A new era for MDI: flexible PU slabstock foam.

Traditionally, tolylene diisocyanate (TDI) has been used for the production of flexible slabstock foams for the last 40 years or so. Almost exclusively an 80/20 blend of the 2,4 and 2,6 TDI isomers is used in the United States for producing polyether polyol based slabstock foam while in other parts of the globe some foam is also poured with a 65/35 ratio of 2,4 and 2,6 TDI (high load bearing and polyester foams).

Methylene bis diphenyl diisocyanate (MDI) has typically been used in rigid PU foams and elastomers but also in flexible molded foams for high quality high resilience applications (ref. 1 ). In those applications, MDI was often used in the form of a prepolymer, a quasi-prepolymer or a blend of TDI and MDI (ref. 2).

In all those applications. the foam produced was of the high resiliency type obtained via highly reactive polyols, crosslinkers (mainly diethanolamine) and low potency silicones.

Hence, it was quite natural to apply this HR technology to the slabstock area and a lot of research effort was invested in this field over the last few years, especially in Europe where the market demand seems stronger than in the U.S.

The larger size of the slabstock buns led to further formulation challenges especially in terms of processability and cell-opening. Nevertheless, MDI has now been successfully run with various technologies well described in the literature (refs. 3-5). However, all of these technologies are based on MDI prepolymers and were designed for the top end of the bedding and furnishings industries like replacement of latex foams, TDI high resilience foams or pocketed springs.

BASF has shown that when MDI is used in conventional foam applications (ref. 6), it usually needs to be blended with TDI 80/20 in order to improve the processing latitude of the system. The advantage of this technology is to make very firm foams without the need for any copolymer polyol but is generally limited to high densities (2.5 to 3.5 pcf) where an improved foam feeling, believed to be due to a finer cell structure, is obtained.

This study was initiated to assess the value of MDI in conventional foams in terms of load bearing capabilities, but also to gain some insight on more fundamental parameters like reactivity, compatibility and the polymer structure-foam properties relationships of those foams.

A series of MDI variants were screened and two were selected based on their enhanced processability and a closer look was then taken to the polymer network formed and how this affected the final foam properties. The main challenge was to produce low density foams with reasonable processing latitude and acceptable foam properties using slabstock polyols, surfactants and catalysts originally designed for TDI applications; furthermore, the higher equivalent weight of MDI and poorer blowing efficiency made this development even more challenging.

A few additives were studied which improve both the water/polyol/isocyanate compatibility and foam processing latitude and lead to foams with good properties; however, tear resistance and elongation at break remain borderline.

Materials and methods

The raw materials used in this study:

* Voranol 3137 polyol is a 3000 mw conventional EO/PO heterofed polyol used in slabstock applications.

* Voranol 3512 polyol is a 3500 mw conventional EO/PO heterofed polyol used in slabstock applications.

* Voranal 3943 polyol is a 43% solids copolymer polyol used in slabstock foam as a load builder.

* MDI is diphenylmethane diisocyanate containing a high proportion of the 4,4' isomer. The remainder being the 2,4' isomer or a blend of this isomer with carbodiimide to obtain a liquid version of pure MDI.

* Isonate 50 is a blend of 4,4' and 2,4' MDI.

* Voranate T-80 isocyanate is an 80/20 mixture of the 2,4 and 2,6 isomers of tolylene diisocyanate.

* PMDI stands for polymeric MDI and is a mixture of pure MDI and isocyanates containing 3,4,5 and higher functionalities.

* Additives 1 and 2 are hydroxyl terminated molecules that act as compatibilizing agents for the system polyol/PMDI/water.

* Stannous octoate is a metal salt catalyst used for the production of slabstock foams.

Polyurethane foam preparation

Small scale: Bench-scale box foam machine. The polyol and the additives, except the tin catalyst, were placed in a metal cup and mixed for 15 seconds at 2,400 rpm. The tin catalyst was added and mixing continued for another 15 seconds at 2,400 rpm. The appropriate amount of isocyanate was added to the cup and mixed at 3,200 rpm for five seconds. The incipient foam was poured into a 15" x 15" x 10" PE lined wooden box and the foam was allowed to rise. Formulations used will be given later.

