**4.4. Physical properties**

Physical and mechanical properties of the flexible polyurethane foams are evaluated by different types of tests in order to make an entire picture from the foam part performance during the consuming by the customer. For instance, flexible polyurethane foam is widely used as car seat foam and it has to keep its shape and other properties such as hardness and compression set during the time which is used. The most important properties of the polyurethane flexible foam as car seat foam according to RENAULT specifications are viewing as below.

Core Density


*4.3.2. Flammability* 

investigated.

*Synergetic effect*

considerably.

two categories:

1. Flaming 2. Non-flaming

*4.3.3. Smoke density and toxicity* 

flaming and non-flaming status.

**4.4. Physical properties** 

viewing as below. Core Density

Flammability of the polyurethane foam is running with wide range of test methods depends on the applications and customers specification. Some fire tests standards include: FMVSS NO.302, British Standard 5852, ISO 9772 and FAA/JAA 25.853 Appendix F. as an example the airplane seat foam fire tests according to FAA/JAA 25.853 Appendix F have been

In this test 5 samples with 75mm\*305mm\*13mm dimension have been subjected with flame

Burning time (the time that burning is continuing after removing the flame source) Burned length (the length of the foam which is damaged by the burning process) Time of dripping

The synergetic effect of different types of FR has been observed. for instance, the fire properties of the EG loaded foams is much worse than when it is used by mixing with a liquid FR such halogenated phosphorous flame retardant. **Also when some amount of melamine is added to the TMCP and TDCP containing foams it helps to decrease total** 

Also the mixing of the liquid FR could boost the fire properties of the melamine loaded foam

Smoke density and Toxicity are measured according to Airbus Directive ABD0031 (2005) on

Samples with 76mm\*76mm\*13mm are chosen to do the above mention tests against them in

Physical and mechanical properties of the flexible polyurethane foams are evaluated by different types of tests in order to make an entire picture from the foam part performance during the consuming by the customer. For instance, flexible polyurethane foam is widely used as car seat foam and it has to keep its shape and other properties such as hardness and compression set during the time which is used. The most important properties of the polyurethane flexible foam as car seat foam according to RENAULT specifications are

vertically for 12 sec and the following parameters have been investigated.

**heat evolved, total smoke produced and CO emission significantly[2].** 

(the time which droplet continues to burn).

Resilience in 1st and 5th cycle according to D455128

When the polyurethane flexible foam is going to be fire resisted, some fire retardants in liquid or solid forms are entered in to the foam structure and make some changes in the physical properties of the final foam part. Mostly the valuable changes have been observed by the solid FR addition rather than the liquid one.

Depending on the fire retardant nature, shape and size, their addition may have some positive or negative effect on foam physical-mechanical properties. By loading the solid FR with the same amount, the foams become softer, because both additives have a similar size as cell windows and make the foam inhomogeneous. With EG, the homogeneity would be less than the foam loaded by melamine, because of its bigger size and flake shape which makes the foam much softer [1].

Sag-factor or the comfort index [12] increasing when the percentage of EG and melamine increases. It means that by adding solid FR, the comfort index would change considerably. Compression set, which is another very important factor, has increased by rising the EG percentage, but there was almost no changes in CS by increasing the melamine content. This effect is due to destroying effect of the cells structure by both additives but mainly by the EG.

Tear strensgth of the foams has improved by increasing the EG which could be related to the rigidity of EG flakes but deteriorates when melamine is added.

Finally, the resilience in 1st cycle is decreased for all additives but it is recovered in 5th cycle, because in 1st cycle the polymer chains have lost their flexibility due to rigid particles but after 5 cyclic movements the particles are embedded in struts and joints and the foam restores its flexibility.

### **5. Statistical method**

Principle component analysis (PCA) is a useful method to illustrate relations between different parameters by using STAT-BOX-ITCF [13, 14].

Interpretation of the results consists first in the checking the representation of the variables in the circles of correlation. The correlations between variables are deduced from the relative position and the length of their corresponding vectors on the circle of correlation. An example of interpretation is done in (Fig.12); the angle between two vectors defines the intensity of the correlation (vectors 1 and 5). If α is=90⁰, no relation exists between the variables. The strength of the correlation is higher when the angle is close to 0⁰ or 180⁰. So, orthogonal vectors (vectors 1 and 2) mean no correlation between the variables. Data are strongly correlated if their vectors are collinear (vectors 1 and 3, and vectors 1 and 4). The nature of the correlation also depends on the direction of the vectors: if vectors have the same direction (vectors 1 and 4) the variables are correlated, i.e. an increase in the variable linked to the vector 1 corresponds to an increase in the variable linked to the vector 4. Inversely, if vectors are opposite (vectors 1 and 3), the variables are anti-correlated.

