Investigation of new low GWP blowing agent AFA-11 for PU/PIR

Abstract

The standards for energy efficiency have been constantly rising, and at the same time the rigid polyurethane foam industry is also facing ever increasing pressure to address climate change by eliminating global warming substances. Blowing agents are vital foam components and are responsible for the outstanding thermal performance of PU foams that are commonly used for thermal insulation in appliances and residential and commercial buildings.

To meet these challenges, Arkema is investigating a new range of blowing agents, the AFA series, designed for most PU applications including appliances, pour-in-place (PIP), spray, and polyisocyanurate (PIR) boardstock. AFA blowing agents are both liquid and gas, and have very low global warming and negligible ozone depletion potential.

The focus of this study was to evaluate solutions based on AFA-L1, which is a liquid under ambient conditions and is a candidate for replacement of the hydrofluorocarbon HFC-245fa and hydrocarbons. Using other blowing agents such as cyclopentane, HFC-134a, 245fa, and HCFC-141b (a hydrochlorofluorocarbon) as references, AFA-L1 was evaluated using an Edge Sweets high-pressure foam machine. AFA-L1 displayed a similar blowing efficiency, provided some improvement in dimensional stability, and an advantage in initial kfactor versus HFC-245fa and hydrocarbons. The solubility of AFA-L1 based agents in various polyols, and its compatibility with materials such as plastics, elastomers, and metals, was also evaluated.

Introduction

Since the phase-out of chlorofluorocarbons (CFCs) in the mid 1990s as a result of the Montreal Protocol, the rigid polyurethane foam industry in North America has faced a constantly evolving period of change in the availability and use of different blowing agents.

Although the regulations have imposed a significant cost burden on system suppliers and foam manufacturers due to the need to ensure that new blowing agents perform acceptably and that products conform to building code requirements, one aspect of the enforced changes is that the industry has developed a much greater understanding of the properties and performance attributes of different blowing agents. This has led to a greater degree of sophistication in use of combinations of liquid and low-boiling blowing agents. For example, liquid blowing agents are used in a wide range of applications including spray foams.

CFC-11 was a remarkably versatile blowing agent but was phased out due to its very high ozone-depletion potential (ODP). The hydrofluorochlorocarbon HCFC-141b was a good replacement, but its higher boiling point and its greater solubility in the polymer matrix led to some problems with dimensional stability requiring an increase of the foam density. The inclusion of the low-boiling blowing agent, HCFC-22, gave improvements in dimensional stability and a reduction in overall density.

However, the high ODP of these HCFCs also led to their eventual phase-out.

To fill the void created by these phase-outs, third-generation blowing agents, such as the hydrofluorocarbons HFC-34a and HFC-245fa were developed, commercialised, and have become widely used in many rigid PU foam uses in appliances, pour-in-place and spray.

However, the chemical industry is now facing ever-increasing pressure to address climate changes by focusing on carbon footprint and more specifically on the Global Warming Potential (GWP) of blowing agents. In Europe, the European Parliament has committed to adoption of the Kyoto protocol whose purpose is to reduce emissions of greenhouse gas by 8 percent, compared to the 1990 level, from 2008-2019.

Because of their high GWP values (1300 for 134a and 1020 for 245fa), HFCs are becoming regulated for some applications. For example, the European Mobile Air Conditioning directive prohibits use of HFC-134a in MAC systems for new platforms in 2011 and all new cars in 2017.

To address future market needs, Arkema is therefore contemplating the development of fourth-generation blowing agents to replace HFCs such as 245fa and 134a. To be fully environmentally acceptable, this next generation must have a zero ODP and a very low GWP while maintaining excellent general foam properties, including insulation performance.

Properties Arkema is investigating a range of new blowing agents, called the AFA series. Designed for most PU applications including appliances, PIP (pourin- place), spray, and PIR (polyisocyanurate) boardstock, these molecules possess very low GWP and negligible ODP. Among them, AFA-L1, which is a liquid under ambient conditions, is a potential candidate for replacement of HFC- 245fa and pentanes.

