and the shear strength (σ) of glued joints, if both composite types consisted of similarly aged lamellas, were only slightly significant or insignificant.

1. Introduction

Glued laminated (glulam) beams and similar wooden composites, e.g., cross-laminated timber (CLT), overlayed materials, laminated wood–nonwood composites and multiwood composites (COM-PLY), are commonly used in structural engineering, furniture, sport, musical and decoration products [1]. In glulam, the defects in the sawn timber are cut out, and lamellas with a thickness of ≤50 mm are bonded together [2]. Glulam beams can be made from coniferous woods (Douglas fir, larch, Norway spruce, hemlock, Southern pine, etc.), broadleaved woods (azobé, beech, doussie, keruing, padouk, red maple, resak, etc.), as well as densified low-density woods (batai, poplar, etc.) [2,3,4,5], preferentially using thermosetting glues such as polyurethane (PUR), resorcinol–formaldehyde (RF), phenol–resorcinol–formaldehyde (PRF), melamine–formaldehyde (MF) or urea–melamine–formaldehyde (UMF). In Europe, the Norway spruce is the most common wood species used for glulam [2,5,6].

The final quality of glulam beams is influenced by many factors: (1) the wood species, whether it is used in its natural state or after chemical, thermal, mechanical or biological modifications, and what are its typical structural characteristics and properties, e.g., the density and porosity, the type and amount of extractives, the pH value, roughness, free surface energy, moisture content at bonding, tension and compression strengths, elasticity, and resistance to pests, fire and weathering; (2) the glue type and its physical–chemical characteristics, e.g., the molecular weight, viscosity, surface tension, chemical reactivity, resistance to water and total durability; and (3) the technology of gluing, e.g., the spread rate of glue and its open time and the pressing diagram used [7,8,9,10,11,12,13,14]. The applicability of structural glulam beams depends mainly on their mechanical properties and resistance to biological damage, fire and weathering. Today, when new sources of wood are limited, preparation of glulam beams from older wood products previously aged in various environments is actual.

During the demolition and reconstruction of buildings, a large amount of construction waste is generated, including wood waste. This usually ends up in municipal waste landfills, while only a small part of it is reused in the form of recycling [15,16] or for energy production [17]. In general, to increase the efficiency of the use of waste materials, a method of cascading resources was developed [18]. This is a method in which the efficiency of the use of a material resource is increased through its gradual or repeated use for several applications in individual consecutive cascading steps. By applying cascading in practice, the efficiency of the use of wood raw materials can also be increased, and thus, trees and wood stocks in the forest can be saved. The theoretical concept of wood cascading was first described by Fraanje [19]. Subsequently, the studies by Höglmeier et al. [20,21], Lubke et al. [22] and other researchers emphasized the environmental and economic advantages of this method for wood recycling in practice. Currently, despite all of the favorable aspects of the cascading method, wood waste is primarily used for the production of particleboards [23], cement chipboards [24] and possibly other types of agglomerated materials [25]. However, during their production, it is necessary to first chip the waste wood material, which does not fulfill the basic principle of cascading. In order to maximize efficiency, it is necessary to achieve the greatest possible number of cascading steps, and in the first steps, wood particles from the waste material, even larger than wood chips, should be used. So far, relatively few workplaces have dealt with such research, and most of the research has focused on the production and testing of the properties of CLT panels from recycled solid wood [26,27].

Glued laminated (glulam) wood can be characterized as a construction material that is created by gluing individual wood lamellas under prescribed technological conditions. The lamellas are oriented so that the longitudinal wood fibers coincide with the longitudinal direction of the final wooden element [28]. Currently, glulam wood is used in various construction applications, from load-bearing elements in one- and multi-story residential and non-residential buildings, through arched trusses covering buildings with large spans [29], to load-bearing elements of bridge structures [30,31,32]. In principle, glulam can be produced from any wood species, as long as suitable types of glues are developed and tested for the given species of wood. In European countries, glulam beams are made mainly from coniferous wood species, especially from spruce wood [31,33]. The definitions of and requirements for wood lamellas and glulam products are prescribed by the standard EN 14080:2013 [6].