Large scale: Pilot line evaluation. A few of the box-foam formulations were scaled up using a Varimax pilot line foam machine. A polyol output of 80 to 100 lb./min. was used in order to get a total throughput compatible with the density to be produced, thus leading to an optimal rise profile/conveyor speed which in turn determines the block shape at a given fall plate setting.

Foam physical property testing

Foams were allowed to cure for seven days at room temperature prior to physical testing. Bulk properties, including dry compression sets. were measured using ASTM methodology.

FTIR

Infrared spectroscopy of foam samples used a Nicolet 740 FTIR spectrometer equipped with an MTEC model 200 photoacoustic cell.

Swelling study

Solvent swelling studies were performed by placing a preweighed 1" x 1" x 1" cube of foam in excess of dimethyl-formamide (DMF) and allowing it to stand for seven days. The wet foam is then blotted and weighed to obtain a wet weight. The foam sample is then vacuum dried at 80[degrees]C for 48 hours and a dry weight obtained. Volume swelling and percent extractable are then calculated for the specimen.

Dynamic mechanical spectroscopy evaluation

Dynamic mechanical spectroscopy of foam samples was obtained using a Rheometrics solids analyzer, RSA II.

Microscopy

Optical microscopy was performed to determine the miscibility of some of the reagents using an Olympus SZH with a magnification of 128.

Measure of foam exotherm at blow-off time.

A 40 gauge type J thermocouple (extremely fine) linked to a digital thermometer was placed in the central part of a 500 cc plastic cup by drilling a hole through the bottom part of the cup. This cup was then placed in a 15" x 15" x 10" wooden box and a normal box-foam was made according to the above described method. The reactive liquid completely filled the cup thus allowing an extremely fast response of the thermocouple (good wettability). The rising foam was then allowed to overflow in the box and the temperature at blow off was recorded.

Polyol reactivity test

This test is carried out using a jacketed cone and plate viscometer set at 25[degrees]C. The polyol, isocyanate and catalyst are mixed together at an isocyanate index of 105. The isothermal viscosity build-up is recorded versus time and reflects the rate of polymerization of the polyol with isocyanate in absence of water.

Results and discussion

Compatibility/reactivity study

The first obvious observation made when replacing TDI with MDI in a conventional slabstock formulation is the difference in rise profile. These reaction rate differences can be due to a combination of the following: the position of the NCOs on the aromatic ring, their sterical hindrance and also to the functionality of the isocyanates and the acidity of the system. It is well known that the foaming reaction needs to take place in basic conditions when tertiary amine catalysts are used; the acidity of TDI is very low in all cases while the MDI acidity varies dramatically depending on whether the product in use was distilled or not and also what process was used to manufacture the isocyanate.

Furthermore, the functionality of the MDI dictates the viscosity build-up at a given isocyanate conversion and therefore will also largely influence the apparent reactivity characterized by the gel time or string time.

Another factor influencing the chemical reactivity of an MDI system is the compatibility of the various reagents, especially the ternary system polyol, water and isocyanate.

As a result, the rise profile curves of TDI and MDI based foams are very different, which challenges processing on commercial foaming machines as those machines are designed to pour TDI foams.

To illustrate the difference in rise profiles of MDI and TDI based foams, a series of formulations were run as described in table 1. It can be noted that at 3 pbw of water, a delay in rise time occurs when MDI of higher acidity is used while the difference is less obvious for a low acidity PMDI.

    Table 1 - effect of isocyanate acidity upon foam
                      reactivity
Sample number                  1          2           3
Isocyanate index             105        105         105
TDI                         30.8
Polymeric MDI 1                          59
Polymeric MDI 2                                      59
Voranol 3137 polyol           95         95          95
Water                          3          3           3
Additive 1                     5          5           5
Surfactant                   1.2        1.2         1.2
Dabco 8264 catalyst          0.6        0.6         0.6
Stannous octoate            0.24       0.24        0.24
Isocyanate acidity (ppm)      <5         40         130
Cream time (seconds)          10         14          13
Rise time (seconds)           90         93         104

In order to better differentiate between the effects of all the parameters described above, a series of experiments were carried out:

* Polyol/isocyanate reactivity tests were carried out following the above described method. The results are shown in figures 1 and 2. When looking at the data, one can see that after 7 minutes, the TDI/polyol mixture only reaches 4,000 cPs while the PMDI/polyol one has reached the gel point with 150,000 cPs. Isonate 50, a difunctional MDI containing a high level of ortho -- para isomer shows a somewhat intermediate viscosity build-up with 20,000 cPs after seven minutes. Considering the lower functionality of this pure MDI versus polymeric PMDI, it seems rather logical to observe this trend. This reactivity difference between TDI and MDI has been well described in the literature for model molecules in solution (see for instance reference 7). At room temperature, the reaction of 2,4 TDI with 2-ethylhexanol is about equivalent to 4,4, MDI at the beginning of the reaction while it is about four times slower at 90% conversion; on the other hand 2,6 TDI is consistently less reactive than 4,4, MDI from two fold at the early stage of the reaction to about five fold at 90% conversion. Other more recent studies claim a reactivity difference between 2,4 TDI and 4,4, MDI of about two fold at the beginning of the reaction (ref. 8).

* In order to visualize the system reactivity differences, foam rise profile tests were then performed as shown in figures 3 and 4. The formulations used are grouped in table 2. It can be seen that replacing TDI with PMDI in this case has little impact on the rise time, but significantly changes the rise profile. A higher rise rate is found for TDI at the beginning of the reaction, while PMDI shows a higher rate towards the end of the rise. The other obvious difference is the lower bun height obtained in case of PMDI. This partly reflects the well-known lower blowing efficiency of PMDI vs. TDI.

    Table 2 - formulation used for rate of rise
                      purposes
Foam number                    1          2           3
Voranol 3137                 100        100          95
Additive                                              5
Water                          3          3           3
Surfactant                     1          1           1
Amine cat.                   0.4        0.4         0.4
Stannous oct.               0.15       0.15        0.15
TDI 80/20                   39.6
PMDI                                   59.5        59.5
ISO index                    105        105         105
Cream time (s)                12         12          13
Rise time (s)                 96         97         103
Foam weight (g)            1,270      1,437       1,429

The same amount of polyol was used in both cases and the same isocyanate index, which means that a larger mass of PMDI was used than of TDI (which is a reflection of the higher equivalent weight of PMDI). If the amount of carbon dioxide generated at blow-off was the same, the MDI bun height would still have been slightly lower than the TDI bun due to the lower temperature reached at blow-off for the MDI foam (this is merely a result of the higher mol weight of the MDI molecule leading to a higher heat capacity of the polymer mass compared with the TDI formulation).

Figure 4 shows the influence of an additive developed to improve the processability of PMDI foams. This additive slightly increases the rise time of the foam which shows up near the end of the rise profile leaving the rest of the curve essentially identical to the one without this additive. Furthermore, the bun height is higher with the same amount of polyol and PMDI, which reflects a density decrease of about 5%.

It was observed during the rise profile experiments that the MDI based blends seemed to be hazy while the TDI ones were clear. A more systematic study was therefore undertaken in order to see if this physical phenomenon could explain the differences in rise curves observed earlier.

* Compatibility of the isocyanate with other reagents. 100 pbw polyol were thoroughly mixed with 50 pbw of isocyanate in a plastic beaker for 30 seconds with a tongue depressor. A drop of the mixture was then poured onto a glass plate and observed under an optical microscope before the mixture started to react (at least 5-10 minutes). The results are summarized in table 3 and a few photos are shown in figures 5-8 to highlight the observations made.

Table 3 - optical microscopy study of compatibility of PU
   reagants
Blend                    1     2     3     4     5     6     7
8
PO polyol              100         100           100        100
EO polyol                    100          100          100
100
TDI 80/20               50    50
PMDI                                                         50
50
Modified liquid MDI                 50     50
Isonate 50                                        50    50
Color                    C     C     T      O      C     C    T
C
Phases                   1     1     2      2      1     1    2
1

The modified liquid MDI is a blend of pure MDI containing a high
percentage
of the 4,4' isomer and carbodiimide. Isonate 50 is a pure MDI
containing a
large amount of ortho para MDI. C means clear liquid; T means
turbid, O
opalescent. Note: the opalescent samples are emulsions of very fine
particles
that scatter the light and are difficult to observe by optical
microscopy.
The turbid samples clearly had two phases and discrete particles.

By using a black background and a back light it was easy to differentiate between the air bubbles entrapped in the mixture and the discrete MDI droplets when present (a characteristic thick colored border was observed in case of air bubbles). One can see from the data that in a conventional TDI/polyol binary system, the two components are fully miscible in the 1/2 ratio studied. Figure 6 shows a large air bubble which was used to focus the microscope and demonstrate the transparency of this blend.

