1 Introduction
The use of fossil energy for thermal conditioning of buildings contributes significantly to the global warming by CO2 emissions. There are several techniques established to reduce the energy consumption of a building for thermal conditioning like thermal insulation, shading, orientation of the building or positioning of the windows etc. An aspect, which is often not specifically designed is thermal mass or heat capacity, although it is well known that wood buildings have, especially in the summer, drawbacks compared to heavy buildings made, for example, of concrete or brick due to low thermal mass. This often results in the necessity of active air-conditioning in order to keep the room temperature in a pleasant range.
Heat capacity is defined as the product of mass by specific heat capacity of the material. Thus, in order to increase the heat capacity of a building either the amount of material can be increased or a material with higher specific heat capacity can be used. With regard to limited resources, increasing mass just for the purpose of improving thermal conditions seems not to be desirable as it increases the CO2 footprint and is usually also associated with decreasing usable space or increasing the volume of the building. On the other hand, modifying the specific thermal mass of a material is not easily possible.
Anzeige
However, wood has a cellular structure, which can be used for impregnation to increase the mass and depending on the specific heat capacity of the impregnation material also the heat capacity of the construction element. A special class of materials, so-called phase change materials (PCM), possess a very high specific heat capacity in a narrow temperature range and might be good candidates for impregnation of wood. Especially paraffinic PCM seem to be suitable as they are inert and do not react with the chemicals of the wood. Thus, a long endurance of the PCM is expected. Literature reviews on the impregnation of wood with PCM are given, for example, in Sulaiman et al. (2022) and Hartig et al. (2023).
By applying PCM, a low amount of additional mass leads to a high increase in heat capacity of a timber element. Previous investigations (Hartig et al. 2021, 2023) showed that it is realistic to deposit about 200 kg of paraffinic PCM per m³ of wood. Thus, depending on the applied wood species the mass is increased by about 30 to 50%. At the same time, the storable thermal energy is increased in the phase transition range of the PCM by 400 to 600% (assuming a sensible enthalpy of wood of about 20 J/g and a latent enthalpy of PCM of about 200 J/g in the range of +/- 5 °C around the phase transition temperature). By placing PCM in wood elements, temperature amplitudes are passively reduced inside a room over the day by energy uptake of the PCM. By natural ventilation, the room temperature can be decreased over the night below the phase change temperature of the PCM and the stored thermal energy is released.
Most of the previous investigations from the literature concentrate on the application of PCM-impregnated wood for interior elements without structural functions like flooring (Mathis et al. 2018; Hartig et al. 2021, 2023). In the current investigations, elements like beams, walls or ceilings, bearing structural loads are intended. A similar objective is also followed, for example, in Saavedra et al. (2021). For practical applications, several aspects need to be investigated.
Key aspects for applying timber in structural elements are the mechanical properties and the combustion behaviour. The database in the literature on the mechanical properties of PCM impregnated wood is scarce. Saavedra et al. (2021) presented results for Pinus radiata impregnated with octadecane. Regarding the combustion behaviour of PCM-impregnated wood, information from literature is missing. Results of respective investigations are given in this contribution. Paraffin possesses very high heat release rates during combustion (e.g. Zhang et al. 2013), which in combination with the inflammable material wood gives rise to safety concerns.
Anzeige
2 Materials and methods
2.1 Materials
Four wood species were investigated: spruce (Picea abies Karst.), beech (Fagus sylvatica L.), poplar (Populus nigra L.) and oak (Quercus robur L.). The used PCM was the commercially available paraffin RT35HC produced by the company Rubitherm, Berlin, Germany. It has a melting point of about 35 °C. This melting temperature was chosen for a convenient handling in the tests by having the PCM in the solid state at room temperature. For practical applications similar PCM with the melting temperature in the range of pleasant room temperatures (about 19–23 °C) are available. The wood was impregnated in an autoclave with a maximum pressure of 0.8 MPa for 3 h as described in Hartig et al. (2023). The effectiveness of the impregnation is described by the PCM uptake ρPCM/wood, which is the mass of PCM within a certain volume of wood. It can be calculated as:
$${\rho _{PCM/wood}} = \left( {{m_{imp}} - {\rm{ }}{m_{wood}}} \right)/V$$
(1)
where mwood and mimp are the mass of the sample before and after the impregnation and V the volume. The density of the wood ρwood can be calculated with mimp = 0. The moisture content of the wood was about 10–12% before impregnation. It shall be noted that usually a certain amount of PCM is leaching from the wood after impregnation, which also reduces ρPCM/wood. The leaching behaviour depends on the wood species (Hartig et al. 2023). The values given later on for ρPCM/wood correspond to the values just before the respective test, when the leachable portion of PCM was allowed to leave the wood in large part, not just after the impregnation. Moreover, leaching can be reduced by adding particles to the PCM. In the current investigations, cement particles of about 2% by weight were added to the PCM. The cement was the commercially available product Mikrodur R-X produced by the company Dyckerhoff, Germany. Wood and PCM alone have approximate properties according to Table 1.