Table 1 summarises the properties of AFA-L1 and other blowing agents such as cyclopentane (cC5), isopentane (iC5), normal pentane (nC5), HFC-245fa, HFC-365mfc and HCFC-141b.

Physical properties — Besides its low GWP, AFAL1 did not exhibit either a flashpoint (FP) or flammability limit under ambient conditions.

Table 1Properties of AFA-L1 and references
	MWt	BPt	FP*	Lambda	(mw/m-°K)	LFL/UFL**	GWP
AFA-L1		>15	None	9	10	None
iC5	72	28	-51	13	14	1.4	11
nC5	72	36	-49	14	15	1.5	11
cC5	70	49	-7	11	13	1.5	11
F-365mfc	148	40	-24	11	12	3.4/13.0	782
F-245fa	134	15	None	13	13	None	1020
F-141b	117	33	None	9	10	7.4/15.5	700
*Flash Point **At ambient temperature

AFA-L1 exhibits a low gas-phase thermal conductivity (Lambda), comparable to F141b, which should contribute to the insulation value of the rigid polyurethane foam. Toxicity properties — The use of any new material requires a thorough risk and toxicity assessment. The toxicity assessment has a long path, including acute and chronic testing, which requires a strong commitment. Arkema has already started the toxicity evaluation of several compounds. Initial findings showed that AFA-L1 was not mutagenic in the AMES test. Experimental Last year, Arkema reported the dimensional stability and initial k-factor of PU foams blown with AFA-L1 compared to HCFC-141b, HFC- 245fa and a cyclo/isopentane 80/20 blend1. Since that initial investigation we have continued to accumulate aged k-factor data. Experiments were also performed to assess the compatibility of AFA-L1 with various plastics and elastomers, as well as its solubility/miscibility in a number of rigid foam polyols and polymeric isocyanates. Finally, the performance of AFA-L1 in foams processed on a high-pressure foam machine is being investigated. Table 2 summarises the formulations of that initial study. The blowing level/density and reactivity were set from the results obtained using HCFC-141b; gel time around 50 seconds and free rise density around 2.0 pcf.

Table 2: PU formulations
Formulation	HCFC-141b	HFC-245fa	Cyclo/ isopentane	AFA-L1
Voranol 490	18.27	18.05	18.85	18.09
Jeffol R-425X	10.96	10.83	11.31	10.85
Stepanpol PS-2352	7.31	7.22	7.54	7.24
Polycat 5	0.07	0.07	0.07	0.07
Polycat 8	0.37	0.37	0.37	0.37
Tegostab 8846	0.71	0.71	0.71	0.71
TCPP	2.36	2.36	2.36	2.36
Water	0.64	0.64	0.64	0.64
Forane141-b	6.27	0	0	0
HFC-245fa	0	7.19	0	0
Cyclo/isopentane (80/20)	0	0	3.75	0
AFA-L1	0	0	0	7
Rubinate M	53	52.6	54.4	52.7
Isocyanate index	115	115	115	115
Total blow ml/g	20	20	20	20
% physical blowing	60	60	60	60
% water blowing	40	40	40	40

In that earlier work1, physical test samples were made in 6x6x6” (15x15x15 cm) open-box pours by conventional hand mixing. One box each was made for k-factor testing and dimensional stability at 70°C/97 percent Relative Humidity (RH). Due to the nature of the free rise foams, k-factor samples were cut such that the foam rise was parallel to the test face, in order to minimise the effect of any defects running right through the sample thickness. Also, since the k-factor samples were undersized, 5x5”, each test piece was surrounded by like material in order to test a full 12x12x1” sample. Dimensional stability test samples were cut such that the foam rise was perpendicular with the 4x4” test face. Material compatibility Table 3 lists the elastomers and plastics tested for compatibility with AFA-L1. Three dog-bone shaped samples of typical dimensions 75 mm x 4 mm x 2 mm were prepared from each material. Each piece was measured then introduced into a test tube filled with AFA-L1.