The theoretical analysis of the efficiency of solid wood recycling for the production of glulam beams and other glued composite materials was solved in the study by Risse et al. [34]. The results of this study indicate that the recycling of solid wood for preparing new glulam beams can be environmentally and economically significant. However, relatively few researchers have dealt with this issue so far, and thus, it is not sufficiently reviewed. For example, Syrovátko [35] tested glulam beams made of lamellas prepared from old wooden elements of the roof structure, and he discovered that the mechanical properties of these beams were comparable to those of beams made of new lumber. The work by Mali [36] showed that the mechanical properties of glulam beams from old waste wood decreased by 4.7 to 10.5% compared to glulam beams based on current healthy wood, while the rate of decrease in their strength depended on the proportion and position of old and new wood in individual types of glulam beams. Based on this knowledge, it can be concluded that old waste wood obtained from the demolition of buildings may not always lose its original quality during its use for its initial purpose, and its mechanical properties are often at the level of new wood. These findings were confirmed by the review studies by Cavalli et al. [37] and Kránitz et al. [38], in which the experimental results of several researchers dealing with changes in the properties of old wood were compared. The knowledge described in these review studies showed that the age of old wood does not have a dominant effect on changes in its mechanical properties. This is self-evident as long as the old wood is not simultaneously degraded (1) by biological agents, i.e., especially due to rotting caused by wood-decaying fungi and larval galleries caused by wood-damaging insects, and (2) by abiotic factors, i.e., especially by thermal attacks during fires at ~300 °C, in the vicinity of hot chimneys above 100 °C or specifically in roofs during long-term elevated summer temperatures of 50–70 °C, and specifically by chemical attacks due to aggressive substances such as inorganic biocides and flame retardants.

However, in practice, older wood elements in the roofs, ceilings or log-walls are, in several cases, damaged by brown or white rot, insect galleries and high temperatures or contain chemical preservatives that served to increase their durability during use [39]. Therefore, in the cascading method of wood source utilization for glulam or other supporting composite elements, the correct choice of wood material is very important. Decreases in the strength of wood deteriorated by decaying fungi, insects, high temperatures or aggressive chemicals are usually evidently higher as decrease of its density [40]. Bio-attacked wood quickly loses its stiffness, as well as its tension and bending strength [41,42]. Wood exposed to high temperatures above 160 °C loses its strength quickly and its stiffness slowly [43], it is less hygroscopic–more hydrophobic, and its surfaces are less wettable by polar liquid glues [43,44,45]. Bastani et al. [46] found that the processing time needed for the glue to be absorbed into thermally modified wood is longer due to the slower penetration rate. Changes in the pH value of the thermally modified wood surface might retard or accelerate the curing of glues, depending on their type [47]. Finally, the mechanical properties of glulam from thermally modified wood lamellas are usually lowered [48]. In many cases, chemical preservatives—many of which are aggressive chemical substances or such substances can be formed over time in a wet environment—negatively affect the wettability of wood lamellas with polar glues [12] and deteriorate the lignin–saccharide matrix of wood cells in lamellas [39]. So, such chemicals can worsen the adhesion of glues to wood lamellas and/or decrease the strength of wood lamellas and finally degrade the mechanical properties of glulam beams [49]. For example, in the study by Piao et al. [50], who dealt with the use of wood material from old pine telecommunication poles in the production of glulam beams, the strength of glulam prepared from aged wood was reduced due to the presence of the biocidal preservative CCA (chromated copper arsenate). However, today, in several European countries, old wood containing CCA, creosote, Cu-azole, pentachlorophenol or other more toxic contaminants of the fourth healthy-environmental class cannot be used for new panels including glulam [51].

The basic aim of this work was to examine the influence of the primary aging of spruce lamellas—natural aging under the action of wood-destroying insects, as well as artificial aging due to high temperatures or inorganic wood preservatives commonly used in the past—on the bending characteristics of glulam beams. Simultaneously, the relationship between the bending characteristics of glulam beams prepared from aged and sound spruce lamellas and the shear strength of glued joints between these lamellas was evaluated.

