CORK AS A LIGHTWEIGHT PARTITION MATERIAL. ECONOMICAL AND ACOUSTICAL ANALYSES

Since 10.10.97: 

Civil Engineering Department, College of Engineering, University of Porto, Portugal
Tel.: 351.225081931, Fax: 351.225081.940/525

SUMMARY

The purpose of this paper is to present an indigenous material (cork) in a construction technology innovation. The building construction development in Portugal has been changed by the very slow but gradual move to a general use of partitions made with increasingly light materials from the traditional heavy interior walls. A socioeconomic analysis of the Portuguese situation as it is concerned with the lodging policies and building construction industry, is briefly presented and confronted to 37 other countries using some economic parameters. The main goal of this paper is to present a study of the acoustical behavior of lightweight partitions, especially those usually called sandwich-type and its experimental application to a new and very specific kind of partition made of gypsum board and cork, a traditional material in Portugal. This newly developed sandwich lightweight partition is presented and acoustically characterized. A new and simple mathematical model is presented to evaluate the sound isolation of this kind of partition. The results obtained in reverberant chamber tests are presented as well as the comparisons with the predicted values using the new model proposed. Different single and double wall types were tested giving STC values up to 44 dB.
 

KEYWORDS: Partitions; Sound Isolation; Acoustics; Cork; Sandwich; Lightweight
 

INTRODUCTION

Portugal is the world largest producer and exporter of cork. Nevertheless cork has never been widely used in Portugal in the building construction industry as a sound isolation material. Its main use in this field has been as a thermal insulator or as an acoustic absorptive material.

Cork is a natural product that constitutes the outer bark of a tree named the cork-oak (Quercus suber L), that is disseminated all over Portugal (mainly in the southern area). The product is formed by the grouping of cells, in successive layers, each layer being the result of one year's growth. The first cork stripping is carried out only at the end of 20/25 years of life of the tree. The cork stripping is made every nine or ten years. The lifetime average of a cork-oak is about 170 years.

Cork is a unique material, thanks to its low specific weight, great elasticity, flexibility and durability, its impermeability to liquid and gases, its resistance to wear and fire, its high mechanical resistance and dimensional stability and resistance to reactive agents, microorganisms, etc.

The world cork production (1995) is about 377 000 tons per year and Portugal is the first world producer with 51 %. The second world producer is Spain with 26 %. Italy with 7 % is third and Morocco with 6 % is fourth.

The main goal of this paper is to present the acoustical behavior of a newly designed lightweight partition made of gypsum board and cork.
 

RESEARCH JUSTIFICATION

The main use sought for the type of partitions described in this paper is as interior walls in dwellings or office buildings.

The traditional interior wall type in the Portuguese building construction industry has been and still is, the heavy 7 cm width brick (hollow) wall with 3 cm plaster. This type of solution is not cheaper, regarding total cost (materials used and labor costs), than the proposed lightweight partition. However the changing from the traditional heavy walls has been slowly due to the very low labor costs in the Portuguese building construction industry (see Table 1).

The building construction development in Portugal has not followed other western European countries and has been changed by the very slow and gradual move to the general use of partitions made with light materials.

The use of lightweight partitions in the building construction is the standard rule in many western and industrialized countries. This is due mainly to the importance that the labor costs have in the final cost of the building. In Portugal the labor costs are still very low (see Table 1) and this is one reason not to have the same changing move as seen in other countries since many years ago.

In Table 1 is seen that the skilled and unskilled labor costs in Portugal are only higher than some countries in Eastern Europe, Africa, Middle East and Asia. We can note than, for instance, Portugal had in 1991 labor costs 7 to 8 times higher than those in the USA, or 7 to 9 times higher than the ones in Nordic European countries 3. These values cannot be directly compared due to the difference of levels of life but they can give a clear picture for future developments in this field.

Another aspect to justify the idea that lightweight partitions will be increasingly popular in Portugal is the present costs per square meter for housing. Although the way of life and salaries in Portugal are quite inferior to many other western countries, the housing costs per square meter are not so different that those from these countries. In Table 1 we can also see that the costs per square meter for single family housing and apartments or flats are still very low. However the highest values are now just only around 3 times higher than the ones in Portugal.

Building construction in Portugal is not yet as optimized and rationalized than in other countries in Europe or North America. Very low labor costs do not imply a correspondent low final costs per square meter for housing. The use of composite cork and gypsum board in partitions, with a high labor rentability, can then present a lower final cost than the traditional brick wall. This lowering of costs will naturally push down the square meter housing cost in Portugal. As we see in Table 1, the square meter costs in the USA are even lower than in Portugal when labor costs are seven times higher.

Therefore these socioeconomic parameters indicate the interest on having a less expensive and highly effective lightweight partition.
 