Table 3: Elastomers and plastics
Elastomers
Neoprene
Polyacrylate
Viton fluoroelastomer
EPDM (ethylene propylene diene M-class rubber)
Hypalon (chlorosulphonated polyethylene)
Natural rubber
Silicone rubber
SBR (styrene-butadiene rubber)
NBR (nitrile rubber)
Plastics
PBT (polybutylene terephthalate)
PTFE (polytetrafluoroethylene)
PVC (polyvinylchloride)

#
The tube was sealed and placed in a water bath kept at slightly above the boiling point of AFAL1 for 5 minutes, 24 or 100 hours. Then the polymer was removed from the test tube and any change in dimensions measured. The samples were then subjected to a tensile test with crosshead speed of 50 mm/min and the distance between grips set at 30 mm. Solubility/miscibility experimental details One of the first considerations in using a new blowing agent in PU and PIR foams is to evaluate its solubility in polyols and isocyanates. This parameter for AFA-L1 was determined by two methods: 1) a simple miscibility experiment, where the amount of blowing agent in the polyol was steadily increased until incompatibility was reached, as determined visually; and 2) a more involved determination from measuring the vapour pressure of the blend at various levels of AFA-L1 addition. Solubility experiments The setup consists of a pressure vessel with magnetic stirring, a pressure and a temperature transducer. Temperature inside the vessel is controlled to within 0.1°C and the pressure to within 0.1 percent. In the vessel (volume about 100ml), 50g of polyol was loaded. The vessel was then placed under vacuum to remove air. The change of pressure in the metal cylinder was monitored to ensure that there was no leak. The blowing agent was introduced in the vessel with a specially designed gas syringe. The amount of blowing agent loaded was verified by measuring the weight of the syringe before and after introduction. The temperature of the vessel was maintained at 50°C (above the boiling point of the different blowing agents) and the speed was maintained at 300 rpm. The vapour pressure of the blowing agent was recorded as a function of time. It was important to allow time for the system to reach equilibrium. After reaching equilibrium, the amount of blowing agent dissolved in the polyol was the difference between the added blowing agent present in the polyol and the blowing agent present in the gas phase of the vessel. Then another small amount of blowing agent was added. The procedure was repeated several times until the pressure in the vessel was equal to the liquid-gas equilibrium vapour pressure of the blowing agent at this temperature (maximum attainable pressure at temperatures below the critical temperature of the blowing agent). Miscibility experiments Table 4 lists the various polyols and isocyanates examined for miscibility with AFA-L1. Blends were prepared by adding a predetermined weight of polyol to a 125ml (~4 oz.) clear Boston Round bottle with Taperseal lined cap.

Table 4: Miscibility study
OH	value*	Viscosity**
Glycerine based polyether polyols
Carpol GP-700	230-250	250
Carpol GP-725	230-250	250
Carpol GP-4000	39-42	700
Carpo lGP-4520	34-38	890
Amine based polyether polyols
Carpol TEAP-265	625-645	470
Carpol EDAP-770	757-783	56000
Jeffo lAD-310	2400
Sucrose based polyether polyols
Jeffol SG-360	360	3500
Jeffo lSD-361	360	2500
Jeffol SG-522	520	27000
Voranol 490	490	5500
Carpol SPA-357	335-365	2500
Mannich based polyether polyols
Jeffol R-425X	425	4500
Jeffol R-470X	470	8200
Sorbitol based polyether polyols		polyols
Jeffol S-490	490	9000
Aromatic polyester polyols
Terate 2451	240	3200
Terate 3510	240	6000
Stepanpo lPS-2352	240	3000
Terol TR-925	295-315	11000
Polymeric MDI	%	NCO
Rubinate M	31.2	190
Papi 580N	30.8	700