On the other hand, if TDI is replaced by either polymeric MDI or a liquefied pure MDI containing a high proportion of 4,4' MDI, a discrete phase of MDI is readily formed. This carbodiimide modified MDI was used for convenience reasons as the pure 4,4' MDI is a solid at room temperature and tends to precipitate out and crystallize. However, the experiment was carried out with pure MDI containing a high proportion of the 4,4' isomer at 45[degrees]C and equally two phases were observed. If this MDI is now replaced by a blend containing a high amount of the 2,4' isomer of MDI, the blend stays liquid at room temperature and the 1/2 ratio mixture with conventional polyol is now a clear solution characteristic of a monophase system.

PMDI can be dissolved in a high EO containing polyol. Based on this observation, additives I and 2 were developed to compatibilize PMDI and Voranol 3137. The addition of 5 to 10 pbw of additive 1 or 2 which improves the processability of the PMDI systems makes the polyol/PMDI blend turn from turbid to transparent and it is observed by microscopy that almost all of the discrete PMDI droplets disappear.

Also, Tabor et al (ref. 9) showed in a DSC study for a series of EO and PO based polyols that water was precipitating out of the system when PMDI was added to a clear polyol/water blend (100/6 ratio). Unfortunately, no easy direct observation of this phenomenon is possible as this ternary system readily reacts at room temperature and thus renders the microscope experiments difficult to interpret. However, it can still be seen that in case of TDI, as soon as water is added to the polyol/lso blend, gas is formed and the sample starts to foam, but between the growing bubbles, the solution is still clear. In case of PMDI in a blend of V-3137 compatibilized by 10 pbw of the additive, the homogeneous composition becomes hazy as soon as water is added and many black spots, that look more like PMDI than water, based on their number and volume, precipitate out and stay as discrete droplets during most of the foam rise. This seems to tell that the water is readily accepted by the polyol plus emulsifier system and drives PMDI out of phase, which could possibly explain the limited success of the additive in decreasing the density at higher water levels.

Calculation of the solubility parameters

These light microscope observations can be related to the Hildebrand parameters ([delta]) of the molecules studied. There are many ways of calculating those parameters and we will present a few to show the scattering of the data. The values are grouped in table 4.

[TABULAR DATA 4 OMITTED]

The most simple way of calculating this parameter is based on the heat of vaporization AHv, but is unfortunately only practical for small molecules having a boiling point higher than room temperature. This parameter originally introduced by Hildebrand and Scott (ref. 10) was intended only for nonpolar nonassociating systems:

[delta] = (AHv -RT/V[).sup.1/2]

R is the ideal gas constant and equals 8.31J/K mol. T is the temperature expressed in [degrees]K and V is the molar volume.

The values for polyether polyols based on either ethylene oxide (EO) or propylene oxide (PO) can be extrapolated from the values of smaller molecules like monoethylene and propylene glycol, diethylene and propylene glycol, triethylene glycol, tripropylene glycol, etc., as shown in figure 9. Unfortunately, these parameters work well only in non polar or low polar environment, and the presence of water in our systems ([delta] = 47.9 [MPa.sup.1/2]) makes these parameters of little use.

Small et al (ref. 11) introduced a second term to the solubility parameter to take into account the polar contribution:

[Mathematical Expression Omitted]

[[delta].sub.n] being the non-polar contribution equal to the original Hildebrand parameter and op the polar contribution.

Hansen (ref. 12) proposed an extension of the Hildebrand parameter method to polar and hydrogen bonding systems. It was assumed that dispersion, polar and hydrogen bonding parameters were valid simultaneously:

[Mathematical Expression Omitted]

Many of those parameters were compiled by Barton (ref. 11) and are presented in table 4 for polyols and isocyanates.

The closer the solubility parameters of the different components of a system are the more compatible they should be. A semi-empirical rule dictated by Magat (ref. 13) teaches that two substances are miscible in all proportions if the Hildebrand parameters difference is lower or equal to 0.5. In our case, due to the highly polar system studied, this is not exactly the case, but the solubility parameters indicate a correct trend: MDI is less soluble than TDI in polyether polyols, and it is more soluble in EO based polyol than in PO based one. Also, water is more soluble in polyethylene glycols than in polypropylene glycols based on the solubility parameters, which is well known.

