The present invention relates to the field of compositions prepared from a polyethylene resin prepared with a chromium-based catalyst system and from a bimodal polyethylene resin prepared with a Ziegler-Natta or with a metallocene catalyst system.

Polyethylenes (PE) of medium-to-high density (0.94-0.965 g/cm3) produced by chromium-base catalyst systems (Phillips-type) have been used in many blow- moulding applications because of their excellent processing behaviour that can be attributed, at least partly, to the presence of long chain branching (see M. P. Mc Daniel, D. C. Rohlfing, and E. A. Benham, in 'Polym. React. Eng., 11 (2), 101, 2003).

For the past thirty years, emphasis has been placed on improving the yield, but most importantly, on improving the compromise between environmental stress crack resistance (ESCR) and rigidity for polyethylene resins prepared with chromium-based catalyst systems such as described for example in EP-A- 857736 or in Stephenne et al. (V. Stephenne, D. Daoust, G. Debras, M. Dupire, R. Legras, and J. Michel, 'Influence of the molecular structure on slow crack growth resistance and impact fracture toughness in Cr-catalyzed ethylene- hexene copolymers for pipe applications', J. Appl. Polym. Sci. 82, 916, 2001). For a given Young E-modulus or for a given density (for unfilled samples), the ESCR failure time was substantially increased. The resins were thus classified into two groups in terms of an ESCR/rigidity compromise: the first generation Cr- based resins and the second generation Cr-based resins. The second generation Cr-based PE resins had longer failure times than the first generation resins. The second generation Cr-based resins, however had slightly reduced processing performances: this was observed for example by earlier onset of either surface or bulk melt fracture phenomena.

Another class of polyethylene resins, within to the same range of densities as the chromium-based resins has also been developed. These resins were typically designed for pipe applications but they were extended to blow-moulding applications such as, for example, for fuel tank applications. These resins were either bimodal Ziegler-Natta (ZN)-based polyethylene (as described in J. Scheirs, L.L. Bδhm, J. S. Boot, and P. S; Leevers, TRIP vol.4(12), p. 408, 1996) or metallocene-based polyethylene as disclosed for example in WO 02/34829. Such PE resins were endowed with excellent slow crack growth resistance: the ESCR was typically increased by one order of magnitude with respect to second generation chromium-based polyethylene resins. Processing, however was much more difficult and some high-throughput blow-moulding applications could not be carried out, mainly because of early appearance of melt fracture phenomena. Additionally, bimodal ZN-based or metallocene PE were much more expensive than the first generation Cr-based PE resins.

The first and second generation chromium-based polyethylene resins remained far below the ZN-based PE resins in terms of ESCR. A slight but insufficient improvement was obtained by using Cr-based PE resins produced in double-loop reactors as disclosed for example in EP 952165.

Blends of Cr-based polyethylene with low amount of resin having a high ESCR were then used. These other resins were for example selected from linear low density metallocene-produced polyethylene (mLLDPE) having high molecular weight (HMW). Such blends as disclosed for example in EP 1319685 had an improved ESCR. The gain in ESCR with respect to the Cr-based PE resin alone was important for amounts of high ESCR resins typically of more than 10 wt%, but the rigidity was severely reduced. For example, the rigidity incurred a drop of over 0.003 g/cm3 for blends comprising 90 wt% of Cr-based PE having a density of 0.957 g/cm3 and 10 wt% of metallocene-produced linear low density polyethylene (mLLDPE) having a density of 0.927 g/cm3.

Even less desirably, adding a HMW mLLDPE resin to the Cr-based resin led to a considerable loss of processing behaviour. Furthermore, HMW mLLDPE microscopic inclusions can be detected with optical microscopy when the blend is extruded on a typical extruder. The concentration of these inclusions can be quantified by image analysis and amounts to more than 1 %.

Alternatively, blends of first and second generation Cr-based resins could be prepared. In order to obtain a significant increase in ESCR failure time with respect to that of 1st generation Cr-based resin, it was necessary to blend at least 50 wt% of a second generation Cr-based resin to a first generation Cr-based resin, thus resulting in a composition having reduced processing performances.