Table 1
Selected properties of wood and PCM
Property | Spruce | Beech | Poplar | Oak | RT35HC* |
|---|---|---|---|---|---|
Density [kg/m³] | 460 (DIN 68364 2003) | 710 (DIN 68364 2003) | 440 (DIN 68364 2003) | 710 (DIN 68364 2003) | 880 (solid) 770 (liquid) |
Melting temperature [°C] | - | 34–36 | |||
Enthalpy (in the range 27 − 42 °C) [kJ/kg] | (42 °C − 27 °C) ∙ ~2 kJ/(kg °C) ≈ 30 | 240 | |||
Flash point [°C] | 200–275 (Niemz 1993) | 177 | |||
2.2 Mechanical behaviour
2.2.1 Bending behaviour
For structural purposes, the mechanical behaviour of the impregnated timber is important. As the PCM might be available in the solid and the liquid state, the mechanical behaviour is determined for both states. As a first indicator of structural performance, three-point bending tests (span 420 mm) were carried out according to DIN 52186 (1978). The dimensions of the samples were 450 mm by 20 mm by 25 mm (length by width by height). Both, clear and impregnated samples of spruce, beech and poplar were tested. The impregnated samples of beech and spruce had a PCM uptake of ρPCM/wood ≈ 200 kg/m³. Poplar had with ρPCM/wood ≈ 250 kg/m³ a higher PCM uptake. Oak was not tested since the PCM uptake in large samples representative for timber is low and limited to the regions close to the surface. Thus, it can be assumed that it does not substantially influence the mechanical behaviour for timber elements. The tests were performed at room temperature with the PCM either in the solid state or after a conditioning of the respective samples in a climate chamber at 50 °C (60% relative humidity) in the liquid state. For each parametric combination, 10 samples, thus, in total 120 samples, were tested.
2.2.2 Surface hardness
A mechanical property, which has less importance for structural applications, but might be interesting for aesthetic and functional reasons in non-structural elements like flooring, is the surface hardness. Tests corresponding to EN 1534 (2020) were performed to determine values of Brinell hardness. In this test, a ball-shaped indenter with a diameter of D = 10 mm is pressed with a force of F = 1 kN on the surface of the sample. The force is applied within 15 s and kept for 25 s. The Brinell hardness HB can be determined with the diameter of indentation d with
$$HB\, = \,2F/({\mkern 1mu} (\mu D(D - {\left( {{D^2} - {d^2}} \right)^{0.5}}))\,for\,0 < d < D.$$
(2)
Deviating from EN 1534 (2020), d was not determined 3 min after removing F on the surface, but directly calculated from the indentation depth t, which was determined from the cross-head displacement of the testing machine, using equation.