2. Materials and Methods

2.1. Spruce Wood Lamellas

For the experiment were used prisms of Norway spruce (Picea abies Karst. L.) wood with the dimensions of 2000–3000 mm (length) × 140–300 mm (width) × 24–30 mm (thickness). The prisms were cut (1) from several naturally aged 60-year-old bio-attacked roof trusses of a reconstructed house, No. 320 in village Brodzany, Slovakia, and (2) from one sound 96-year-old tree trunk obtained from the National Forest Centrum in Zvolen, Slovakia. Then, the prisms were conditioned for 1 year outside under shelter at Technical University in Zvolen. From the prism were cut wide pre-lamellas of type I with the dimensions of 580 mm × 80 mm × 10 mm (longitudinal × radial × tangential), which were used for three-layer glulam beams (Section 2.3 and Section 2.4), and lamellas of type II with the dimensions of 80 mm × 20 mm × 5 mm (longitudinal × radial × tangential), which were used for the double-bonded sets (Section 2.5). Both of the top surfaces of lamellas were gradually milled in a milling machine (D630-EL; Robland, Belgium) and subsequently machine-sanded with 120-grit sandpaper.

2.2. Aging of Lamellas

The natural aging of lamellas was implemented directly in the 60-year-old roof trusses from which they were prepared. The trusses were subjected to non-homogeneous bio-attacks by larval galleries of woodworm—the common furniture beetle (Anobium punctatum De Geer) with a diameter up to 1.5–2 mm. The degree of the bio-attack of prepared lamellas was determined visually, i.e., a greater degree (GD) for lamellas from the outer zone, usually the sap zone, of truss beams, with 50 or more larval galleries per m², and a lesser degree (LD) for lamellas prepared from the inner zone of truss beams, with no or less than 50 larval galleries per m² (Figure 1).

For the artificial aging, lamellas were cut from a 96-year tree trunk. They did not have sap zones, knots, fungal rots or insect galleries. The thermal aging of lamellas took place at atmospheric pressure and temperatures of 160, 180, 200 or 220 °C for 4 h in a drying oven (Memmert UFE 500; Schwabach, Germany). The chemical aging of lamellas was performed using a double-coating treatment (2 × 120 ± 10 g/m², using a 24 h break) with 2% water solutions of inorganic wood preservatives traditionally used in the past (copper sulfate CuSO₄ × 5H₂O, zinc chloride ZnCl₂, boric acid H₃BO₃ and ammonium sulfate (NH₄)₂SO₄)), followed by their 28-day exposure at 20 °C ± 2 °C and 65% ± 5% RH.

Aged and sound spruce lamellas for glulam beams were conditioned at 20 °C ± 2 °C and 65% ± 5% RH, with achieving their moisture content (MC) of 9% ± 3%. For example, lamellas aged at 220 °C had an MC of ~6%, while the MC of the sound and naturally aged lamellas was ~12%.

The density of spruce lamellas in their conditioned state ranged in a narrow interval from 418 to 467 kg·m⁻³, i.e., 459 kg·m⁻³ for sound lamellas; 433 and 418 kg·m⁻³ for naturally aged lamellas from the inner lesser (LD) and outer greater (GD) bio-attacked zones of old trusses; 451 to 426 kg·m⁻³ for lamellas thermally aged at 160 to 220 °C; and 448 to 467 kg·m⁻³ for chemically aged lamellas. So, these small differences in the densities of the lamellas should not affect the mechanical properties of glued wood elements.