It becomes evident that the poor solubility of MDI in a conventional polyol system affects its chemical reactivity as the blend is not homogenous but consists of regions rich in MDI (discrete droplets) and a continuous phase poor in MDI. This could result in a longer rise time and a lower maximum exotherm, assuming that the isocyanate reaction was incomplete due to the separate phases. It is well known that the foams made out of MDI are of higher densities than the ones made out of TDI. We will try in the next section to gain some insight into this problem by studying the influence of the nature of the isocyanate upon the foam density.

Blowing efficiency of TDI and MDI based foams

This study is somehow complicated by the fact that good quality foam needs to be manufactured in order to measure a meaningful density. All foams included in this study were sufficiently open to prevent the bun from shrinking and stable enough to avoid settling after the end of the rise ([+ or -] 0.5% max. of either shrink or settle is considered normal at a laboratory scale). This often leads to limitations in terms of formulation variability or requires major formulation alterations which make back to back comparisons very difficult. Bearing this in mind, we made a series of box-foams according to the above described methodology. The formulation details can be found in table 5. The assumption is made at this point that the ideal gas law is followed. As the large majority of the gas present in the cells during the rise is essentially a blend of carbon dioxide generated during the reaction, air either entrapped or injected at the machine head and water vapor (which concentration will increase with the water concentrating in the system and the temperature in the cell), this seems to be a fair hypothesis.

[TABULAR DATA 5 OMITTED]

PV = nRT

With P being the pressure and V the volume of n moles of gas at a temperature T (in [degrees]K). Based on the formulations given in table 5, the volume of the polymer and the number of mols of [CO.sub.2] generated can be calculated. The number of moles of [CO.sub.2] can be translated into a volume at the time of achieving the maximum foam temperature assuming that the pressure has equilibrated and is equal to the atmospheric pressure. This volume is then added to the volume of the polymer and a theoretical minimum density can thus be calculated. The blowing efficiency is then simply calculated as the ratio of the theoretical density over the actual density. It can be seen from table 5 that the blowing efficiency for TDI foams is approximately 72% while for MDI foams it is around 69%.

All the data shown here for MDI foams were obtained from formulations containing a compatibilizing agent; It was shown previously (see figure 4 ) that when Voranol 3137 was used on its own, the density was higher by 5%. For this standard MDI foam without compatabilizer the blowing efficiency would therefore be 8% less than a TDI based control.

A closer look at the maximum exotherm measured for MDI and TDI foams at the same isocyanate index, i.e., at the same number of moles of NCO shows that in all cases Tmax is higher for TDI foams than MDI ones. Again this observation is partly a reflection of the higher heat capacity of MDI due to the higher weight of the molecule. The decrease of maximum exotherm can be calculated from the heat of reaction of polyol/isocyanate which is -24 kcal/mol of NCO and the heat of reaction of the water/isocyanate reaction (-22.5 kcal/mol of NCO). It can be determined by DSC that the heat capacity for foams is in the range of 0.5 cal/g of foam for all the samples studied. Taking into account the mass of foam produced per mole of NCO and the heat of reaction quoted here, it can be calculated that replacing TDI with PMDI (31% NCO) should lead to a decrease of the maximum exotherm (based on the extra mass heated up) of about 5-10[degrees]C which is very close to our experimental findings (refs. 14 and 15).

To summarize this reactivity/compatibility study, it can be drawn from the above experiments that MDI is more reactive than TDI in the polyol/isocyanate reaction (especially when looking at Isonate 50, which is a two functional MDI like TDI and is soluble in the EO/PO based polyols like TDI but has a lower overall reactivity in the foaming process). This observation along with the different rise nrofile (early reaction faster with TDI and late reaction faster with MDI) suggests that the water/isocyanate reaction is slower in the case of MDI. Calculations of solubility parameters and microscopic observations indicate that the system polyol/water/MDI is not homogeneous which contributes to the worse water/MDI reactivity confirmed by the poorer blowing efficiency. The lower maximum exotherm and temperature at blow-off in case of MDI due to its higher heat capacity also partially explain the higher densities observed (lower gas volume due to lower temperature). The use of a compatibilizing agent helped to decrease the density by about 5% at 3 pbw water and improved the processability of the system.

Structure/property relationship.