Polyurethane Flexible Foam Fire Behavior 115

**Figure 13.** Correlation circle—relationship: cone calorimeter/FMVSS.

As a first hypothesis, we can consider the following relation:

combustion (Fig.14).

**Figure 14.** Principle of FMVSS.

dripping.

From the energy assessment of the foam consumption during 1s, we can find a relation between the propagation speed of the flame and the energy of the tar produced by the

*q*1 + *q*2 - *∆Q* = *Q* = constant *∆Q* corresponds to the part of heat used to melt the polymeric matrix leading to

The correlation between two variables is also a function of the length of the vectors. As example, vectors 2 and 6 are co-linear and so should be anti-correlated. But the weak length of the vector 6 means that its corresponding variable does not influence the variable linked to vector 2 [2].

**Figure 12.** Interpretation of principal components analysis

#### **5.1. Cone calorimeter–FMVSS 302**

The principal components analysis from cone calorimeter and FMVSS 302 data shows the following correlations (Fig.13)


**Figure 13.** Correlation circle—relationship: cone calorimeter/FMVSS.

From the energy assessment of the foam consumption during 1s, we can find a relation between the propagation speed of the flame and the energy of the tar produced by the combustion (Fig.14).

**Figure 14.** Principle of FMVSS.

114 Polyurethane

to vector 2 [2].

same direction (vectors 1 and 4) the variables are correlated, i.e. an increase in the variable linked to the vector 1 corresponds to an increase in the variable linked to the vector 4.

The correlation between two variables is also a function of the length of the vectors. As example, vectors 2 and 6 are co-linear and so should be anti-correlated. But the weak length of the vector 6 means that its corresponding variable does not influence the variable linked

The principal components analysis from cone calorimeter and FMVSS 302 data shows the

 *RHR1* is moderately correlated with *Figra1*: Figra1 is a variable that depends on the first peak of HRR (also called q1max), *d* the time it occurs. So, it seems quite coherent to find

 *RHR1* is correlated with *Figra2*. In the cone calorimeter, the foam degradation occurs in two main steps. It is obvious that an important consumption of fuel in the first step

 FMVSS is strongly correlated with Figra1 and Figra2 and inversely correlated with *RHR2*. The lower Figra1 and Figra2, the slower the flame spread. A high RHR2 means

this kind of relation if the relative variation of the time is low.

Inversely, if vectors are opposite (vectors 1 and 3), the variables are anti-correlated.

**Figure 12.** Interpretation of principal components analysis

**5.1. Cone calorimeter–FMVSS 302** 

following correlations (Fig.13)

leads to a lower Figra2.

loss of heat by dripping.

As a first hypothesis, we can consider the following relation:

$$q1 + q2 \cdot \Delta Q = Q = \text{constant}$$

 *∆Q* corresponds to the part of heat used to melt the polymeric matrix leading to dripping.

 *q*1 corresponds to the energy released during the first stage of the combustion that leads to the formation of the tar (Figra1).

Polyurethane Flexible Foam Fire Behavior 117

**5.2. Cone calorimeter–British Standard** 

*Ignition Source Crib 5 test to SI 1324 Sch. 1 Pt. 1* 

correlation.

the

The statistical computation was made considering the two different sets of foams: the foams containing TMCP–melamine and the ones containing TDCP–melamine. The level of fire retardant additives has been included in the computation but is not shown on the circles of

Considering the TMCP–melamine foams (Fig.16) it is of interest to note that the lower are Figra1 and Figra2, the lower are the burn times, TWL and DWL. We also note that *T*2 is strongly inversely correlated with the data of SI 1324 Sch. 1 Pt. 1, that is to say the higher *T*2

The statistical computation of the data from the formulations TDCP–melamine clearly shows that the fire behavior of these foams in the SI 1324 test is linked to the second stage of degradation of the foam in the cone calorimeter (Figra2 and *T*2). Indeed, the Figra curve

A high Figra means a high rate of flame propagation and so leads to a high weight loss of the material. Hence, it is not surprising that Figra curves are strongly linked to the BS5852

Better results under the SI 1324 test (lower TWL, DWL and burn time).

represents the fire growth rate of foam during combustion.

**Figure 16.** Correlation circle—relationship: cone calorimeter/SI

*1324, TMCP–melamine formulations* 

*q*2 corresponds to the energy released by the combustion of the tar (Figra2).

