A study on normal incident sound absorption characteristics of Japa- nese traditional cedar board, yakisugi Akiko Sugahara 1 , Kentaro Okamura 2 , Yasuhiro Hiraguri 3 Kindai University 3-4-1 Kowakae, Higashiosaka, Osaka, 577-8502, Japan Noboru Yasui 4 , Chihiro Kaku 5 Team SAKURA #501, 2-21-10 Yoyogi, Shibuya, Tokyo, 151-0053, Japan

ABSTRACT Yakisugi is made by charring the surface of a cedar board to form a thin carbonized layer, which is said to improve durability, weatherability and fire resistance. It is a traditional Japanese exterior material that has become popular overseas due to its performance and visual texture. Recently, it has been attracting attention for its use as an interior material, but its basic performance is still unknown. In this study, we focus on the sound absorption characteristics of yakisugi as one of the investigations to understand its performance as an interior material. The carbonized layer of yakisugi is expected to have sound absorbing performance because of its porous nature. As a basic investigation, the sound absorption characteristics at normal incidence are measured using an impedance tube and compared with the carbonized layer conditions. It is found that yakisugi has sound absorption per- formance in the high frequency range. Moreover, the characteristics are depending on the thickness and surface properties of the carbonized layer.

1. INTRODUCTION

Japan is a country with a large forest area, and there have been many wooden buildings since ancient times. Timber has been used for structural materials, interior and exterior materials, and many other purposes.

“ Yakisugi ” is a traditional exterior material commonly used in Japan, especially in the middle and western parts of the country [1]. This is a Japanese cedar board whose surface is charred to form a thin carbonized layer (see Figure 1). It has been said that the formation of this carbonized layer pro- tects the surface of the cedar board and improves the durability, weatherability, fire and water re-

1 sugahara@arch.kindai.ac.jp

2 okamura@arch.kindai.ac.jp

3 hiraguri@arch.kindai.ac.jp

4 yasui@teamsakura.jp

5 chihiro.k@teamsakura.jp

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

sistance, and biological resistance to mold and decay fungi. Recently, it has become popular in west- ern countries due to its beautiful texture and appearance. Some researchers showed the possibility of applying the traditional Japanese yakisugi construction method to European building materials, such as pine and oak, and the physical properties of the charred surface as the exterior materials are inves- tigated [2-6]. However, there are few such studies on the physical properties of yakisugi , and much remains to be learned. In addition to the use as exterior materials, demand for interior materials has been increasing in recent years. However, its performance as an interior material has not been well studied and is little understood.

In this study, experimental investigations are conducted to determine its potential use as an interior material, especially as sound-absorbing material. The porousness of yakisugi ’s carbonized layer may have the sound absorption performance. In this paper, the sound absorption characteristics of yakisugi at normal incidence are investigated using the impedance tube. The effects of various parameters of yakisugi , such as the thickness of the carbonized layer, the charred surface, and the presence or ab- sence of knots and cracks, on the sound absorption coefficient are examined.

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

Section 2 provides an overview of the yakisugi production process and case studies. Section 3 describes experiments using the impedance tube. Finally, Section 4 summarizes this study.

(a) Exterior wall of yakisugi in Hiroshima, Japan (b) Yakisugi surface tecture Figure 1: Appearance examples of yakisugi

2. MANUFACTURE OF SPECIMENS

2.1. Surface carbonization procedure

There are two types of yakisugi production methods: one is the traditional hand-burning method, and the other is machine-burning in a factory. In this study, a traditional hand-burning method called “ sankaku-yaki ” is used to make samples. The procedure is illustrated in Figure 2. (1) Three cedar boards are assembled and bound with straw rope to form a triangular chimney. (2) Wasted paper placed at the bottom of the chimney is ignited. (3) Then, the board surface is evenly baked due to the chimney effect. (4) When the surface is sufficiently browned, cut off the straw rope and let it cool down by exposing to the air and sprinkling water over it. The burning duration is about 5 minutes at most, depending on the moisture content of the board and the thickness of the carbonized layer. More- over, when the wood is heated, and the temperature exceeds about 200 ℃ , gasification occurs. The combustible gases produced during this process combine with oxygen in the air to cause combustion. This causes a volume reduction, resulting in shrinkage of the material and a distinct pattern of cracks on the carbonized layer. Finally, (5) in order to measure the sound absorption coefficient at normal incidence using a 30 mm-diameter impedance tube, yakisugi board was hollowed out with a hole saw.