We will now concentrate on the physical properties of the MDI foams and try to relate them to the polymer network.

After having selected a suitable PMDI, a solubility study was undertaken as described earlier to try and get a better compatibility of MDI with Voranol 3137 and water. It can be seen from the data in table 6 that the additive selected not only enhances the processing latitude but also helps to lower the density quite substantially by about 10% with 10 pbw of additive. It was seen earlier that only 5% density reduction was obtained at 3 pbw water, which is believed to be due to the limited increase of water/MDI solubility reached. The other obvious observation that can be made is the softening effect of the compatibilizer; however, it can be noted that these PMDI based foams are still very hard even at reduced copolymer polyol (Vornal 3943) levels and therefore, this point is not felt to be a drawback. Again, the foam porosity is rather low in all cases and foams with very low breathability can be manufactured with this technology without having the buns shrink. Compression sets obtained with MDI foams are also quite good but the other mechanical properties (T,T & E's) are lower than average for comparable density TDI foams.

        Table 6 - formulations and physical properties of MDI
              foams based on PMDI and conventional polyol
                 Voranol 3137 - ow water formulations
Formulation                        Unit     1      2      3
4
Voranol 3137                        pbw    80     75     70      95
Voranol 3943                        pbw    20     20     20
Additive 1                          pbw            5     10
5
Isocyanate index                          110    110    110     105
Water                               pbw   2.2    2.2    2.2     2.8
Airflow                             cfm   1.1    0.3     .2     1.2
Compression set 90%                   %   4.3    2.9    3.5     4.2
Density                  lb./[ft.sup.3]   3.5   3.04    2.9    2.65
IFD 25%                            lbs.   101   64.5     54    61.2
IFD 65%                            lbs.   214    131    106     127
Modulus                                   2.1      2      2     2.1
Hysteresis                                 74     79   80.4      74
Resilience                            %    29     31     22      36
Tear resistance                     pli   1.3   1.16   0.85     1.1
Tensile                  lb./[ft.sup.2]  22.6   18.2   16.8    17.6
Elongation                            %    97     98    104      88

Based on these results, another series of foams was run at higher water levels on our Varimax pilot machine. The formulations and results are gathered in table 7. 4.5 to 6 pbw were run in order to assess this technology in high water/low density ABA-free foams. It can be seen from the data that the physical properties vary depending on the formulation and the additives selected. The processability was still manageable on our Varimax machine at a total throughput of about 200 lb./min..

Table 7 - formulations and physical properties of MDI
foams based foams - high water/low density foams
Formulation            Unit

Voranol 3147           pbw                92    90
Voranol 3512           pbw                            95    95
Additive 1             pbw                             5     5
Additive 2             pbw                 8    10
Additive 1             pbw                           0.3   0.5
Isocyanate index       pbw               105    80   105   115
Water                  pbw               4.5     6   4.5   4.5

Airflow                cfm               0.9   1.0   0.3   0.9
Compression set 90%      %                30    37   8.4   113
Density               lb./[ft.sup.3]     2.1   1.5   1.9   2.1
IFD 25%               lbs.               172    85    73   106
IFD 65%               lbs.               332   176   137   236
Modulus                                 1.93   2.1     2   2.2
Hysteresis                                53  49.6    53    56
Resilience               %                37    35    33    40
Tear resistance        pli               1.4   1.1   1.2   1.4
Tensile                lb./[ft.sup.2]   25.4  14.5    23    28
Elongation               %                76    85    88    80

Very firm grades can be manufactured with this technology down to densities of at least 1.5 Ib./[ft..sup.3], i.e., at 6 pbw water with still safe exotherms (as discussed before MDI having a higher heat capacity than TDI, higher water levels can be used at same isocyanate index than TDI in order to reach the same exotherm.) Furthermore, lower isocyanate indices can be run with MDI based foams due to the better covalent network that can be built because of the higher functionality of PMDI.

The compression sets can be greatly altered by varying the formulation in terms of compatibilizing agent and amine package (via an additive).

The mechanical properties are still compromised, which might well be a trade-off of this technology, due to the higher functionality of the isocyanate used and the increased rigidity of the MDI molecule and resultant hard segments. The cell structure of the MDI foams is still very fine, giving the foams a rather plush feeling. This fine cell structure could be resulting from the higher viscosity of the initial polyol/water/MDI emulsion leading to better nucleation.