$$d\, = \,2{\left( {Dt - {t^2}} \right)^{0.5}}\,for\,0 < t < D/2.$$
(3)
As the recovery of the indentation is not considered, the remaining indentation is overestimated and, thus, the Brinell hardness compared to EN 1534 (2020) underestimated. This deviation was accepted since the aim of the investigation was the qualitative determination of the influence of the PCM deposition and the aggregate state of the PCM on the surface hardness. The samples had dimensions of 100 mm by 100 mm by 30 mm (length by width by thickness). The grain direction corresponded to the length direction of the sample. The indentation tests were performed transverse to the grain but there was no special orientation of radial and tangential direction of the wood. Besides clear reference samples of spruce, beech, poplar and oak, respective samples with PCM and with cement-containing PCM were prepared. The samples were stored at room climate. The density and PCM uptake of the samples are given in Table 2. For the investigations in the liquid state, the samples were conditioned in a climate chamber at 50 °C and 60% relative humidity. On each sample, 16 indentation tests were performed on a 4 by 4 pattern, see also Fig. 1. Thus, in total 192 indentation tests were carried out. At first, 8 tests with a chessboard like allocation of the test pattern were carried out at room temperature for each sample to investigate the surface hardness in the solid state of the PCM. After storage at elevated temperature in the climate chamber, the remaining 8 indentation tests were performed on each sample to determine the surface hardness when the PCM is in the liquid state.
Fig. 1
Test setup for investigations on the surface hardness
×
![]()
Table 2
Properties of the samples of the surface hardness tests and characteristic values of the Brinell hardness
Wood species/ impregnation/ temperature | ρPCM/wood [kg/m³] | ρwood [kg/m³] | HB* [N/mm²] |
|---|---|---|---|
Spruce | |||
Impregnated (PCM & cement) cold | 144 | 560 | 20.64; 2.69; 1.30 |
Impregnated (PCM & cement) warm | 14.24; 1.25; 0.89 | ||
Impregnated (PCM) cold | 159 | 570 | 18.63; 2.51; 1.17 |
Impregnated (PCM) warm | 13.04; 1.49; 0.82 | ||
Clear cold | - | 586 | 15.93; 3.50; 1.00 |
Clear warm | 14.14; 3.06; 0.89 | ||
Beech | |||
Impregnated (PCM & cement) cold | 169 | 645 | 37.79; 5.00; 1.51 |
Impregnated (PCM & cement) warm | 22.35; 1.79; 0.89 | ||
Impregnated (PCM) cold | 136 | 723 | 38.73; 3.45; 1.55 |
Impregnated (PCM) warm | 27.33; 2.14; 1.09 | ||
Clear cold | - | 675 | 25.03; 2.75; 1.00 |
Clear warm | 22.88; 1.60; 0.91 | ||
Poplar | |||
Impregnated (PCM & cement) cold | 296 | 382 | 18.73; 2.15; 1.90 |
Impregnated (PCM & cement) warm | 9.85; 1.61; 1.00 | ||
Impregnated (PCM) cold | 268 | 339 | 11.98; 0.73; 1.22 |
Impregnated (PCM) warm | 7.41; 1.36; 0.75 | ||
Clear cold | - | 406 | 9.85; 1.65; 1.00 |
Clear warm | 9.26; 1.74; 0.94 | ||
Oak | |||
Impregnated (PCM & cement) cold | 130 | 625 | 27.24; 6.36; 1.15 |
Impregnated (PCM & cement) warm | 21.20; 5.22; 0.89 | ||
Impregnated (PCM) cold | 143 | 706 | 32.78; 1.53; 1.38 |
Impregnated (PCM) warm | 24.03; 1.70; 1.01 | ||
Clear cold | - | 617 | 23.73; 2.51; 1.00 |
Clear warm | 21.60; 1.73; 0.91 | ||
2.3 Combustion behaviour
2.3.1 Cone calorimetry according to ISO 5660-1
The combustion behaviour of materials can be tested with a so-called cone calorimeter according to ISO 5660-1 (2015). In this test, a quadratic board shaped sample is heated at the top side. The sample is wrapped with aluminium foil to avoid the loss of leached PCM during the test, while the top side remains open. Afterwards, the sample is placed on the sample holder, which is attached to a weighing cell continuously measuring the mass of the sample during the test. An electrical radiant heater is placed above the sample, which heats the sample continuously with 50 kW/m². Figure 2 shows the test setup.
For the ignition, a spark plug is additionally placed on top of the sample. It is removed as soon as the sample starts burning. The time to ignition tig is documented for later evaluation. After ignition, the radiant heater is heating the sample for 30 min. Besides the continuous measuring of the weight of the sample, the oxygen content of the air, which is collected via the conical head above the sample, is determined. Based on the oxygen content of the air, the heat release by the sample is calculated. Moreover, the smoke production is determined with a photocell.