DESCRIPTION OF PARTITIONS TESTED

Five different sandwich partitions were tested in this study - 4 single and 1 double (Figure 1):

Partition D1:	Face layers Core  Total thickness and weigh	10 mm gypsum board (each face) 6 mm agglomerated composition cork 26 mm - 16 kg/m2
Partition D2:	Face layers Core Total thickness and weigh	10 mm gypsum board (each face) 6 mm rubbercork 26 mm - 20 kg/m2
Partition D3:	Face layers Core Total thickness and weigh	10 mm gypsum board (each face) 40 mm agglomerated composition cork 60 mm - 20 kg/m2
Partition D4:	Face layers Core Total thickness and weigh	10 mm gypsum board (each face) ISO500 (2 layers of 25 mm agglomerated composition cork with a 3 mm thick rubbercork sheet in between) 73 mm - 24 kg/m2
Partition DD1/2:	Partition D1 (26 mm) + airspace (100 mm) + Partition D2 (26 mm) Total thickness and weigh 152 mm - 36 kg/m2 (since there was no symmetry in this partition it was tested regarding the sound transmission through both ways of emission named 1 and 2 (DD1/2-1 and DD1/2-2, see Figure 1)

	Skilled LABOR costs (/hour)				Unskilled LABOR costs (/hour)				Single Family housing (cost/m2)				APARTMENTS/flats (cost/m2)
COUNTRY	1987		1991		1987		1991		1987		1991		1987		1991
	NC	US $	US $	NC	NC	US $	US$	NC	NC	US $	US $	NC	NC	US $	US $	NC
Australia	21	14.0	-	-	20	13.3	-	-	375	250	-	-	710	473	-	-
Austria	-	-	25.4	270	-	-	20.7	220	-	-	-	-	-	-	1240	13200
Bahrain	0.62	1.6	-	-	0.45	1.2	-	-	-	-	-	-	-	-	-	-
Belgium	795	20.4	-	-	645	16.5	-	-	-	-	1058	33k	-	-	1026	32k
Canada	22.5	16.7	-	-	19.5	14.4	-	-	420	311	-	-	410	304	-	-
Cyprus	-	-	10.4	4.48	-	-	8.4	3.64	-	-	394	170	-	-	348	150
Czechoslovakia	-	-	1.5	40.3	-	-	1.1	30.3	-	-	-	-	-	-	-	-
Denmark	-	-	28.0	163	-	-	28.0	163	-	-	1030	6k	-	-	1717	10k
Egypt	3.25	4.6	-	-	0.98	1.4	-	-	-	-	-	-	275	393	-	-
Finland	-	-	30.6	112	-	-	21.0	76.7	-	-	-	-	-	-	1412	5150
France	-	-	18.6	95.9	-	-	13.8	71.0	-	-	989	5090	-	-	779	4010
Germany	-	-	15.1	22.8	-	-	-	-	-	-	819	1240	-	-	1254	1900
Greece	-	-	15.0	2391	-	-	10.8	1733	-	-	1001	160k	-	-	826	132k
Holland	37.2	17.7	-	-	-	-	-	-	1200	571	-	-	1100	524	-	-
Hong Kong	29.5	3.8	-	-	19	2.4	-	-	-	-	-	-	2250	288	-	-
Ireland	-	-	12.2	6.93	-	-	10.2	5.79	-	-	529	300	-	-	846	480
Italy	-	-	25.3	28800	-	-	22.8	26k	-	-	1055	1200k	-	-	1055	1200k
Japan	-	-	-	-	-	-	-	-	-	-	-	-	-	-	2612	350k
Kenya	10.8	0.7	-	-	6.85	0.4	-	-	3600	225	-	-	4k	250	-	-
Luxembourg	-	-	22.2	692	-	-	17.3	540	-	-	1122	35k	-	-	1603	50k
Malaysia	5	1.9	-	-	2.5	1.0	-	-	340	131	-	-	315	121	-	-
Malta	-	-	6.4	1.92	-	-	5.3	1.58	-	-	-	-	-	-	-	-
New Zealand	19	10.2	-	-	17	9.1	-	-	790	425	-	-	1075	578	-	-
Nigeria	1.5	0.4	-	-	1.46	0.4	-	-	1500	429	-	-	2375	679	-	-
Norway	-	-	31.4	186	-	-	27.0	160	-	-	845	5k	-	-	1183	7k
Poland	-	-	0.9	8186	-	-	-	-	-	-	444	4250	-	-	268	2567k
Portugal	-	-	4.8	645	-	-	3.8	515	-	-	504	68k	-	-	630	85k
Saudi Arabia	8	2.1	-	-	5	1.3	-	-	-	-	-	-	-	-	-	-
Singapore	8.67	4.0	-	-	6	2.8	-	-	600	279	-	-	775	360	-	-
South Africa	8.25	1.8	-	-	2.22	0.5	-	-	475	106	-	-	525	117	-	-
Spain	-	-	19.1	1821	-	-	15.0	1428	-	-	578	55k	-	-	683	65k
Sweden	-	-	33.6	190	-	-	32.8	185	-	-	1496	8450	-	-	1169	6600
Switzerland	51.5	33.2	-	-	34.7	22.4	-	-	-	-	1081	1375	-	-	1416	1800
Turkey	-	-	0.4	1352	-	-	0.3	825	-	-	110	344k	-	-	265	828k
UK	-	-	11.4	5.89	-	-	9.0	4.63	-	-	686	355	-	-	821	425
USA	-	-	37.2	37.2	-	-	26.5	26.5	-	-	450	450	-	-	595	595
Zambia	2.34	0.2	-	-	1.97	0.1	-	-	1705	122	-	-	1550	111	-	-
Zimbabwe	2.88	1.8	-	-	1.08	0.7	-	-	240	150	-	-	670	419	-	-