2.3. PUR Glue and Gluing of Lamellas

The one-component PUR glue Kestopur 1030 (Kiilto, Tampere, Finland) is characterized by the following technical parameters—viscosity of 7000 mPa·s, density of 1200 kg/m³, open time of 30 min and pressing time of 90–120 min. PUR glue was applied using coating technology on the interconnected surfaces of spruce lamellas at the recommended rate of 180 g/m² ± 10 g/m². The three-layer glulam beams (Figure 2), as well as the double-bonded sets (see figure in [12]), were prepared by pressing spruce lamellas coated with PUR glue. Pressing was performed in steel bolts at a pressure of 1.2 MPa, a temperature of 20 °C ± 2 °C and an RH of 65% ± 5% for 2 h (Figure 2a). Before testing the quality of the prepared glulam beams and double-bonded sets (see Section 2.4 and Section 2.5), the 7-day conditioning of these beams to set equilibrium MCs of 9% ± 3% was performed under the same air conditions.

2.4. Bending Characteristics—MOR and MOE

The conditioned three-layer glulam beams, with the dimensions of 580 mm × 20 mm × 30 mm, were subjected to the 3-point bending test in a TiraTest 2200 (VEB TIW Rauenstein, Germany) device with a maximum force of 10 kN. The distance of the supports was 540 mm, which is 18 times the beam height. The values of MOR and MOE in bending were determined according to the standards EN 408:2010+A1:2012 [52] and EN 310:1993 [53], using Equations (1) and (2):

$${\mathsf{\sigma }}_{MOR}=\frac{3 \times F_{max} \times l_0}{2 \times b \times h^2}\left(\mathrm{Mpa}\right)$$

(1)

$${\mathsf{\sigma }}_{MOE}=\frac{\Delta F \times l_0{}^3}{4 \times \Delta \mathrm{y}\times b \times h^3}\left(\mathrm{Mpa}\right)$$

(2)

where MOR is the modulus of rupture in bending (Mpa), MOE is the modulus of elasticity in bending (Mpa), F_max is the force corresponding to the breaking strength (N), l₀ is the distance between the supports (mm), b is the width of the glulam beam (mm), h is the height of the glulam beam (mm), ΔF is the load increment, and Δy is the deflection increment.

The 1st (reference) group of glulam beams consisted of three sound lamellas, the 2nd group consisted of one aged lamella in the middle layer and two sound lamellas in the surface layers, the 3rd group consisted of two aged lamellas in the surface layers and one sound lamella in the middle layer, and the 4th group consisted of three aged lamellas (Figure 2 and Figure 3).

2.5. Shear Strength

The shear strength (σ) of glued joints in the conditioned double-bonded sets, which consisted of two lamellas of 80 mm × 20 mm × 5 mm with a mutual contact area of 10 mm × 20 mm, was determined according to the standard EN 205:2016 [54] using a LabTech 4.050 (LaborTech s.r.o., Opava, Czech Republic) device. The 1st (reference) double set consisted of two sound lamellas, the 2nd double set consisted of one aged lamella and one sound lamella, and the 3rd double set consisted of two aged lamellas. Double sets from naturally aged lamellas were prepared and tested in this work, while those from artificially aged lamellas were prepared and tested previously in the works of Vidholdová et al. [55] and Ciglian and Reinprecht [12].

2.6. Statistical Analyses

The evaluations of the measured data (arithmetic means and standard deviations) and the processing of linear correlations between the bending characteristics (MOR or MOE) of glulam beams and the shear strength of glued joints in double-bonded sets were performed using Microsoft Excel (Version 2010; Microsoft Corporation, Redmond, WA, USA). The statistical software STATISTICA 12 (StatSoft Inc., Tulsa, OK, USA) was used to evaluate Duncan’s tests to measure specific differences between pairs of means (with indexes of significance: a—very significant decrease at the 99.9% level; b—significant decrease at the 99% level; c—less significant decrease at the 95% level; and d—insignificant change (decrease or increase) at the <95% level) of the bending characteristics of glulam beam replicates containing various combinations of aged and sound lamellas (2nd–4th groups) in comparison to reference replicates containing sound lamellas (1st group).