Fig. 2
Combustion test of wood by cone calorimetry according to ISO 5660-1 (2015)
×
![]()
Each two samples were prepared for the wood species spruce, beech, poplar and oak, which were either impregnated with clear PCM or a PCM-cement mixture. As a reference, also clear wood samples were tested. The samples had dimensions of 100 mm by 100 mm by 30 mm (length by width by thickness). While the longitudinal material orientation was parallel to the length direction, there was no special alignment of radial and tangential direction.
The PCM uptake of the samples was different depending on the wood species: the beech samples possessed an uptake of about 130–180 kg/m³, the spruce samples of about 140–180 kg/m³, the poplar samples of about 260–280 kg/m³ and the oak samples of 120–150 kg/m³. The samples were conditioned before testing at 23 °C and 50% relative humidity in a climate chamber. The tests were performed in the testing facility of EPH Dresden GmbH, Germany.
2.3.2 Reaction to fire test according to EN ISO 9239-1
In a second test, the combustion behaviour of a construction element was tested in a setup according to EN ISO 9239-1 (2010). The element was a three-layer wood flooring with a PCM-impregnated middle layer of 8 mm thickness. The PCM uptake was about 200 kg of PCM per m³ of wood. The PCM contained cement particles. The top layer and the counteracting lower layer were not impregnated and had thicknesses of 3 mm and 2 mm, respectively. The middle layer and the lower layer were made of spruce. The top layer, which was coated with a commercial lacquer, was made of oak. As a reference, elements without PCM were tested. The reference elements are commercially available.
The specimens had a length of 1050 mm and a width of 230 mm. Since the flooring elements are connected mechanically in practice, these joints have to be considered also in the tests. Thus, for the PCM-impregnated and the reference elements one test with a connection in longitudinal direction and one test with joints in transverse direction were performed. It shall be noted that for a valid result of the fire class of the element according to EN ISO 9239-1 at least three tests for each joint direction are necessary. Thus, the presented results correspond only to an orienting investigation. During the tests, the samples were placed on a fibre cement board of 8 mm thickness.
The detailed testing procedure is given in EN ISO 9239-1 (2010). Essentially, the sample is heated for the first 10 min with a gas burner at one end of the sample over the entire width. Figure 3 shows a view into the test chamber at the beginning of the test. After 10 min, the burner is turned off. The flame propagation at the sample is observed for another 20 min and recorded over the entire testing time (30 min) by means of the flame spread L.
Fig. 3
View into the test chamber of the reaction to fire test according to EN ISO 9239-1 (2010) at the beginning of a test with active gas burner
×
![]()
Moreover, the smoke development is recorded. For achieving a certain fire class according to EN 13501-1 (2018), a certain limit of the critical heat flux CHF as well as the smoke density integral ∫R needs to be complied. The limits for the fire classes according to EN 13501-1 (2018) are summarized in Table 3. The tests were performed in the testing facility of EPH Dresden GmbH, Germany.
3 Results and discussion
3.1 Mechanical behaviour
3.1.1 Bending behaviour
Figure 4 and Table 4 show the evaluation of the results of the bending tests in terms of bending strength fb and bending modulus Eb. It can be seen that, as well known, the three wood species have different strength and elasticity values.