There is thus a need to provide, with a simple and cost effective method, polyethylene resins having a better ESCR/rigidity compromise than Cr-based PE resins while keeping an excellent processing behaviour in blow moulding applications.

It is an aim of the present invention to prepare a composition, based primarily on a chromium-based polyethylene resin that has improved ESCR while keeping a good rigidity.

It is also an aim of the present invention to prepare a polyethylene composition that is easy to process.

It is another aim of the present invention to provide a method for preparing a Cr- based polyethylene composition that is simple and cost effective.

Accordingly, the present invention provides a blend comprising - from 70 to 98 wt%, preferably from 70 to 95 wt% and more preferably from 75 to 92 wt%, based on the weight of the total composition, of a chromium-based polyethylene resin; and - from 2 to 30 wt%, preferably from 5 to 30 wt% and more preferably from 8 to 25 wt%, based on the weight of the total composition, of a bimodal polyethylene resin prepared either with a Ziegler-Natta or with a metallocene catalyst system.

The chromium-based polyethylene resin typically has a density of from 0.934 to 0.965 g/cm3, preferably of 0.940 g/cm3 to 0.965 g/cm3, and a high load melt index (HLMI) of from 1 to 100 dg/min.

The bimodal resin is preferably prepared with a Ziegler-Natta catalyst system. It has a density ranging from 0.940 to 0.962 g/cm3 and a high load melt flow index (HLMI) of from 4 to 40 dg/min, preferably of from 6 to 26 dg/min.

The density is measured following the method of standard test ASTM 1505 at a temperature of 23 0C and the melt flow index is measured following the method of standard test ASTM D 1238 at a temperature of 190 0C, under a load of 21.6 kg for the high load melt index HLMI and under a load of 2.16 kg for MI2.

Rigidity as measured by tensile or flexural tests correlates linearly with density for polyethylenes with densities above 0.93 g/cm3 and usual alpha-olefins as comonomers, moulded by compression. This was proven by many authors (see for example, K. Jordens, G. L. Wilkes, J. Janzen, D. C. Rohling, and M. B. Welch, Polymer, 41, p. 7175-7192 (2000)). For our PE grades, the following correlations were obtained respectively between the flexural ASTM D-790 E modulus (0.4 % strain) and density and between the ISO R527/A tensile E modulus (1 %) and density:

Flexural ASTM E Modulus (MPa)= 36,101 x (density in g/cm3) -33,068. The relationship was derived from 24 data points in the density range of from 0.930 to 0.964 g/cm3, and with a correlation coefficient r2 of 0.96.

ISO Tensile E Modulus (MPa)= 27,020 x (density in g/cm3) - 24,656. The relationship was derived from 23 data points in the density range of from 0.930 to 0.960 g/cm3 and with a correlation coefficient r2 of 0.96.

The environmental stress crack resistance was measured following the method of standard test ASTM D-1693-70 condition B with 10 % antarox (ESCR10) or with 100 % antarox (ESCR100). The failure time for ESCR10 is from 2 to 3 times smaller than that for ESCR100. The test ESCR100 is thus preferably selected for the least resistant resins in order to better discriminate their ESCR performances.

In the present invention, the bimodal resin added to the chromium-based polyethylene resin has a higher density than the mLLDPE added in the prior art compositions. Contrary to the prior art compositions, the blends of the present invention have an excellent rigidity. In addition, the mLLDPE of the prior art had a high molecular weight that resulted in a lower HLMI of the final blend than that of the starting chromium resin. In addition the presence of this high molecular weight fraction resulted in reduced processing performances. The bimodal resins added in the present invention have both a low molecular weight HDPE fraction favourable to easy processing and a high molecular weight LLDPE fraction favourable to good ESCR characteristics. Usuallly, ESCR of these bimodal ZNPE resins is so good that no failure occurs when submitting the bimodal PE resins to ESCR conditions of Cr resins of same density. In this invention, failure time for given ESCR condition of the bimodal PE resin is preferably at least ten times longer than the corresponding ESCR of Cr resin of same density, more preferably, it is at least 50 times longer and most preferably, it is at least hundred time longer. When the bimodal resin is extruded on a typical extruder under normal conditions, HMW LLDPE microscopic inclusions can be detected by optical microscopy. The amount of such HMW LLDPE inclusions is preferably less than 5 % and more preferably less than 2 %.