3. Results and Discussion

3.1. Bending Characteristics of Glulam Beams Affected by Aging of Lamellas

3.1.1. Influence of Lamella Aging Methods and Their Locations in Beams

For the 1st–4th groups of three-layer glulam beams, consisting of spruce lamellas of type I (Section 2.1) and PUR glue (Figure 2 and Figure 3), the bending characteristics MOR and MOE were determined (

Table 1. The bending characteristics MOR and MOE of the three-layer glulam beams in the 1st, 2nd, 3rd and 4th groups (Figure 3) prepared from Norway spruce lamellas of type I and PUR glue (Figure 2).

Glulam Beam	MOR (MPa)	MOE (MPa)
1st group (S-S-S)
Sound	81.63 (2.66)	10,432 (742)
2nd group (S-A-S)
GD bio-attack	72.78 (5.38) b	10,864 (693) d
LD bio-attack	74.45 (0.82) c	9956 (1298) d
T-160 °C	81.74 (1.63) d	9566 (332) d
T-180 °C	82.65 (1.93) d	11,574 (391) d
T-200 °C	77.53 (2.65) d	9540 (1360) d
T-220 °C	68.29 (4.75) a	10,564 (1470) d
CuSO4·5H2O	75.29 (5.69) c	9701 (598) d
ZnCl2	70.78 (5.60) a	9806 (674) d
H3BO3	80.03 (3.05) d	10,905 (1055) d
(NH4)2SO4	73.08 (7.18) b	8856 (384) b
3rd group (A-S-A)
GD bio-attack	60.98 (3.21) a	8292 (913) a
LD bio-attack	67.37 (3.74) a	9614 (426) d
T-160 °C	80.85 (3.50) d	9145 (1456) c
T-180 °C	79.88 (6.54) d	11,554 (499) d
T-200 °C	70.38 (6.50) a	10,443 (484) d
T-220 °C	62.60 (4.63) a	9518 (731) d
CuSO4·5H2O	71.09 (4.58) a	10,591 (704) d
ZnCl2	68.92 (1.36) a	9151 (414) c
H3BO3	71.20 (1.63) a	8415 (517) a
(NH4)2SO4	70.82 (5.38) a	9052 (269) c
4th group (A-A-A)
GD bio-attack	60.02 (9.86) a	7540 (1032) a
LD bio-attack	74.99 (7.18) c	9738 (670) d
T-160 °C	80.97 (7.51) d	9841 (860) d
T-180 °C	76.03 (6.19) d	9826 (1436) d
T-200 °C	69.29 (7.38) a	10,155 (543) d
T-220 °C	62.56 (2.17) a	8508 (811) a
CuSO4·5H2O	69.20 (4.20) a	9686 (396) d
ZnCl2	70.32 (3.55) a	10,542 (1689) d
H3BO3	75.45 (2.54) a	10,825 (920) d
(NH4)2SO4	69.87 (1.54) a	9012 (1399) c

Notes: Mean values are from 6 measurements. Standard deviations are in parentheses. Duncan’s tests were used to evaluate the bending characteristics of the 2nd, 3rd or 4th groups of glulam beams consisting of 1, 2 or 3 aged “A” lamellas, relative to the 1st (reference) group of glulam beams consisting of 3 sound “S” lamellas, with indexes of significance: a—very significant decrease at 99.9% level; b—significant decrease at 99% level; c—less significant decrease at 95% level; and d—insignificant change (decrease or increase) at <95% level.

and Figure 4).

The MOR of glulam beams, in many cases, significantly depended—as determined by Duncan’s tests—both on the method of lamella aging and their locations in the beams (Table 1). The highest MORs were determined for glulam beams in the first (reference) group consisting only of sound spruce lamellas (81.63 MPa) and those in the second group, in which the middle lamella was initially aged at lower temperatures of 160 or 180 °C (81.74 or 82.65 MPa). However, the MOR in the second group of glulam beams was evidently smaller at the 99.9% or 99% significance level if the middle lamellas were made from naturally aged prisms with a greater degree (GD) of bio-attack (72.78 MPa; a decrease of 10.8%) or if the middle lamellas were artificially aged at a maximum temperature of 220 °C or in the presence of ZnCl₂ and (NH₄)₂SO₄ (68.29–73.08 MPa; decreases of 16.3–10.5%).