Fig. 4
Results of the bending tests; (a) Bending strength, (b) Bending modulus
×
![]()
Table 4
Properties of the samples of the bending tests (values are mean values, standard deviation and mean value related to mean value „clear cold“)
Wood species/ impregnation/ temperature | ρPCM/wood [kg/m³] | ρwood [kg/m³] | fb [N/mm²] | Eb [N/mm²] |
|---|---|---|---|---|
Spruce | ||||
Impregnated (PCM & cement) cold | 188; 25; - | 452; 13; 1.08 | 105; 8; 1.15 | 14,267; 915; 1.03 |
Impregnated (PCM & cement) warm | 228; 33; - | 443; 25; 1.06 | 88; 5; 0.96 | 13,430; 1222; 0.97 |
Clear cold | - | 419; 90; 1,00 | 91; 9; 1.00 | 13,790; 876; 1.00 |
Clear warm | - | 424; 75; 1.01 | 84; 5; 0.92 | 12,482; 1507; 0.91 |
Beech | ||||
Impregnated (PCM & cement) cold | 204; 18; - | 692; 29; 1.01 | 145; 19; 1.06 | 16,473; 1978; 1.00 |
Impregnated (PCM & cement) warm | 218; 16; - | 692; 16; 1.01 | 126; 7; 0.92 | 15,913; 1055; 0.97 |
Clear cold | - | 686; 14; 1,00 | 137; 9; 1.00 | 16,473; 1153; 1.00 |
Clear warm | - | 687; 12; 1.00 | 125; 5; 0.91 | 16,328; 1332; 0.99 |
Poplar | ||||
Impregnated (PCM & cement) cold | 261; 32; - | 398; 16; 0.96 | 62; 12; 0.84 | 8104; 869; 0.78 |
Impregnated (PCM & cement) warm | 252; 30; - | 386; 32; 0.93 | 58; 15; 0.78 | 8974; 1446; 0.87 |
Clear cold | - | 416; 14; 1.00 | 74; 7; 1.00 | 10,359; 1247; 1.00 |
Clear warm | - | 421; 29; 1.01 | 63; 6; 0.85 | 9609; 1016; 0.93 |
Moreover, the known effect of a reduction of fb and Eb with increasing temperature is observable. In the solid state, the impregnation with PCM results in increasing fb compared to clear wood for beech and spruce. To a lower extent, this is also observable for spruce for Eb. In the liquid state of the PCM, the increase of fb and Eb vanishes. For poplar, both fb and Eb were smaller for the impregnated samples than the clear samples. This was also associated with a more brittle failure behaviour. While the increased brittleness can explain the lower fb values of the impregnated poplar samples, it does not explain the lower Eb values. The Eb values of the cold impregnated poplar samples are also smaller than of the warm impregnated samples. The reason for this effect remains unclear.
For the practical application, it can be assumed that at least for the common species for construction purposes spruce and beech, the mechanical properties are not significantly influenced by the impregnation with PCM. This also corresponds to findings in Saavedra et al. (2021). For poplar, a reduced PCM uptake might result in more favourable mechanical properties and less brittleness.
3.1.2 Surface hardness
Figure 5 shows the results of the surface hardness tests as mean values of the Brinell hardness of 8 tests each. As well known, the surface hardness depends on the wood species, which results from different density and cell structure. Moreover, a temperature dependency is observable. The warm material has a lower surface hardness than the cold material. This also applies to the clear reference samples. The largest increase in surface hardness due to solid PCM is observable for beech and poplar. At least with the applied amount, the addition of cement to the PCM does not influence the surface hardness. In the liquid state of the PCM, the increase in surface hardness is lost and values corresponding to the clear reference samples are achieved. Especially in the PCM impregnated samples, the indentation is substantially recovering during the storage at elevated temperature in the climate chamber.
Fig. 5
Brinell hardness of clear and PCM-impregnated samples with the PCM in the solid or liquid state
×
![]()
For a practical application, these results are advantageous. At least for the time in service where the PCM is in the solid state, the surface hardness is compared to clear wood significantly larger and, thus, the surface is more resistant against indentations. Moreover, the PCM supports recovery of indentations by the phase change and the associated density changes.
3.2 Combustion behaviour
3.2.1 Cone calorimetry according to ISO 5660-1
The combustion behaviour of PCM-containing wood was investigated with two different tests. Figure 6 shows the results of the cone calorimetry for the different wood species and impregnations. It can be seen that the mass of the samples decrease quasi-linearly during the combustion process. Only shortly before the end of the combustion an increasing mass loss rate is observable. Schartel and Hull (2007) explain the second peak by “cracking of the char or an increase in the effective pyrolysis” due to the rear boundary conditions.