Note that the 10-cm-diameter circles pre-cut in Figure 2(5) are for use in other experiments and are not relevant to this study.

In this study, 170 mm × 2,000 mm Japanese cedar boards from Nara are charred using “sankaku- yaki.” The thicknesses are approximately 21 mm and 35mm. Although it is common practice to char the sap-side of the board, we prepared samples carbonized on both the sap-side and the heart side to examine the difference in charring surfaces. Because the sound absorption performance of yakisugi is presumed to be highly dependent on the thickness of the carbonized layer, samples with three different carbonized layer thicknesses are prepared; “thin,” “normal,” and “thick.” The charring du- ration determines the depth of the carbonized layer. Therefore, the thickness of the carbonized layer “thin,” “normal,” and “thick” are formed by varying the duration from the time the fire blew out from the top of the chimney until the rope is cut and dismantled: 1.5 (thin), 2.5 (normal), and 3.5 minutes (thick).

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

(1) Form the trian- gular chimney

(2) Ignition from the bottom of the chimney

(3) Flame eruption from the top, start measuring time

(4) Dismantle, ex- posing to the air and watering

(5) Hollow out to sample size

Figure 2: Steps from charring the cedar board to preparation of yakisugi samples

2.2. Case study

At first, whether it has sound-absorbing properties is examined in the first place by comparing the cases with and without the carbonized layer. Next, the effects of the board and the carbonized layer thicknesses and the surface irregularity, such as knots, cracks, and peelings are investigated. Sixteen yakisugi samples are prepared according to the procedure in Subsection 2.1. The case study conditions and appearance are shown in Tables 1 and 2. Of these, nine cases are made of 35 mm-thick boards, and seven cases are made of 21 mm-thick boards. Geometrical parameters of 35-mm- and 21-mm- thick samples are shown in Tables 1 and 2, respectively. Of the 35-mm-thick cases, 09_F is the one where the carbonized layer was cut away with a cutter.

The thickness of the carbonized layer is measured with a digital caliper. Because of uneven firing, the thicknesses are measured at four locations. The averaged values are shown in Tables 1 and 2: 1.1- 2.7 mm for the “thin” cases, 2.5-4.0 mm for the “normal” cases, and 4.9-5.9 mm for the “thick” cases. Generally, the thickness of the carbonized layer is larger in the order of “thick,” “normal,” and “thin” with longer charring duration. Note that 08_TN and 17_TN have relatively thick layers even though they are “thin” cases because of uneven firing.

Because the inhomogeneity of the carbonized layer may affect the sound absorption performance, the relationship between them and the sound absorption coefficient will also be discussed.

Samples with knots are marked with a “ ○ ” in the Knot column. The Cracks column represents the interval and the maximum depth and width of the cracks. A “-“ in the Interval column indicates a case with only one crack or a case with no carbonized layer. The specimens with a “ ○ ” in the Peelings column peeled off at the edges because of the carbonized layer softness. Overall, the cracks are larger in those with the thicker carbonized layer, knots, or peelings.

Table 1: Case studies of 35-mm-thick samples. The number in () in the Board column of 09_F in- dicates the real thickness.

Thickness Charred side Knot

Cracks Peel- ings Board Carbonized

Name

layer Interval Depth Width

01_TK

5.9 Sap - 4, 7 2.2 0.5 - 02_TK_P 5.2 Heart - 7.5 5 2.4 〇 03_M 3.2 Sap - 7.5 2.7 1.5 - 04_M_K 3.4 Sap ○ 8 2.5 1 - 05_M_P 4.0 Heart - 8 3.5 1.5 〇 06_TN 1.8 Sap - 5 1.5 0.7 - 07_TN_K 1.1 Sap ○ - 1.4 0.7 - 08_TN 2.4 Heart - 8 1 0.7 - 09_F 35(30) 0 Heart - - - - -