In accordance with the theoretical assumption, the lowest MOR were obtained for glulam beams of the third group and fourth group, in which aged lamellas were localized in the surface “tension and compression” zones. For example, the MOR decreased by 25.9–23.3% if such lamellas were prepared from the outer greater bio-attacked zone of naturally aged prisms (~60.5 MPa), or if these lamellas were artificially aged at a maximum temperature of 220 °C (~62.6 MPa). The determination of similar values of the MOR for the third group (A-S-A) and the fourth group (A-A-A) of glulam beams created from lamellas aged (A) artificially at 220 °C or aged naturally with a greater attack by A. punctatum is in good accordance with the theory that for the bending strength of beams, including glulam beams, the most important zones are their tension and compression zones.

The MOE of glulam beams depended less, compared to the MOR, on the method of spruce lamella aging and their locations in glulam beams. This was determined by Duncan’s tests (Table 1). Quite high MOEs, approximately from 9000 to 11,000 MPa, were obtained for various glulam beam types, i.e., not only in the first and second groups but, in some cases, also from the third and fourth groups. Apparently lower MOEs, from 8000 and 9000 MPa, were observed for some glulam beams in which aged lamellas were localized (1) in the middle “neutral” zone of the second group—aged by (NH₄)₂SO₄; (2) in the surface “tension and compression” zones of the third group—with a GD of bio-attack or aged by H₃BO₃; (3) in all three zones of the fourth group—aged at 220 °C. The lowest MOE, only 7540 MPa, was obtained for glulam beams in the fourth group consisting of three lamellas with a greater degree (GD) of bio-attack by A. punctatum.

Generally, a decrease in the bending properties of glulam beams made from aged lamellas can be explained (1) by the creation of poorly glued joints, as well as (2) by the presence of deteriorated wood in beams having lower strength and stiffness. Poorly glued joints are created when using wood with a more hydrophobic surface or a surface weakened by bio-, thermal and/or chemical attacks. Worse bending properties of wood damaged by larval galleries were found by Reinprecht and Pánek [42]. Temperatures from 160 °C to 220 °C cause only a minor change (“increase or decrease”) in the MOE but a large decrease in MOR from 10% to 50% with a rise in temperature [56,57,58]. The mechanical properties of wood decrease after its attack by acids, bases and other aggressive chemicals [40]. For example, the negative effect of wood impregnation with compounds of trivalent boron on the strength of glued joints and the strength of wooden composites significantly depends on the type of wood and the type of polymer used as the glue or coating [59,60]. Keskin and Mutlu [61] researched the effect of impregnating beech, oak and pine woods with boron preservatives on wood strength, and they found that wood containing sodium tetraborate (Na₂B₄O₇·10H₂O) had increased bending strength, whereas many mechanical properties of wood decreased in presence of H₃BO₃ and other trivalent boron types.

3.1.2. Influence of Temperatures during Lamella Aging

With the increase in the temperature applied during the initial aging of spruce lamellas from 160 to 220 °C, the MOR of the prepared glulam beams gradually decreased. This was proved by the individual linear correlations in the second group (MOR = 84.63 − |0.040 × T|; R² = 0.24), third group (MOR = 85.97 − |0.070 × T|; R² = 0.38) and fourth group (MOR = 85.90 − |0.076 × T|; R² = 0.34) of beams, as well as by the summary linear correlation evaluating all three groups of glulam beams together (MOR = 90.39 + |0.087 × T|; R² = 0.29) (Figure 5a). On the contrary, the MOE values of glulam beams were not apparently influenced by the initial thermal aging of spruce lamellas, as was confirmed based on the small coefficients of determination: R² from 0.0003 to 0.19 for the individual linear correlations of the 2nd–4th groups and R² = 0.015 for the summary linear correlation (Figure 5b).