Fig. 6
Results of the cone calorimetry tests according to ISO 5660-1 (2015)
×
![]()
The mass loss rate during the undisturbed combustion is approximately the same for the impregnated and the clear reference samples. Except for beech, the mass loss rate is also similar for all the tested wood species. Nevertheless, it can be seen that the combustion time is longer for the PCM-containing samples due to the higher total mass and has for some samples not ended at the end of the test after 1800 s. The heat release is larger for the PCM-impregnated samples than for the clear reference samples. While the heat release rate HRR during undisturbed combustion is, except for beech, approximately the same for the clear samples of the different wood species, it depends for the impregnated samples on the PCM uptake. Corresponding to the PCM uptake, HRR of the PCM-impregnated samples is the lowest for oak and the highest for poplar. There is no considerable difference due to the addition of cement observable. The somewhat higher HRR for the cement-containing samples of beech can be explained by a higher PCM uptake in this case compared to the impregnated samples without cement. The smoke production is much larger for the impregnated samples. During the undisturbed combustion, in some cases no smoke development is observable for the clear reference samples, which indicates a complete combustion. Contrary, the smoke production of the impregnated samples is high over the complete testing period. This indicates an incomplete combustion of the PCM.
Based on the measured quantities, indicators characterizing the fire ignition hazard and the provided energy for fire propagation can be calculated (Schartel and Hull 2007). A first indicator is the ratio of the peak heat release rate PHRR and the time to ignition tig. Both values as well as the ratio are given in Table 5. It can be seen that tig is substantially smaller for the impregnated samples compared to the clear reference samples while the PHRR values are considerably larger. The ratio PHRR / tig is approximately four to five times larger for the impregnated samples, which means that the PCM acts as fire accelerant.
Table 5
Characteristics of samples (2 for each parameter combination) tested with cone calorimetry according to ISO 5660-1 (2015) (in parentheses: mean values normalised by mean values of clear spruce)
Sample | ρPCM/wood [kg/m³] | ρwood [kg/m³] | PHRR [kW/m²] | tig [s] | PHRR / tig [kW/ (m²s)] | THE [MJ/m²] | ML [g] | THE / ML [MJ/ (g m²)] |
|---|---|---|---|---|---|---|---|---|
Spruce with PCM & cement | 185; 140 | 580; 577 (1.00) | 271; 254 (1.63) | 6; 5 (0.30) | 45; 51 (5.52) | 332; 303 (2.04) | 154; 159 (1.18) | 2.1; 1.9 (1.72) |
Spruce with PCM | 149; 144 | 574; 571 (0.99) | 244; 260 (1.57) | 5; 6 (0.30) | 49; 43 (5.29) | 309; 311 (1.99) | 161; 162 (1.22) | 1.9; 1.9 (1.63) |
Clear spruce | - | 577; 576 (1.00) | 170; 151 (1.00) | 20; 17 (1.00) | 9; 9 (1.00) | 153; 158 (1.00) | 133; 131 (1.00) | 1.1; 1.2 (1.00) |
Beech with PCM & cement | 178; 176 | 664; 661 (1.15) | 324; 416 (2.30) | 8; 6 (0.38) | 40; 69 (6.32) | 403; 414 (2.63) | 210; 209 (1.59) | 1.9; 2.0 (1.66) |
Beech with PCM | 131; 134 | 723; 714 (1.25) | 252; 245 (1.55) | 8; 8 (0.43) | 31; 31 (3.57) | 370; 364 (2.36) | 209; 207 (1.57) | 1.8; 1.8 (1.50) |
Clear beech | - | 689; 686 (1.19) | 177; 185 (1.13) | 24; 23 (1.27) | 7; 8 (0.89) | 212; 211 (1.36) | 168; 167 (1.26) | 1.3; 1.3 (1.07) |
Poplar with PCM & cement | 283; 280 | 382; 381 (0.66) | 330; 342 (2.09) | 6; 5 (0.30) | 55; 68 (7.09) | 456; 440 (2.88) | 168; 165 (1.26) | 2.7; 2.7 (2.29) |
Poplar with PCM | 269; 267 | 339; 339 (0.59) | 354; 374 (2.27) | 4; 5 (0.24) | 89; 75 (9.39) | 427; 416 (2.71) | 156; 153 (1.17) | 2.7; 2.7 (2.31) |
Clear poplar | - | 400; 384 (0.68) | 175; 174 (1.08) | 12; 10 (0.59) | 15; 17 (1.84) | 148; 137 (0.92) | 105; 101 (0.78) | 1.4; 1.4 (1.17) |
Oak with PCM & cement | 125; 124 | 664; 637 (1.13) | 240; 247 (1.51) | 6; 6 (0.32) | 40; 41 (4.66) | 268; 266 (1.72) | 153; 152 (1.15) | 1.8; 1.8 (1.49) |
Oak with PCM | 147; 142 | 707; 703 (1.22) | 199; 211 (1.28) | 9; 7 (0.43) | 22; 30 (3.01) | 305; 309 (1.97) | 174; 179 (1.34) | 1.7; 1.7 (1.48) |
Clear oak | - | 611; 620 (1.07) | 171; 174 (1.08) | 18; 19 (1.00) | 10; 9 (1.07) | 153; 153 (0.99) | 139; 140 (1.06) | 1.1; 1.1 (0.93) |
A second indicator is the ratio of the total heat release at the end of the test THE and the mass loss ML of the sample. This ratio is for the impregnated samples less than twice as large as for the clear reference samples. Thus, for the propagating fire the PCM has a less significant influence than for ignition.