35

Table 2: Case studies of 21-mm-thick samples

Thickness Charred side Knot

Cracks Peel- ings Board Carbonized

Name

layer Interval Depth Width

4.9 Heart - - 4 1.4 〇

11_TK_P

12_TK_P 5.2 Heart - 7.5 4 1.4 〇

13_M_P 3.5 Sap - 6.5 1.2 1.3 〇

14_M 2.5 Heart - 6 2.5 1.7 -

21

15_TN_K 2.0 Sap 〇 - 3.5 1.9 -

16_TN 1.7 Sap - 7.7 1.2 1.7 -

17_TN 2.7 Heart - 7.5 2.5 2.0 - 3. MEASUREMENTS OF ABSORPTION COEFFICIENT AT NORMAL INCIDENCE

3.1. Measurement settings

The sound absorption coefficient at normal incidence is measured using a self-made acrylic im- pedance tube shown in Figure 3. The inner diameter of the tube is 30 mm. The measurements are conducted according to the transfer function method specified by the ISO standard 10534-2 [6]. The settings of the impedance tube are shown in Table 3. Because the specimens do not form a perfect cylinder due to their creation’s precision, Vaseline and clay are applied to the surrounding area to fill the gap with the impedance tube. Recordings are made on a PC through an audio interface (Fireface UC, RME) with a sampling rate of 48 kHz and a quantization bit rate of 16 bit.

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

Figure 3: Self-made impedance tube

Table 3: Self-made impedance tube settings Tube Diameter 30 mm Microphone spacing 25 mm Frequency range 500 Hz – 6 kHz Source Loudspeaker FF105WK / Fostex Source signal White noise (60 seconds) Receiver Pair microphone QTC-50mp / Earthworks

3.2. Results and discussion

3.2.1. The cedar board thickness

Figure 4 shows the results of 35-mm- and 21-mm-thick samples with no knots and peelings. The cases in each figure have similar carbonized layer thicknesses, “normal” and “thin,” respectively. The results show that there is not much difference between the sound absorption coefficients of the two data in each figure. Therefore, one can see that the thickness of the cedar board itself does not affect the sound absorption performance.

(a) “Normal” carbonized layer cases (b) “Thin” carbonized layer cases Figure 4: Sound absorption coefficient of samples with different cedar board thickness

10 09 Zos 202 —0.M —13MP 500 1500 2500 3500 Frequency [Hz] 4500 5500

3.2.2. The presence and thickness of the carbonized layer

Figure 5 shows the results of 35-mm-thick specimens without knots or peelings; four specimens with different thicknesses of the carbonized layer and 09_F with no carbonized layer. The case 09_F has a peak at around 3.5 kHz but overall has a lower sound absorption coefficient than the other samples. This peak may be due to the installation condition of this sample. For the other four cases with the carbonized layers, the sound absorption coefficient generally increases with frequency. In addition, when focusing on the thickness of the carbonized layer, the sound absorption coefficients of 01_TK and 08_TN, which have a larger thickness, are generally higher than those of 03_M and

10 09 Tos B07 206 05 E04 202 ou 00 —06_TN —16_1N 500 1500 2500 3500 Frequency [Hz] 4500 5500

06_TN, which are relatively thinner. These are characteristics of porous sound-absorbing materials, and this suggests that the presence of the carbonized layer can bring this about. Figure 6 shows the relationship between the average sound absorption coefficient per one-octave band and the thickness of the carbonized layer for all cases. The linear approximation curve and the correlation coefficient are presented in each subfigure. The correlation coefficient increases with increasing frequency. Thus, the effect of the carbonized layer thickness on sound absorption performance is more signifi- cant at higher frequencies.

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

Figure 5: Sound absorption coefficient of 35-mm-thick samples

with and without the carbonized layer

ROG R070 ‘Carbonized layer thickness [mum] ° 2 4 Carbonized layer thickness [mum] 2 4 6 Carbonized layer thickness fm]

(a) 1 kHz (b) 2 kHz (c) 4 kHz Figure 6: Relationship between the carbonized layer thickness and average sound absorption coefficient at each frequency band (a) 1 kHz, (b) 2 kHz, (c) 4 kHz

3.2.3. Depth and width of cracks

Figure 7 shows the relationship between the average sound absorption coefficient per one-octave band and the maximum depth and width of cracks. 𝑅𝑅 D and 𝑅𝑅 W are the correlation coefficients be- tween each parameter and the sound absorption coefficient. The results show that the maximum width of the cracks exhibits a small correlation with the sound absorption in all bands. On the other hand, crack depth affects the sound absorption coefficient in the 4 kHz band, and this suggests that the depth of the cracks affects the sound absorption characteristics in the high frequencies.