3.2. Shear Strength of Glued Joints Affected by Aging of Lamellas

The shear strengths (σ) of glued joints in the first, second and third double-bonded sets prepared from naturally aged lamellas of type II (Section 2.1) and PUR glue (Section 2.3 and Section 2.5) are presented in

Table 2. The shear strengths (σ) of glued joints in the double-bonded sets consisting of Norway spruce lamellas of type II and PUR glue.

Double-Bonded Set	Shear Strength—σ (MPa)
1st set (S-S)
Sound	10.82 (1.39)
2nd set (S-A)
GD bio-attack	10.85 (1.24) d
LD bio-attack	11.02 (0.86) d
3rd set (A-A)
GD bio-attack	9.90 (1.30) d
LD bio-attack	10.94 (1.49) d

Notes: Mean values are from 10 measurements. Standard deviations are in parentheses. Duncan’s tests were used to evaluate the shear strength of the 2nd or 3rd double-bonded set consisting of 1 or 2 aged “A” lamellas, relative to the 1st (reference) double-bonded set consisting of 2 sound “S” lamellas, with indexes of significance: a—very significant decrease at 99.9% level; b—significant decrease at 99% level; c—less significant decrease at 95% level; d—insignificant change (decrease or increase) at <95% level.

.

According to Duncan’s test for the shear strength, no significant differences occurred between the first set prepared from two sound lamellas (σ = 10.82 MPa) and the second or third sets prepared from one or two naturally aged lamellas (σ from 9.90 to 11.02 MPa). This unexpected result was potentially caused by PUR glue possibly seeping into A. punctatum larval holes when, after glue curing, quasi-pin joints between the jointed lamellas can be created with their orientation perpendicular or at a defined angle to shear forces, and thanks to which the shear strength of glulam beams from the naturally aged lamellas did not worsen. For this reason, the correlation analyses of MOR or MOE = f (σ) in Section 3.3 had to be performed without the sets of naturally aged lamellas and the groups of glulam beams made from them.

The shear strengths (σ) of glued joints of the first, second and third double-bonded sets prepared from artificially aged spruce lamellas were published and discussed by Vidholdová et al. [55], who used thermally aged lamellas, and by Ciglian and Reinprecht [12], who used chemically aged lamellas. In this work, the achieved shear strength values at Section 3.3 were applied in the correlation analyses of MOR or MOE = f (σ). The strength of the glued joints apparently decreased at the level of significance of 95% and as high as 99.9% when one lamella of the double set was aged at a temperature of ≥200 °C or in the presence of ZnCl₂ and H₃BO₃, or when both lamellas of the double set were aged at a temperature of ≥160 °C or in the presence of one of the four used inorganic preservatives.

3.3. MOR and MOE in Bending of Glulam Beams Versus Shear Strength of Glued Joints

Comparing the bending characteristics of glulam beams with the shear strength (σ) of glued joints, certain significant links were observed only for the MOR; however, none were observed for the MOE (Figure 6).

Generally, the artificial aging processes of spruce lamellas manifested in glulam beams as an apparent loss of MOR; however, they underwent less significant changes in their MOE (Table 1). This observation is in accordance with a greater negative effect of high temperatures above 160 °C or several aggressive chemicals on the wood strength than on the wood deflection or other deformation types [43]. As the bending characteristics of glulam beams depend not only on the used wooden materials-lamellas (species, quality, amount and location in beam, etc.) but also on the used glue (type, quality, amount, etc.), the mechanical properties of glulam beams significantly depend on the “wood-glue” interfaces. Their quality is reflected in the adhesion strength, which, in practice, is usually determined by the shear strength.

The shear strength (σ) of sets created from spruce lamellas and PUR glue was negatively influenced by the primary artificial aging of lamellas at high temperatures of ≥200 °C [55], as well as by the presence of inorganic preservatives, namely, CuSO₄ × 5H₂O, ZnCl₂, H₃BO₃ and (NH₄)₂SO₄ [12].