NC- National currency
Table 1 - Skilled/unskilled labor costs and estimating costs per square meter for housing (for the normal standards prevailing in that country). 2,3

The materials used were:

• Gypsum board - The gypsum board used was a 10 mm thickness composite panel, fabricated with a volume density of about 800 kg/m³ and a surface density of about 7.5 kg/m². Each panel was 1.20 x 2.80 m.
• Cork - Three types of cork materials were used:

Agglomerated composition cork (or white agglomerated cork) - Made of cork granules (1 to 3 mm) bound together under heat and pressure. It has a specific weight of about 125 kg/m³ and a modulus of elasticity of about 16 kg/cm²;

Rubbercork - A cork derived material done with different sizes of granulated cork and synthetic rubber (binder: nitrile, neoprene or cloroprene - NBR or SBR). The one used in this study had a specific weight of about 900 kg/m³ and a modulus of elasticity of about 110 kg/cm²;

ISO500 - Two layers of 25 mm agglomerate composition cork glued with a 3 mm thick rubbercork sheet in between.

EXPERIMENTAL PROCEDURES

Transmission loss measurements were performed in the sound transmission facility at the National Laboratory for Civil Engineering (LNEC) in Lisbon, Portugal. The facility consists of two adjacent reverberation rooms each with a volume of 112 m³ and a surface area of 140 m². The interior dimensions of each chamber were 5.47 x 4.87 x 4.20 m. The test opening wall between the rooms was 3.46 x 2.79 m. The procedures for measuring transmission loss (TL) followed the Portuguese standard NP-669 (similar to the standards ISO 140 and ASTM E90). The equipment used were 1" microphones Brüel & Kjaer (B&K) model 4145 with B&K 2615 preamplifiers and B&K 2607 amplifier.
 

RESULTS

The Transmission Loss (TL) values obtained for each partition tested are presented in Table 3 and Figure 2. The Sound Transmission Class (STC), a single-number rating, was calculated according to NP-2073, a similar method as described in ISO 717 and ASTM E413. Table 2 presents the summary of results.

All the partitions show a small drop (critical frequency) near the 160 Hz frequency band. The partition D3 also presents another drop in the TL values for the 1000 Hz frequency band.

PARTITION	STC	TRANSMISSION LOSS TL (dB)
	(dB)	minimum	maximum
single D1	28	17	31
D2	34	21	39
D3	32	21	45
D4	34	16	44
double DD1/2-1	44	25	58
DD1/2-2	43	24	57
traditional brick 7 cm + plaster	37	27	51

DD1/2 -1 or 2 = D1 + D2 (double) with both ways of emission of sound
Table 2 Comparison of Sound Transmission Class (STC) and Transmission Loss (TL) among the partitions tested and the traditional single brick wall.

Figure 1 Partitions tested.

Figure 2 Spectra of TL for the 6 partitions tested.

Frequency	Transmission Loss - TL (dB)
(Hz)	D1	D2	D3	D4	DD1/2-1	DD1/2-2
125	18	24	23	18	27	26
160	17	21	21	16	25	24
200	20	26	25	21	29	28
250	20	27	28	21	32	33
315	22	26	28	27	35	33
400	24	29	28	30	38	37
500	25	29	29	30	42	38
630	26	30	28	33	44	42
800	27	31	27	36	48	45
1000	28	33	24	35	49	47
1250	28	34	31	38	51	48
1600	30	36	38	40	55	54
2000	31	36	41	41	56	54
2500	31	37	45	41	57	55
3150	30	39	45	41	58	57
4000	30	39	44	44	56	55
S T C	28	34	32	34	44	43

Table 3 Transmission Loss (TL) values.
 

COMPLIANCE WITH THE PORTUGUESE NOISE CODE

The Portuguese Noise Code (Regulamento Geral sobre o Ruído, D.L. 251/87-24.06) in use since January 1988, states that partitions between rooms in the same dwelling should have a STC greater or equal to 40 dB. Therefore (Table 3) the double partition DD1/2 complies with this law. The traditional brick wall with STC of 37 does not conform with the minimum legal STC value.
 