The bimodal resins additionally have a broad molecular weight distribution that is compatible with that of chromium-based resins. The molecular weight distribution is defined by the polydispersity index D that is equal to the ratio Mw/Mn wherein Mw is the weight average molecular weight and Mn is the number average molecular weight. The polydispersity of the bimodal polyethylene resin typically ranges from 6 to 25, the low end being associated with the metal locene-prepared resins and the high end with the Ziegler-Natta resins. The polydispersity of the chromium-based polyethylene resin ranges from 7 to 10 for the first generation resins and from 10 to 25 for the second generation resins.

The preferred bimodal polyethylene resins are Ziegler-Natta resins produced in a double loop reactor

The polyethylene blend of the present invention simultaneously - keeps the good processing properties of chromium-based polyethylene resins, preferably of first generation chromium-based polyethylene resins - gains the good ESCR characteristics of the bimodal polyethylene resin keeps good rigidity properties - exhibits homogeneous dispersion of high MW LLDPE species and contains less than 1 %, preferably less than 0.5 % of inclusions.

In the prior art compositions, the added high molecular weight linear low density polyethylene was difficult to disperse into the Cr-based resin and inclusions could be detected by optical microscopy. The bimodal polyethylene resin of the present invention can easily be dispersed into the Cr-based resin and thus does not suffer from the prior art problems. The composition can be prepared by introducing the bimodal resin directly into the main extruder hopper: it can be introduced either as a fluff or as pellets or as a master batch if fillers and pigments are added at the same time. If this method is used, the amount of added polyethylene is preferably of less than 20 wt% based on the weight of the total composition.

Alternatively, any conventional method of melt blending can be used.

The resin blends of the present invention are useful in blow moulding applications requiring high ESCR resistance together with easy processing and good impact strength such as for example fuel tanks or large containers and in the production of pipes or tubing at high throughput rates.

List of Figures.

Figure 1 represents the compromise environmental stress crack resistance ESCR expressed in hours at a temperature of 500C and with 100 % antarox, as a function of density of the resin or of the resin blend expressed in g/cm3.

Figure 2 represents the die swell expressed in % as a function of shear rate expressed in s'1 for several resins and resin blends. It must be noted that the tests are typically interrupted just before melt fracture occurs.

Figure 3 represents the shear stress expressed in Pa as a function of shear rate expressed in s"1 for several resins and resin blends. The vertical arrows indicate the occurrence of melt fracture for each blend.

Figure 4 represents the critical shear rate at the onset of melt fracture expressed in s-1 as a function of the logarithm of HLMI expressed in dg/min for several blends. A good correlation between critical shear rate at onset of melt fracture and logarithm of HLMI is observed for blends having a concentration of polyethylene modifier of 8 wt%.

Examples.

Several resins and resin blends were tested for environmental stress crack resistance, for die swell, for onset of melt fracture and for traction.

The starting first and second generation chromium-based polyethylene resins were respectively: Resin R1 is a first generation resin sold by ATOFINA under the name Lacqtene® 2002 SN58; Resin R2 is a first generation resin sold by ATOFINA under the name Lacqtene® 2003 SN53; - Resin R3 is a first generation resin sold by ATOFINA Research under the name Finathene® 47100 Resin R14 is a second generation resin sold by TOTAL PETROCHEMICALS under the name Finathene® SR572 Resin R15 is second generation resin sold by TOTAL PETROCHEMICALS under the name Finathene® 53080 Resin R16 is a high MW chromium-based resin of HLMI =2 dg/min sold by TOTAL PETROCHEMICALS as fluff HDPE 56020 S.