In order to better understand these physical properties differences, a more in depth structure -- property relationship study was undertaken.

FTIR study

Several FTIR spectra are grouped in figure 10. Our interest will focus on the carbonyl region, from 1,600 to 1,800 [cm.sup.-1]; the literature is rich in studies in this area and all the peaks have been thoroughly assigned (see for instance references 16 and 17).

It is very clear from the spectra that the TDI reference has highly organized hard domains and possesses strong hydrogen bonding within those domains, characterized by the intense peak at 1,645 [cm.sup.-1] (bidentate urea). This last peak shifts to the higher wavelengths when TDI is replaced by pure MDI foam (very difficult to process). It almost disappears when PMDI is used. This denotes a different hard segment organization in those networks. To try to understand this observation, dynamical mechanical spectroscopy was performed on those samples.

DMS study

DMS spectra of foams made at 3 pbw water are shown in figure 11. It can be clearly seen by looking at tan o that for the MDI based foam, the glass transition temperature of the soft phase is shifted to higher temperatures by 10-20[degrees]C and is very broad; this is characteristic for phase-mixed systems with a strong contribution of a higher crosslink density reached with MDI as compared to TDI. In other words, the polymers made out of MDI tend to be more phase mixed than the ones made out of TDI. This is further confirmed by the rubbery plateau of the storage modulus which is very horizontal for the TDI foam and more inclined with the MDI foam, which again is characteristic of thermal transitions due to poor phase segregation.

Swelling study

The foams were swollen in DMF ([delta] = 24.1 [MPa.sup.1/2][cm.sup.3] [mol.sup.-1]) which is a good solvent for the polyurethane networks. The maximum equilibrium swelling were recorded as well as the percentage of extractable polymer (solution fraction). The results are gathered in table 8. At 3 pbw of water the solution fraction for PMDI based foam is much lower than would be expected for a comparable TDI based foam (1.1% vs. 10%) (ref. 17). This reflects an improved covalent network in the PMDI foam, which can be related to the increased isocyanate functionality.

Table 8 - swell studies of MDI based forams in DMF

Isocyanate   Isocynate   Water level      Sol.        Swelling
               Index       (pbw)       fractions(%)      (%)
PMDI            105          3            1.1             13
PMDI            105        4.5            1.2           10.5
PMDI            105          6            1.7           16.1

At higher water levels. the solution fraction of PMDI foams remains very low, around 1-2%, while the swelling ratio increase substantially. This means that the PMDI helps to create a good covalent polymer network but the virtual network, usually referred to as hard segments, hydrogen bonded network or urethane-urea aggregates, is rather poor and can be partially swollen in DMF, thus increasing the percentage of swelling. Swelling experiments in DMF of linear elastomers (ref. 18) have shown that urethane networks can usually be completely dissolved in DMF, while urethane urea ones swell but do not dissolve, proving the difference of cohesion of these two bonds.

Skorpenske et al. in a previous study (ref. 17) showed that the compression set values depend on the quality of both the virtual and covalent network for TDI foams. In the case of MDI foams, it seems that the covalent network is of good quality even at high water/low MDI indices; however, the virtual network, consisting of the hard segments, is very weak as reflected by the high swelling and nonexistent bidentate carbonyl links (determined by FTIR). This is believed to be the consequence of a different hard phase configuration highlighted in the DMS study. This difference in phase composition could be due to:

* A different reactivity of MDI systems where the polyol/MDI reaction is favored versus the water/MDI one. As a consequence, the polyurea precipitation could be delayed or even absent.

* The higher functionality of PMDI versus TDI which will shorten the gel point thus favoring a thermodynamically less stable configuration to be gelled.

* The difference in compatibility of TDI and MDI in the fully formulated system.

General conclusions

This study was carried out to assess the value of MDI based slabstock conventional technologies and to study the structure-property relationship of those foams. The following conclusions could be drawn from this work:

* The foams can be processed on a large scale Maxfoam machine but the processing is more challenging than with TDI foams.

* Very firm grades can be produced without the need for any copolymer polyol. Special grades like low porosity foams or packaging foams can be manufactured with these technologies.

* MDI is less compatible with polyol/water and needs therefore a compatibilizing agent to achieve a reasonable processing latitude.