MATHEMATICAL MODEL

Model

A mathematical model was searched for the estimation of the Transmission Loss (TL) for this type of lightweight non homogeneous sandwich partition.

Since the effect of the critical frequency is small in this type of partition and because it will be present in the low end of the frequency range of interest, the developed model will show just one expression for the entire range of frequencies (125-4000 Hz).

The general appearance for the model will be 1: TL = 10 log (A.B.C) + D

where A, B, C and D are terms representing: mass effect, internal damping, partition stiffness and a constant.

Using the work of Cremer and Fahy 4,5 as the basis for the mathematical elaboration, and after some computations 1 the model searched is:

TL (dB) = 10 log | (m / 700) ^1.5 * f ^2.5 * B ^0.5 - (m² * f ^1.5) | - 37.7

where the sandwich partition is transformed in one equivalent homogeneous partition. The values for m and B are then not referred to any of the layers of the sandwich panel but they represent the values for an equivalent homogeneous composite panel.

The partition equivalent mass will be found using the following expression: m = 2 r₁ h₁ + r₂ h₂

where: ₁ - index for face sheet; ₂ - index for core; r - volume density (kg/m³); h - thickness (m)

The partition equivalent bending stiffness will be found using the expression 4: B = 2 .B₁ [1 + (3 g) /  (1 + g)]

where: B₁ = E₁ h₁³ /12 and g = 2 G₂ / E₁ h₁ h₂ k²

and correspondingly:

1 - index for face sheet	k - wave number, k = w/c = 2 f/c
2 - index for core	Pn - Poisson number
E - modulus of elasticity (Young's mod.) (kg/m2)	G2 - core transversal elasticity modulus, G2 = E2 / 2(1+Pn)

Figure 3 Four comparison cases for spectra TL measured versus TL model.

Using the value of 0.3 for the Poisson number (reasonable value for the usual building construction materials) and after some mathematical calculation we get: B= (E₁ h₁³ / 6) + c² E₁ E₂ h₁³ / (10.4 Pi ² E₁ h₁ h₂ f ² + 2 c² E₂)

where:

1 - index for face sheet	B - equivalent bending stiffness	c - speed of sound in air (m/s)
2 - index for core	E - modulus of elasticity (Young's mod.) (kg/m2)	f - frequency (Hz)
		h - thickness (m)

Model Agreement

Table 4 presents the results for the STC values using the described model to this sample of partitions and the differences found to the calculated STC values using the measured TL's. As seen in this table, the average difference of +1 dB represents a very reasonable conformity given by the model. The Pearson correlation coefficient (R) between predicted and measured STC's is 0.87. Figure 3 displays four spectra's comparisons between predicted and measured TL's where a good agreement can be seen.

PARTITION	STC (dB)			DIFFERENCE
	Measured	Model		(dB)
D1	28	31		+ 3
D2	34	33		- 1
D3	32	33		+ 1
D4	34	35		+ 1
			Average	+ 1

Table 4 Comparison between STC measured and STC model.
 

SYNOPSIS

A socioeconomic analysis about the Portuguese situation concerned with the lodging policies and its building construction industry and development, justified the study of a newly designed lightweight partition done with cork, a traditional material not widely used in this industry.

This newly developed sandwich lightweight partition, made with cork in the core and gypsum board as face layers, was presented and acoustically characterized. A new and simple mathematical model was presented to evaluate the TL of this kind of partition. The results obtained in laboratory were presented as well as the comparisons with the predicted TL values using the new model proposed. Five partition types were tested giving STC values up to 44 dB. The compliance with the 40 dB minimum STC value for partitions between rooms in the same dwelling stated in the Portuguese Noise Code was displayed for the double partition tested.
 

ACKNOWLEDGMENTS

The author wishes to recognize the University of Porto, the National Laboratory for Civil Engineering (LNEC), the IPOCORK Co. and Prof. Pedro Martins da Silva for their support in this project.
 

REFERENCES

1. Carvalho, António P. O. Estudo Acústico de Divisórias Amovíveis - Análise modelar e desenvolvimento de soluções, MSc. thesis, University of Porto, Portugal, 1988.
2. Davis, Belfield and Everest Spon's International Construction Costs Handbook, E&FN Spon, London, 1988.
3. Davis Langdon & Everest Spon's European Construction Costs Handbook, E&FN Spon, London, 1992.
4. Cremer, Lothar Theorie der Schalldámmung Dunner Wande bei Scragen Einfall, Akustische Zeitschrift, 1942, 81(7).
5. Fahy, Frank Sound and Structural Vibration, Radiation, Transmission, and Response, Academic Press, 1985.