AFA-L1 was then added to the appropriate weight of polyol to obtain 5, 10, 15, 20, 25 or 30 weight percent of blowing agent, i.e. one bottle for each polyol and weight percent level of blowing agent. Weights of both components were adjusted in order to maintain a similar volume and headspace in each bottle. The bottle was immediately capped and placed on a roller mixer for several minutes until thoroughly mixed. The blends were allowed to stand for 24 hours before being reweighed, to ensure no loss of blowing agent, and then observations were made of the blend condition as follows: stable solution (clear); stable emulsion (cloudy, but not separated); or showing signs of separation. Observations were repeated after one week at room temperature. Machine runs Using an Edge-Sweets Flexamatic 25HP-BT highpressure foam machine, foams were run with AFA-L1, HCFC-141b, HFC-245fa, or a cyclo/isopentane (80/20) blend. The polyol blends or ‘B’ sides were blended and mixed with an air mixer in an open pail. For blowing agents with boiling points at or below room temperature, the blend, minus blowing agent, was conditioned at 10°C along with a container of the blowing agent, prior to final mixing. Machine parameters shown in Table 5 were based on the control system using HCFC-141b and kept constant throughout the runs. Table 5: Foam machine parameters Chem Temp (F) 70 Mix pressure (psi) 1800 Total 160 Mould Temp (F) k-factor 115 Lanzen panel 110 Results & discussion: k-factor testing As reported last year1 and reviewed in Table 6, AFA-L1 exhibited similar initial k-factor performance to HCFC-141b over all the temperatures tested and performed slightly better than HFC-245fa. Also AFA-L1 performed significantly better than cyclo/isopentane foam.

Table 6 :11 Initial k-factor for liquid blowing agents (BTU.in/ft2.h.°F)
Mean Temp (F)	HCFC-141b	HFC-245fa	Cyclo/isopentane (80/20)	AFA-L1
18	0.162	0.153	0.172	0.156
32	0.165	0.159	0.171	0.162
50	0.173	0.167	0.173	0.169
75	0.184	0.179	0.183	0.181
104	0.196	0.193	0.198	0.193

Table 7 shows the k-factor results after almost one year of ageing at room temperature (samples are 1” thick core foams).

Table 6 :11 Month aged k-factor for liquid blowing agents (BTU.in/ft2.h.°F)
Mean Temp (F)	HCFC-141b	HFC-245fa	Cyclo/isopentane (80/20)	AFA-L1
18	0.162	0.153	0.172	0.156
32	0.165	0.159	0.171	0.162
50	0.173	0.167	0.173	0.169
75	0.184	0.179	0.183	0.181
104	0.196	0.193	0.198	0.193