The modifier resins were respectively: - Resin R4 is a monomodal metallocene-prepared PE resin, sold by ATOFINA Research under the name Finacene® ER2281. - Resin R5 is a bimodal Ziegler-Natta PE resin, sold by ATOFINA Research under the name Finathene® BM593. - Resin R6 is a bimodal Ziegler-Natta PE resin, sold by ATOFINA Research under the name Finathene® XS1ON - Resin R7 is a second generation Cr resin of low density (0.934 g/cm3) sold by ATOFINA Research under the name Finathene® HF513 - Resin R8 is a monomodal high molecular weight (HMW) ZN LLDPE produced as described in EP 0989141 example 4 - Resin R9 and R10 HMW mPE produced as described in example 7 of EP0989141 - Resin R17 is a ZN LLDPE sold by TOTAL PETROCHEMICALS under the name LLDPE 4010 FE18 - Resin R18 is a bimodal mPE of HLMI = 11 dg/min and density = 0.950 g/cm3 as described in WO02/34829

Several second generation Cr resins sold by ATOFINA (Finathene 53080, 53140, SR572, SR583, SR523) were used to construct ESCR100-density curve represented in Figure 1

For comparison with Finathene 47100 (resin R3), a second generation chromium-based polyethylene resins R11 sold by ATOFINA under the name Lacqtene® 2001 TN46 was used.

Two commercial PE resins of 2nd generation (SR523-type) R12 and R13 were also modified by addition of some of the above modifiers.

The properties of these resins are summarised in Table I. TABLE I.

All blends were produced on a twin-screw Brabender extruder TSE 20/40 under nitrogen blanketing at a temperature of 210 0C, using a screw profile that gave similar state of dispersion of high MW LLDPE as encountered on bimodal ZN or MPE extruded on industrial extruder such as for example Werner Pfeiderer Z5K58. The state of dispersion is estimated by determining with image analysis the area fraction of dispersed high MW nodules as observed by optical microscopy on a thin film of bimodal PE resin. A good dispersion is obtained when area fraction is below 1 %. For all resins and resin compositions, the processing performances were appraised by determination of their die swell behaviour and of their critical shear rate and stress at the onset of melt fracture. For the die swell determination the die had a length L over diameter D ratio L/D of 10/2. For the onset of melt fracture determinations, the die had a ratio L/D of 15/0.752 and the test was carried out at a temperature of 210 0C in a Gottfert capillary rheometer.

A dispersion analysis was carried out on all the blends containing 8 wt% of modifier PE resin: the dispersion was excellent and less than 0.4 % of inclusions were observed. This was much lower than the percentage of inclusions observed for blends of the same starting Cr-based PE resin with 8 wt% of HMW mPE or ZN PE resins having similar HLMI.

The ESCR results are displayed in Figure 1 and Tables Il to IV. Table Il displays for resins R1 and R2 (1st generation Cr PE) and their blends, the ESCR, density and, in some cases, tensile test results as well as shear rate at onset of melt fracture (MF). TABLE II.

1 R2 in the second batch had a density of 0.952 g/cm3 instead of 0.954 g/cm3 in the first batch 2RI 8 had 16 % of HMW inclusions; blend 92/8 of R2/R18 had 4 % HMW inclusions; blends 92/8 of R2 with HMW PE R8, R9 and R10 had respectively 2.3, 3.3 and 7.3 % of HMW inclusions. Figure 1 represents ESCR100 expressed in hours at a temperature of 50 0C as a function of density expressed in g/cm3. The first generation chromium-based polyethylene resin all fell on the left curve labelled 1st generation type and exhibited the expected behaviour of decreasing ESCR with increasing density. The second generation chromium-based polyethylene resin, such as resins R12 and R13, all fell on the right curve labelled 2nd generation-type and exhibited the same trend as the first generation resins, but with substantially higher ESCR values at equivalent densities. As can be gathered from Tables Il to IV and illustrated in Figure 1, when the first generation chromium-based polyethylene was compounded with 8 wt% of a monomodal metallocene linear low density polyethylene (mLLDPE), the ESCR was substantially increased with respect to that of the starting chromium-based PE resin. The same first generation chromium-based polyethylene resin, when compounded with 8 wt% of bimodal Ziegler-Natta polyethylene resin (blend C2) had about the same ESCR but without reduction of density leading therefore to a ESCR/density curve substantially shifted to the right. When the first generation chromium-based polyethylene was compounded with more bimodal Ziegler-Natta polyethylene resin, with amounts up to 25 wt%, the ESCR performances increased nearly exponentially. For example, ESCR performances were found to equal those of the second generation chromium-based polyethylene resins such as R12 or R13 at 16 % level. Adding 24 % of bimodal Ziegler-Natta polyethylene (ZN PE) resin further increased the ESCR100 of the blends while keeping processing performances at levels above those registered for the Cr-based resins having the best ESCR performances, at similar values of HLMI.