— 01 TK —03_M —— 06_TN ——08_TN. | | | | I I | 1 ! 5001500 2500 3500 4500500 Frequency [Hz]

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW

(a) 1 kHz (b) 2 kHz (c) 4 kHz Figure 7: Relationship between between the maximum depth and width of cracks and average sound absorption coefficient at each frequency band (a) 1 kHz, (b) 2 kHz, (c) 4 kHz

3.2.4. The knots and peelings

Figure 8 shows the sound absorption coefficients of the cases with (a) the knot and (b) peelings. The three cases in Figure 8(a) exhibit a peak around 1.2 kHz due to the knots, but there are cases where the peak appears even without a knot. However, the cases with knots have large cracks. Thus, they show high sound absorption in the higher frequency range. On the other hand, Figure 8(b) shows that the values with peelings are higher than that of 01_TK without peeling, especially in the higher frequency range. This suggests that inhomogeneity of surface properties such as peel- ing may affect the sound absorption performance.

TT) Dep] [R028 Ror © Width : a Ry=0.09 RyH0.34 lo With Bry ee oO 7 ° <2 as —~1 ©, ° °o 4 —

(a) With knots (b) With peelings Figure 8: Sound absorption coefficient of the samples with knots and peelings 4. CONCLUSION

10 09 Tos B07 206 05 E04 202 ou 00 500 1500 2500 3500 Frequency [Hz] 4500 5500

In this study, the sound absorption coefficient at normal incidence is measured assuming the use of yakisugi as an interior material. The results show that the presence of a carbonized layer improves the sound absorption performance due to its characteristics as a porous material. The thicker the car- bonized layer, the higher the sound absorption coefficient. It is also suggested that the inhomogeneity of surface properties affects the performance. The cases with thicker carbonized layers and those with

09 Dos B07 206 Bos E04 202 ou 00 —04_ MK —07INK —15_INK 500 1500 2500 3500 Frequency [Hz] 4500 5500

knots or peelings are found to have larger cracks, which improves the sound absorption performance, especially in the high-frequency range.

One issue to be addressed in the future is the problem of surface coatings, and it is necessary to consider its effect on the sound absorption performance, and what kind of coating is suitable. Addi- tionally, the reverberation sound absorption coefficient, commonly used in room acoustics, will be measured. 5. ACKNOWLEDGEMENTS

This research was supported by Research and Practice Grants No. 2109 from Housing Research Foun- dation JUSOKEN. 6. REFERENCES

1. Okamura, K., Yasui, N., Kaku, C., Koshihara, M., Imamoto, K. & Oshima, K. A research on yakisugi – Performance evaluation and feasibility study for dissemination – (in Japanese). Jour- nal of Housing Research Foundation “Jusoken”, No. 44, 13-24, (2017). 2. White, R. H. & Dietenberger, M. A. Wood Handbook, Chapter 18: Fire Safety of Wood Con- struction. United States Department of Agriculture Forest Service, Madison, WI, USA (2010). 3. Kymäläinen, M., Turunen, H., Čermák, P., Hautamäki, S. & Rautkari, L. Sorption-related char- acteristics of surface charred spruce wood, Materials , 11(11) , 2083, 1-15(2018). 4. Kymäläinen, M., Hautamäki, S., Lillqvist, K., Segerholm, K. Surface modification of solid wood by charring, Journal of Materials Science , 52(10), 1-9 (2017). 5. Hasburgh, L., Zelinka, S. L., Bishell, A. & Kirker, G. Durability and fire performance of charred wood siding (sho sugi ban), Forests, 12(9), 1262, 1-13 (2021). 6. ISO 10534-2: 1998, Acoustics – Determination of sound absorption coefficient and impedance in impedance tubes – Part 2: Transfer-function method., 1998.

‘inter.a 21-24 AUGUST SCOTTISH EVENT CAMPUS ? O? ? GLASGOW