This relation indicates the different strategies to decrease the value of RHR1 (and so Figra1), that is to say the flame spread in the FMVSS 302 tests:


Hence, we may propose that the flame propagation rate in FMVSS 302 testing is much lower when easy melting and dripping allows heat reduction and tar dripping. It may be proposed that *q*2 corresponds in fact to the almost complete combustion of the tar.

Comparing the RHR curves of foams processed with variable water level, we note that the density of the foam strongly influences the first RHR peak (Fig.15). The higher the water content (the lower the density) the faster the step of melting under cone calorimeter conditions.

The effect of density on RHR1 may explain the previous correlation found between density and FMVSS 302. Low density leads to rapid melting and to a high flame propagation rate.

**Figure 15.** Effect of density on the melting stage of polyurethane during combustion

#### **5.2. Cone calorimeter–British Standard**

116 Polyurethane

to the formation of the tar (Figra1).

degrades at a later stage.

conditions.

that is to say the flame spread in the FMVSS 302 tests:

*q*1 corresponds to the energy released during the first stage of the combustion that leads

This relation indicates the different strategies to decrease the value of RHR1 (and so Figra1),

 To decrease the heat released during the first stage of degradation of the foam and as a consequence to decrease the heat fed back to the virgin polymer (decrease in Figra1).

 To delay the heat released by the tar. When the foam is molten, the tar starts to burn and this tar is not immediately lost by dripping. Hence, it is of interest to delay the combustion of this tar to enable it to drip (decrease in Figra2). An increase in RHR2 is not sufficient to reduce the flame spread and it is important that the high energy tar

Hence, we may propose that the flame propagation rate in FMVSS 302 testing is much lower when easy melting and dripping allows heat reduction and tar dripping. It may be proposed

Comparing the RHR curves of foams processed with variable water level, we note that the density of the foam strongly influences the first RHR peak (Fig.15). The higher the water content (the lower the density) the faster the step of melting under cone calorimeter

The effect of density on RHR1 may explain the previous correlation found between density and FMVSS 302. Low density leads to rapid melting and to a high flame propagation rate.

To increase RHR2, that is to say to reduce the energy of combustion by dripping.

*q*2 corresponds to the energy released by the combustion of the tar (Figra2).

To decrease the total heat evolved *Q* using specific FR additives.

that *q*2 corresponds in fact to the almost complete combustion of the tar.

**Figure 15.** Effect of density on the melting stage of polyurethane during combustion

*Ignition Source Crib 5 test to SI 1324 Sch. 1 Pt. 1* 

The statistical computation was made considering the two different sets of foams: the foams containing TMCP–melamine and the ones containing TDCP–melamine. The level of fire retardant additives has been included in the computation but is not shown on the circles of correlation.

Considering the TMCP–melamine foams (Fig.16) it is of interest to note that the lower are Figra1 and Figra2, the lower are the burn times, TWL and DWL. We also note that *T*2 is strongly inversely correlated with the data of SI 1324 Sch. 1 Pt. 1, that is to say the higher *T*2 the

Better results under the SI 1324 test (lower TWL, DWL and burn time).

The statistical computation of the data from the formulations TDCP–melamine clearly shows that the fire behavior of these foams in the SI 1324 test is linked to the second stage of degradation of the foam in the cone calorimeter (Figra2 and *T*2). Indeed, the Figra curve represents the fire growth rate of foam during combustion.

**Figure 16.** Correlation circle—relationship: cone calorimeter/SI

#### *1324, TMCP–melamine formulations*

A high Figra means a high rate of flame propagation and so leads to a high weight loss of the material. Hence, it is not surprising that Figra curves are strongly linked to the BS5852 results. The combustion of PU foam occurs in two steps: the "melting" of the foam and the combustion of the tar. The tar combustion is the most exothermic part of the combustion. A decrease in the heat released by the tar reduces the flame propagation and leads to a decrease in the weight loss of the foam (Fig. 17).

Polyurethane Flexible Foam Fire Behavior 119

Regarding the TDCP–melamine formulations, we note a positive effect of the TDCP amount on the RHR1 peak. Even if TDCP degrades later than TMCP, a part of the TDCP is efficient

The melamine content is strongly correlated with the SI 1324 data. High melamine content

The statistical treatment shows that the FMVSS 302 rating is an inverse function of the density of the foam which is itself a function of the water index (Fig. 18). No significant relations may be proposed between FMVSS 302 and porosity or TDI index because data did not show any variation of the porosity (same SnOct content) and only a low variation of the TDI index.