In wood treated at high temperatures, there in polysaccharides and lignin are created new substances containing fewer polar -OH groups [43]. Such wood is more hydrophobic, has worse wettability and, finally, also forms weaker adhesion to polar glues. According to Taghiyari et al. [62], a reduction in the shear strength of glued joints created from thermally modified wood can be attributed to (1) a reduction in polar groups in the cell walls of wood due to the degradation of amorphous polysaccharides by heat treatment, resulting in fewer sites available for bonding; (2) the increased stiffness of cell walls after heat treatment, which results in a reduction in internal surfaces for the chemical bonding or mechanical interlocking of adhesives; and (3) a reduction in the wettability of wood, which may retard the proper penetration of glues. The formation of micro-cracks and checks, as well as the decrease in wood density due to heat treatment at temperatures above 180 °C [63], also might contribute to a decline in the shear strength of bonded wood.

A decrease in the shear strength of glued joints for wood impregnated with chemicals can be explained by many factors, e.g., by the reduced penetration of glue into lumens and intercellular spaces in wood filled with chemicals, followed by reduced adhesion, as documented by Atar et al. [64] and Özçifçi and Okcu [65].

Generally, due to the action of high temperatures or aggressive chemicals with a low or high pH value, the oxidative character of the lignin–saccharide matrix of wood can possibly deteriorate, the individual wood cells can be defibrated and, finally, a decrease in the mechanical properties of wood occurs [40,66]. A decrease in adhesion strength also occurs owing to the presence of new improper, especially hydrophobic, substances on the surfaces of thermally aged wood, as well as owing to the lower penetration of glue into cell lumens filled with preservatives or other chemical substances [67,68,69]. Subsequently, a decrease in the mechanical properties of wood and changes in the wood–glue phase interface can negatively affect the strength of glued joints [48,65].

From the above-mentioned views, it is possible to raise the question of why only slightly significant relationships were found between the MOR and σ (Figure 6a), or even none between the MOE and σ (Figure 6b). This question can be explained by other types of forces acting during bending, in addition to shear forces in the adhesive joint, e.g., by the action of tension and compression forces localized directly in the wood cells of surface lamellas, as well as by the combination of shear and non-shear loadings of the adhesive layers. At the same time, it is necessary to point out that while during the thermal aging of spruce lamellas, the quite homogeneous degradation of wood anatomical elements in their outer and inner zones could take place [70], on the contrary, during the chemical aging of spruce lamellas, the deeper penetration of inorganic chemicals into the refractory spruce wood is limited [71], and therefore, the degradation of the anatomical elements could occur only in the surface parts of spruce lamellas exposed to bonding, i.e., without a negative effect on the strength of the lamellas [12].

4. Conclusions

• The type of natural and artificial aging of spruce lamellas and the location of aged lamellas in three-layer glulam beams affected the MOR more than the MOE.
• The largest negative effect on the MOR of glulam beams was exerted by the primary thermal aging of lamellas at a maximum temperature of 220 °C.
• The smallest MOR was acquired by glulam beams in which the location of aged lamellas was in their surface “pressure and tension” zones, in accordance with bending theory.
• The linear correlation analyses confirmed certain relationships between the MOR of glulam beams and the shear strength (σ) of bonded joints also containing aged lamellas. However, in this experiment, these relationships were significant only in one case, i.e., if glulam beams had a sound lamella in the middle (A-S-A) and the bonded joints consisted of a sound lamella and aged lamella (S-A).
• In general, it can be stated that lamellas from recycled wood sources with damages caused by a higher number of insect galleries or by temperatures above 180 °C or the presence of certain inorganic wood preservatives should not be used for glulam beams designed for load-bearing structures.
• In practice, individual lamellas from aged woods should be selected and then, in the glulam beams, positioned according to their defect type, degree and range to ensure minimal decreases in the MOR and MOE. Before gluing, the lamellas should be dried to a suitable moisture content (usually 6–15%), and the surface of each lamella should be accurately machined to ensure that the thickness of the glue layer will be even throughout and that no dirt, grease and clumps of chemical substances will be present there.