As a conclusion, it can be stated that unprotected PCM-containing wood ignites faster and provides more heat during combustion for fire propagation than clear wood. However, the results of the cone calorimeter tests cannot directly be transferred to real fire events due to a large dependency on length scale (Schartel and Hull 2007). Thus, further tests are necessary.
3.2.2 Reaction to fire test according to EN ISO 9239-1
As described in Section 2.3.2, fire tests according to EN ISO 9239-1 (2010) on wood flooring elements containing PCM were also performed. Table 6 shows the achieved values of the fire tests with the flooring samples. Besides the CHF and ∫R values, also the maximum flame spread Lmax and the maximum light absorption Rmax are given. The flame spread propagated in all samples until the end of the test after 1800 s.
Table 6
Results of the reaction to fire test according to EN ISO 9239-1 (2010) on wood flooring with PCM
Value | Sample without PCM (longitudinal joint) | Sample without PCM (transverse joints) | Sample with PCM (longitudinal joint) | Sample without PCM (transverse joints) |
|---|---|---|---|---|
CHF [kW/m²] | 3.98 (Dfl.) | 3.71 (Dfl.) | 4.58 (Cfl.) | 4.67 (Cfl.) |
∫R [min %] | 42.1 (s1) | 50.4 (s1) | 175.7 (s1) | 110.7 (s1) |
Lmax [mm] | 480 | 498 | 440 | 434 |
Rmax [%] | 8.7 | 11.1 | 20.7 | 15.8 |
Figure 7 shows diagrams of the flame spread vs. time and the characteristic relation for determining CHF. The input quantity for the characteristic relation in Fig. 7b is the maximum flame spread Lmax, which is also shown in Fig. 8. From Table 6 and Fig. 7, it can be seen that CHF of the samples containing PCM has a higher value than the reference samples without PCM. Thus, the elements containing PCM have with Cfl. a more favourable fire class than the reference elements, which achieve only Dfl.. This probably unexpected result can be explained with the higher heat capacity of the PCM-containing elements, which initially absorb the heat and cool the flamed top layer. This is possible because the PCM is not directly flamed but protected by the not impregnated top layer. Thus, the flash point of the PCM at 177 °C is not instantaneously reached like in the tests with the cone calorimeter previously described.
Fig. 7
Results of the reaction to fire tests according to EN ISO 9239-1 (2010) on flooring elements; (a) Flame spread vs. time, (b) heat flux vs. flame spread (characteristic relation)
×
![]()
Fig. 8
Charring as indicator for flame spread on the specimens after the reaction to fire tests according to EN ISO 9239-1 (2010)
×
![]()
It should be considered that the PCM was in its solid state at the beginning of the test and, thus, a significant additional heat uptake was possible. It remains to be tested, how the behaviour is when the PCM is in its liquid state right from the beginning of the test and, thus, the cooling effect is not existing. It can be assumed that in this case also the PCM-containing elements achieve only fire class Dfl..
The second characteristic for the definition of the fire class is the smoke development. As Table 6 shows, the PCM-containing elements behave less favourable with stronger smoke development in terms of ∫R and Rmax compared to the reference samples. Still, the values are considerably below the acceptable values as given in Table 3 for classification s1, which, thus, both the PCM-containing samples and the reference samples reach.