For comparison, blending the same first generation resins with 8 % high molecular weight (HMW) PE resin (HLMI <2) led to a drop of processing that rendered these resins unsuitable for typical applications of Cr PE resins. Furthermore, the dispersion became insufficient, leading to inclusions of HMW PE nodules. Blends were also prepared from 8 wt% of bimodal ZN PE (R6) and 92 wt% of 2nd generation Cr PE, as displayed in Table 111. Their ESCR100 was equivalent to that of blends prepared with 24 wt% of the same ZN PE resin (R6) and 76 wt% of 1st generation Cr PE resins.

TABLE III.

From these observations, it was concluded that for both types of chromium- based PE resins, at equivalent density, blends with bimodal ZN PE according to this invention have higher ESCR than the starting chromium resin or than blends with HMW resins.

Shear rate at onset of melt fracture was not strongly reduced by the addition of a bimodal ZN PE and it remained acceptable: the lowest onset of melt fracture shear rate (for 24 % resin R6) was found to be close to 850 s -1, a value that is close to that observed for a second generation Cr with excellent ESCR performance and having same HLMI.

Furthermore, as observed for blends with the 1st generation chromium-based PE resins, the die swell behaviour (percentage and maximum shear rate before melt fracture) was not strongly affected by blending up to 20-24 % of bimodal PE resin of this invention.

Table IV presents ESCR results obtained from blends of first generation Cr- based PE with ZN PE or rnPE as compared to second generation Cr-based PE. First generation Cr-based PE resin R3 modified with 8 % bimodal ZN PE (R5 and R6) of this invention exhibit higher ESCR10 at comparable density than those modified with mPE (R4) or than unmodified second generation Cr-based PE resins (R11). Also, for trie low HLMI resin R3 and its blends, the gain in processing with respect to a second generation Cr-based PE resin (R11) was substantial.

TABLE IV.

In conclusion, adding up to 16 wt% of bimodal ZN PE (R6-type) to first generation Cr-based resin (R3-type) produced resins with ESCR behaviour comparable to that of second generation Cr-based PE resins (R11-type) with a substantially improved processing behaviour.

This can be seen in Figures 2 and 3 representing the die swell and melt fracture onset as a function of shear rate.

Figure 2 represents the die swell as a function of shear rate. It shows that the bimodal ZN resin R6 alone had a very small die swell but that melt fracture occurred at very low values of shear rate, of the order of 150 s"1. The second generation chromium-based PE resins R12 had a large die swell, rapidly increasing with shear rate and melt fracture occurred at fairly low values of shear rate, of the order of 275 s"1. The first generation chromium-based PE resins R1, R2 and R3 had a die swell that fell between that of the bimodal ZN resin R5 and that of the second generation chromium-based resin R12, but no melt fracture was observed at the shear rates that were tested (up to 800 s"1). As can be seen, blends prepared from first generation chromium-based PE resin R2 with respectively 8 wt%, 16 wt% and 24 wt% of bimodal ZN PE resin R6 all exhibited the same excellent melt fracture behaviour as the first generation chromium- based PE resins, i.e., no melt fractures were recorded up to 800 s"1, and additionally, their die swell was reduced at low shear rate. Figure 2 also shows that the difference in die swell behaviour is primarily related to the type of starting resin. Indeed, it was observed that all blends prepared with the same type of starting resin remain grouped and were significantly separated from groups prepared with different types of starting resins.

Figure 3 represents the shear stress as a function of shear rate at 210 0C, for a die having a length (L) over diameter (D) ratio L/D of 15/0.752. Figure 3 also shows that for the same blends the shear stress versus shear rate curve was not strongly affected at shear rates typically encountered during processing (from 100 to 1500 s"1). It was fu rther observed that shear rate at onset of melt fracture linearly decreased with increasing concentration of bimodal ZN PE R6 (see Table II) but this decrease was smaller than that observed for blends of the same first generation Cr-based PE resins with HMW rnLLDPE (resins R9 or R10) or with HMW ZN PE (resin R8). Also at a given shear rate, related to the throughput of the extruder, the shear stress, or pressure at the die, was much higher for blends with HMW LLDPE modifiers.