* The densities of MDI foams are always higher than those of TDI, because of the higher equivalent weight of MDI leading to a lower blowing efficiency. Major contributing factors are:

1) Poor solubility of MDI in polyol/water which can be somewhat improved by adding a compatibilizing agent (5 to 10% density decrease depending on the water level ) and

2) The cell opening which occurs at a lower water conversion.

* The reactivity of TDI and MDI are different: TDI reacts faster at the beginning of the reaction, while MDI is quicker at the end of the rise. Furthermore, MDI reacts faster in the polyol/MDI reaction but slower in the water/MDI one, suggesting that the water compatibility might be the key to lower foam densities.

* MDI based grades showed very promising physical properties including excellent compression sets at densities down to 2 lb./[ft..sup.3] and in very firm grades, without the need for any copolymer polyol. Mechanical properties were always compromised and so far, foams at the lowest densities (1.5 lb./ [ft..sup.3]) had borderline 90% compression sets of about 30%.

* Structure-property relationship studies highlighted that MDI foams have a much better covalent network than TDI foams (swelling, FTIR). On the other hand, the virtual crosslinking is much weaker and the morphology is much more phased mixed particularly in formulations with high water content.

[Figures 1 to 11 ILLUSTRATION OMITTED]

References

(1.) Duff; A. and M. Gansow, 1988. "Polyurethane foam chemistry for automotive seating," proceedings of the SPI 31st Annual Technical/Marketing Conference, Philadelphia, PA, pp. 420-25. (2.) Almqvist, A. "CFC elimination in flexible molded PU-foam for furniture and automotive applications. " (3.) Mispreuve, H. and P.M. Knaub. 1993. "New high resilience slabstock technologies based on high performance polyols," proceedings of the SPI Polyurethanes World Congress, Nice, France, pp. 297-304. (4.) Mispreuve, H. and P.M. Knaub. 1994. "MDI slabstock foams in luxury bedding: A perfect match," proceedings of Utech 94, Den Hagen, NL, pp. 37/1-5. (5.) Casey, M., I. Mueller, A. Parfondry and A. Elliot. 1992 "Comfort cushioning from ICl's MDI-based flexible slabstock technology, " proceedings of Utech 92, Den Hagen, NL, pp. 191-193 (6.) Smiecinski, T.M., S.E. Wujcik, D.C. Mente and E.H. McKenna. 1994. "Mixed isocyanates in flexible polyurethane slabstock foams." Proceedings of the 35th annual polyurethane technical/marketing conference, Boston, MA, pp. 326-331. (7.) Bailey, M.E., V. Kirss and R.G. Spaunburgh. "Reactivity of organic isocyanates," 1956. Industrial and engineering chemistry, 48-4 pp. 794-797. (8.) Lowenkron, S., private communication. (9.) Tabor, R.L., K.J. Hinze, R.D. Priester and R.B. Turner "The compatibility of water with polyols, " 1992. Proceedings of the SPI 34th annual polyurethane technical/marketing conference, New Orleans (LA) pp 514 525. (10.) Hildebrand, J.H. and R.L. Scott "The solubility of nonelectrolytes, " 1950. Reinhold Publishing Corp. New York. (11.) Barton, A.F.M. "Handbook of solubility parameters and other cohesion parameters, " 1983, CRC Press, Florida. (12.) Hansen, C.M. "The three dimensional solubility parameter - key to paint component affinities-I," J. Paint technol., 39, 104, 1967. (13.) Magat M., "Thermodynamics of solutions of high polymers," J. Chim. Phys., 46, 334, 1949. (14.) Woods G. "Flexible polyurethane foams - chemistry and technology," 1982. Applied Science Publishers Ltd. Essex, U.K. (15.) Lidy Werner. Internal publication (16.) Priester R.D. Jr., J.V. McClusky, R.E. O'Neill, R.B. Turner, M.A. Harthcock and B.L. Davis. "FT-JR - A probe into the reaction kinetics and morphology development of urethane foams," 1990. Proceedings of the SPI 33rd annual technical/marketing conference, Sept. 30-Oct. 3, pp. 527535. (17.) Skorpenske R.G., R. Solis, R.A. Kuklies, A.K Schrock and R.B. Turner "Compression set mechanisms in flexible polyurethane foam," 1992. Proceedings of the SPI 34th annual polyurethane technical and marketing conference, October 21-24, pp. 650-659. (18.) Knaub P.M.A. unpublished data.