As a conclusion, it can be stated that a protection of the PCM-containing wood improves the fire behaviour substantially and can compensate the fire accelerating properties of the PCM when directly flamed.
3.3 Construction elements
The previously described investigations were conducted to evaluate whether it is possible to design PCM-containing construction elements, which meet crucial building requirements. Figure 8a shows a flooring element as an example of a non-structural construction element. For structural elements, more requirements need to be met. A key aspect are the mechanical properties. The results of the bending tests showed that the PCM impregnation does not impair the strength and the modulus for commonly used wood like spruce and beech. At the state of the art, timber elements are produced by glue lamination. Although respective investigations are not presented here, preliminary results as published in Hartig and Haller (2023) indicate that the gluing of PCM-impregnated timber is possible with a sufficient quality of the glue line.
An exemplary glue-laminated timber element is shown in Fig. 9b. The element takes the aforementioned findings into account. The element consists of a core of glue-laminated spruce without PCM. The core remains without PCM since the heat penetration over the day is assumed to be limited and the PCM would not be activated at this position. On the side and the bottom faces of the core, PCM-impregnated spruce boards are laminated for locally increased heat capacity. In the case, that the side faces are not ventilated in the room, the PCM-containing layer could also be placed only at the bottom face. As cover and PCM-free sealing layer, a thin board, which could be also a veneer, of oak is glue-laminated. Oak is hard to impregnate and serves as a barrier against leakage of the PCM. On the other hand, it also has decorative value. If the latter is not intended or necessary, other wood species, which possess a bad impregnability, like spruce could be used. However, it should be considered that knots and cracks could impair the barrier against leakage.
The PCM-free cover layer also improves the fire behaviour. As the respective investigations showed, the direct flaming of the PCM-containing pieces should be prevented. Then, the PCM could delay the fire spread by absorbing heat and, thus, cooling the flamed surface of the element. The general concept might be applied also to other engineered timber elements like cross-laminated timber.
Fig. 9
PCM-containing construction elements; (a) Wood flooring with a PCM-containing middle layer as non-structural element, (b) Glue-laminated timber element with PCM impregnation in the outer layer and a PCM-free cover layer as structural element
×
![]()
4 Conclusion
Different investigations with regard to an application of PCM-containing wood in structural and non-structural timber and wood elements were presented. The investigations regarding the mechanical behaviour showed in terms of bending strength and modulus that the mechanical properties are not impaired by the PCM impregnation if the PCM uptake is moderate. For poplar, which allows for a very high PCM uptake an embrittlement was observed, especially when the PCM was in its solid state. The surface hardness, which has a certain importance for flooring elements is substantially improved when the PCM is in its solid state. In the liquid state, the surface hardness is at same level as in wood free of PCM. Nevertheless, this is advantageous since for a considerable share in time the PCM is assumed to be in its solid state where the proneness for indentations is reduced.
The combustion tests with cone calorimetry showed that PCM-containing wood ignites substantially faster than wood without PCM. Moreover, it also provides more heat during combustion for fire propagation depending on the existing PCM uptake. Thus, PCM-containing wood should not be applied unprotected if there is a substantial risk for fire development in a building. On the other hand, the reaction to fire of PCM-containing wood can be substantially improved by a proper protection from direct contact to fire. The crucial point is to keep the temperature at the PCM-containing wood as long as possible below the flash point of the PCM.
Finally, a multi-layer flooring element with PCM-containing wood in the middle layer as an example for a non-structural construction element and a glulam element with partial impregnation with PCM as an example for a structural construction element were developed, which show options for the practical application of PCM-impregnated wood for construction purposes.
Acknowledgements
The authors gratefully acknowledge the financial support from German Federal Ministry of Food and Agriculture for the FNR research project “PCM-WOOD: Impregnation of wood with phase change materials for latent heat storage in buildings” (grant number 22014318). The authors also thank the project partner ter Hürne GmbH & Co. KG, Südlohn, Germany for collaboration.
Declarations
Financial interests
Technische Universität Dresden applied for a patent related to parts of the content of this paper with author Jens U. Hartig as inventor. Apart from that, the authors declare they have no financial interests.
Non-financial interests
None.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.