Figure 4 representing the critical shear rate at the onset of melt fracture as a function of HLMI shows a semi-logarithmic relationship between the critical shear rate at melt fracture and HLMI for a given concentration of polyethylene modifier of 8 wt% in this test. On the same Figure, the drop of shear rate at onset of melt fracture with concentration can be visualized for blends of resin R2 with various concentrations of bimodal ZN PE resin R6. One can see that a blend with 8 % modifier having HLM1 1.5 would give similar shear rate at onset of melt fracture than the blend containing 24 % R6. It is thus much less detrimental for processing to blend a given Cr resin with a bimodal PE resin of this invention. During extrusion of the blends, the extruder head pressure torque, the melt temperature at the die and the mass throug hput were recorded. From these measurements, the extruder specific energy was calculated using formula I developed by Martin (Martin C. : Twin-screw extruder, ch.2, in "The.SPE guide on extrusion technology and troubleshooting", J. Vlachopoulos and J. R. Wagner, eds, SPE, Brookfield, 2001.)

Specific energy (kWh/kg) = ( kWm x ε x T% x RPM ) / (Q x RPMmax) wherein - kWm is the motor kilowatt rating = 2O.76 kW for Brabender TSE 20/40 - ε is the system efficiency = 0.954 - T% is the torque ratio = torque/max torque, with the maximum torque = 200 Nm - Q is the output rate expressed in kg/ri - RPM is rotation per minute The results are summarised in Table \/ for an extruder Brabender TSE20, having a screw profile adapted to give same dispersion as industrial extruder and temperature settings of: 200-200-200-200-200 0C and a die diameter of 4 mm..

TABLEV.

It can be concluded from Table V that blends with second generation chromium- based PE resins require more specific energy than those prepared with first generation chromium-based PE resins: typically about 0.4 to 0.5 kWh/g for second generation chromium-based PE resins versus 0.3 to 0.35 kWh/g for first generation chromium-based PE resins. For comparison, blends prepared from first and second generation chromium- based polyethylene resins were prepared in order to obtain compositions having improved ESCR. Observable increase in ESCR was observed for blends containing at least 50 wt% of second generation resin R13. The addition of such large quantities of second generation resin led to deterioration of processing compared to 1st generation Cr resin.

Traction tests were also carried out on the polyethylene resin compositions according to the present invention (Tab Ie II). The addition of 8 wt% of bimodal ZN PE resin R6 to first generation Cr-based resins produced blends having an E- modulus and a yield stress higher than ihose observed for blends prepared from the same first generation Cr-based resins with 8 wt% of either HMW mPE or HMW ZN PE. Interestingly, when the amount of bimodal ZN PE (resin R6) in blends with first generation chromium-based polyethylene resin R2 increased, ductility of the resulting composition also increased as testified by the increase of percentage strain at break. The same ductility behaviour was observed with the addition of HMW PE (resins R8 to R10) but these blends suffered from reduced processing capability, rigidity and homogeneity.

Another application of this invention is to produce blend of very high MW chromium-based PE with bimodal ZN PE as described above that can be pelletised and still have acceptable impact strength with enhanced ESCR and good processing. These polyethylene compositions can be used in extrusion blow moulding of large containers that can withstand impact at low temperature. For example, extrusion at 215°C, 80 RPM of HMW chromium-based PE resin R16 on Brabender TSE20/40 as described above resulted in melt fracture. Blends comprising 76 % of the same R16 resin and 24 % of bimodal resin R6 could be extruded without melt fracture and with a reduced specific energy: it dropped from 0.58 kWh/kg for R16 alone down to 0.3 kWh/kg for the blend. ESCR10 was increased from 46 h for R16 alone to 70 h for the blend. The impact strength {notched Charpy ISO) at -250C dropped somewhat but remained at an acceptable level.