After almost a year, AFA-L1 is still performing well, exhibiting similar results to HFC-245fa and slightly better than both HCFC-141b and pentane, especially at the lower mean test temperatures. It is well known that the mean temperature at which a k-factor is measured can greatly affect the thermal performance of a foam depending on the state of the blowing agent in the cells2. In other words, a blowing agent with a higher boiling point will not perform as well at lower test temperatures due to condensation of the gas in the foam cells, making it a less effective insulating material. Also, it should be noted that since last year, we have had the opportunity to reproduce those initial results for AFA-L1 several times. Material compatibility results The materials shown in Table 3 were subjected to 5 minutes, 24 or 100 hours emersion in AFAL1 at a temperature slightly above its boiling point. Once the test sample was removed from the test tube, at the designated time, it was checked for percent change in weight, volume, elongation at break and tensile strength at break. Most of the materials were not degraded in AFA-L1, especially EPDM, Hypalon, SBR, NBR, PBT, PTFE, PVC and PA 6. However, at the longest exposure, some materials showed more than 35 percent swelling, such as polyacrylate, Viton, natural rubber and silicone rubber. Concerning mechanical properties, we noted a slight decrease in tensile strength at break and elongation at break, respectively, for plastics and elastomers. Solubility experiment results The increase of vapour pressure for different blowing agents in a polyester polyol. The solubility limit is reached when the pressure is equal to the vapour pressure of the blowing agent. For example, with AFA-L1 the solubility limit in this polyester polyol is ~35 parts or roughly 25 weight percent. Similar data was generated for a sucrose- based polyether polyol, resulting in higher levels of solubility, as is typical for polyether polyols, for all the tested blowing agents. Combining these two data sets shows the vapour pressure of AFA-L1, HFC-245fa and HCFC-141b with the two different polyols. We can see that for each blowing agent the increase of the vapour pressure will depend on the nature of the polyol. The vapour pressure will increase gradually with the blowing agent concentration when the affinity is high (for example polyether polyol); conversely, when affinity is low, vapour pressure increases more quickly at lower blowing agent concentrations (for example polyester polyol). For typical polyol compositions using a polyether and a polyester polyol, the vapour pressure will therefore be somewhere in the area between the two curves. Note that this area is very similar for AFA-L1 and HCFC-141b, meaning that the solubility of AFA-L1 is very close to the solubility of HCFC-141b. Also, this area is much larger for these two blowing agents than for HFC-245fa, leading to a higher flexibility when designing the foam formulation, especially in situations, such as in the spray foam industry, where the blowing agent is blended into the ‘B’ side or polyol blend of the system prior to packaging for shipment. Miscibility experiment results AFA-L1 was completely miscible in all the polyether polyols and both polymeric MDIs up to and including the 30 weight percent level, giving a stable clear premix. It was felt that the 30 percent level was a reasonable upper limit based on typical levels needed for most application densities. For the polyester polyols, AFA-L1 was completely miscible in the Terate and Terol polyols. With the Stepanpol, AFA-L1 was miscible up to and including the 20 weight percent level. However, both the 25 and 30 percent blends gave an opaque emulsion upon mixing, which separated overnight. This level was similar to that seen in the solubility experiment reported earlier. Machine runs A generic PiP formula as discussed previously was used. Total blowing agent level was 23.0 ml/g, free rise density was about 1.65 to 1.70 pcf, and gel time was approximately 30 sec. We used a 3”x14”x14” mould to determine k-factor. Linear regression was used to calculate minimum fill weight. The panel was packed to about 15 percent above minimum fill. The initial k-factor of AFA-L1 foam is better than those of cC5/iC5 and HFC-245fa at the full range of mean temperatures. Above 50°F, however, AFA-L1 foam has a slightly higher k-factor than HCFC- 141b foam while under 50°F AFA-L1 foam has a lower k-factor. Conclusions and path forward It was shown last year that AFA-L1 displayed a similar blowing efficiency, provided some improvement in dimensional stability, and exhibited a slight advantage on initial k-factor versus HFC-245fa. It was also shown to have a significant advantage in k-factor versus hydrocarbons. Those initial results have been confirmed in our continued study of this low GWP material. In this study, AFA-L1 continued to perform well as a potential blowing agent for PU/PIR foams. After a year of ageing, foams made with AFA-L1 showed comparable insulation performance to foams blown with HFC-245fa and better than those using HCFC-141b and pentane. Additionally, AFA-L1 showed excellent solubility with various polyols and good compatibility with various polymers Arkema has already announced its commitment to providing solutions in helping reduce the carbon footprint of HFCs. In the PU foam blowing agent area, Arkema’s intent is to pursue the development of the best low GWP blowing agent candidate. So far that candidate appears to be AFA-L1, although several others have shown improved effectiveness as a foam blowing agent. AFA-L3, L4 and L5 have shown some advantages over L1 and L2 as a low GWP foam-blowing agent for the entire range of rigid polyurethane foam segments and applications. References 1. Chen B., P. Bonnet, L. Abbas, J. Costa, and M.Elsheikh. 2008. “Investigation of New Low GWP Blowing Agents for Rigid Polyurethane Foams,”Proceedings of Polyurethanes Conference 2008. 2. Bogdan M., F. Moore Jr., and Jason Hoerter. 2003. “Meeting the Insulation Requirements of the Building Envelope with Polyurethane and Polyisocyanurate Foam,” Proceedings of the Polyurethane Expo 2003. 47-56. Suppliers Forane foam blowing agents, solvents and refrigerants, Arkema Inc.; Rubinate M, a polymeric methylene diphenyl diisocyanate (pMDI), Huntsman International; Jeffol polyols, Jeffcat catalysts, Huntsman Petrochemical Corp.; Carpol polyols, Carpenter Co.; Voranol polyols and Papi pMDI, Dow Chemical Co.; Terate polyols, Invista North America; Terol polyols, Oxid LP; Stepanpol polyols, Stepan Co.; Tegostab surfactants, Goldschmidt GMBH; Polycat catalysts, Air Products and Chemicals Inc.; Hypalon, Viton elastomers, DuPont Performance Elastomers llc; Zytel plastics, E.I.DuPont De Nemours and Co.