The porosity index of the foam is strongly correlated with the SnOct range used in the foam

The previous study of conventional foams has revealed correlations between the FMVSS 302 testing and these parameters. The PCA study shows the absence of correlation between the EO content, the porosity (and so the SnOct range) and the index of the foam with the FMVSS 302 testing. However, it clearly shows that FMVSS 302 is strongly and inversely correlated

leads to a decrease in TWL, DWL burn time and maximum rate of weight loss.

The Figra2 and RHR2 peaks are also correlated with these data.

with the density of the foam as it has been previously supposed.

**Figure 18.** Correlation circle—relationship: physical properties/FMVSS.

in the first stage of the combustion.

**5.3. Properties–FMVSS 302** 

manufacturing.

The TDCP and TMCP additives differ in their chlorine and phosphorus content and also in their temperature of degradation. TMCP degrades earlier than TDCP (150 ⁰C and 210 ⁰C, respectively); this temperature corresponds to a 5 wt. % weight loss under thermo gravimetric analysis conditions). A previous study [15] has clearly shown that TMCP is efficient in the early stage of combustion but no interaction with melamine is observed (temperature of 5 wt. % weight loss of melamine is 290 ⁰C). TDCP acts later and when melamine starts to degrade about 50 wt. % of TDCP is available in the system, so a strong TDCP–melamine synergy is observed. The use of TDCP or TMCP in combination or not with melamine leads to very distinctive fire properties of the foams.

**Figure 17.** Correlation circle—relationship: cone calorimeter/SI

#### *1324, TDCP–melamine formulations*

Considering the TMCP–melamine foams, it is of interest to note that the higher the TMCP content the lower is RHR1. That confirms the early effect of TMCP that acts by decreasing the heat released by the foam in the first stage of the combustion. Secondly, the melamine content is inversely correlated with RHR2. As described previously, the temperature of decomposition of melamine is high (290 ⁰C) and this inverse correlation indicates an efficiency of melamine during the combustion of the tar.

Regarding the TDCP–melamine formulations, we note a positive effect of the TDCP amount on the RHR1 peak. Even if TDCP degrades later than TMCP, a part of the TDCP is efficient in the first stage of the combustion.

The melamine content is strongly correlated with the SI 1324 data. High melamine content leads to a decrease in TWL, DWL burn time and maximum rate of weight loss.

The Figra2 and RHR2 peaks are also correlated with these data.

#### **5.3. Properties–FMVSS 302**

118 Polyurethane

decrease in the weight loss of the foam (Fig. 17).

with melamine leads to very distinctive fire properties of the foams.

**Figure 17.** Correlation circle—relationship: cone calorimeter/SI

efficiency of melamine during the combustion of the tar.

Considering the TMCP–melamine foams, it is of interest to note that the higher the TMCP content the lower is RHR1. That confirms the early effect of TMCP that acts by decreasing the heat released by the foam in the first stage of the combustion. Secondly, the melamine content is inversely correlated with RHR2. As described previously, the temperature of decomposition of melamine is high (290 ⁰C) and this inverse correlation indicates an

*1324, TDCP–melamine formulations* 

results. The combustion of PU foam occurs in two steps: the "melting" of the foam and the combustion of the tar. The tar combustion is the most exothermic part of the combustion. A decrease in the heat released by the tar reduces the flame propagation and leads to a

The TDCP and TMCP additives differ in their chlorine and phosphorus content and also in their temperature of degradation. TMCP degrades earlier than TDCP (150 ⁰C and 210 ⁰C, respectively); this temperature corresponds to a 5 wt. % weight loss under thermo gravimetric analysis conditions). A previous study [15] has clearly shown that TMCP is efficient in the early stage of combustion but no interaction with melamine is observed (temperature of 5 wt. % weight loss of melamine is 290 ⁰C). TDCP acts later and when melamine starts to degrade about 50 wt. % of TDCP is available in the system, so a strong TDCP–melamine synergy is observed. The use of TDCP or TMCP in combination or not

The statistical treatment shows that the FMVSS 302 rating is an inverse function of the density of the foam which is itself a function of the water index (Fig. 18). No significant relations may be proposed between FMVSS 302 and porosity or TDI index because data did not show any variation of the porosity (same SnOct content) and only a low variation of the TDI index.

The porosity index of the foam is strongly correlated with the SnOct range used in the foam manufacturing.

The previous study of conventional foams has revealed correlations between the FMVSS 302 testing and these parameters. The PCA study shows the absence of correlation between the EO content, the porosity (and so the SnOct range) and the index of the foam with the FMVSS 302 testing. However, it clearly shows that FMVSS 302 is strongly and inversely correlated with the density of the foam as it has been previously supposed.

**Figure 18.** Correlation circle—relationship: physical properties/FMVSS.
