CA1332987C

CA1332987C - Process for chemical treatment of discontinuous cellulosic fibers and composites of polyethylene and treated fibers

Info

Publication number: CA1332987C
Application number: CA000597923A
Authority: CA
Inventors: Govinda Raj; Bohuslav V. Kokta
Original assignee: Individual
Current assignee: Individual
Priority date: 1989-04-19
Filing date: 1989-04-19
Publication date: 1994-11-08
Anticipated expiration: 2011-11-08
Also published as: US5120776A

Abstract

Cellulosic fibers were pre-treated with maleic anhydride or phthalic anhydride in the presence of an initiator to improve the bonding and dispersibility of the fibers in the polymer matrix. High density polyethylene (HDPE) was filled with chemically treated chemithermomechanical pulp (CTMP) and wood flour. Composites of HDPE-pre-tretaed wood fibers, characterized at different fiber ratios, produced superior mechanical properties than the untreated fiber composites. These composites can be compression or injection molded to produce useful articles.

Description

~ ~ - 1 3 3 2 ~ 8 7 PRa~E;S E~. -'~TlqU. ~ r)T~ L INl~-L`i '~T.T~IT~:Tc E~.S
AND ~__J~ll~ JF pnT.YFTRYTFNR AND TRE~rED FI~ERS

., This invention relates to the process of ~' 'f~Al trl ' ' of cellulosic ' fibers to i ,-ov~ the bonding with the polymer matrix, and thereby in improving the - '- jf~A1 properties of thP llA~tic~-wD~d fiber ~ , 't~
, ' It is well known that in~L~.dtion of ~ nntim~ fibers with polymeric '`~ materials can impart improved properties such as greater strength and ', ,'-; stiffn~ to the polymer matrix. Such mdterials are described, for example, in "~ ' U.S. Patent no. 3,764,456 to ~: '' and in U.S. Patent no. 4,442,243 which describes reinforced mica-th~ tic , , t~ having i _ ~v~d physical '',- ~ ~.~pe.~ies and durability. The ~ ''nAtinn of di~o~ntim]m~Y cellulosic fibers with a variety of vul~Ani~d elastomers i~ described in U.S. patent no.
, ~ 3,697,364 to Ebustany and _oran.
,~ The use of inorganic fillers such as mica and glass fiber posses many -. .
d;ffif~]lties during the fabricAtion. Due to its abrasive nature these fillers cause more wear to the ply~ ..ng ~inf-ry and also since these fillers are ` , 40 ,-.~ brittle they suffer extensive bLc~h_Je during l , ~ing. Many of the above -............................................................................ .
", , ,onf~ problemg can be greatly reduced when organic fillers were used. The great Ahl]nflAnf~e and ~ L.,~s of cellulosic materials make them one of the ', ~ttrA~tive choice as a low cost fillers in polymers. The r]hli~hpd lit,erature , -~
contAin~ nu~ber of references for the u e of cflll]losif~ fillers as additives ~ -' - 50 `~ for both I'~~ - '~ and thenmoplastic polymers.
- Although the use of ~f~ lo~if~ filler~ in th~- -3 t resins has been known for flf~fl~s, their use in thermoplastics has been limited as a result of . . ~
problems in dispersing the filler in thermoplastic matrix and the lack of ~ f~h 'oAl bonding between the filler and polymer matrix. It has b~en shawn by '" " ~ 60 ~; Hamed in U.S. Patent no. 3,943,079 that the dispersion of discontin-'',~ cflll~lo~if fiberg in the polymer matrix can be greatly ' _ ~v~d by ~.,'"' ~

~ ,. ' '~ .

` :

- ; ~ pre-L~. ' ' of the fibers with a polymer and a lubricant. The French patent no. 76,34301 describes the preparation of a new materidl, property of which can be ad~usted to A~ArtAt-ion of different ar~ ationQ, which can be formed by a copolymer or a li~nn~J~ lo~i~ material and it is bond to a polymer by grafting in a manner that can play a reinfo~ role to the polymer.
Goettler in U.S. Patent no. 4,376,144 has shown that the ~A~h~Qion of ! discontinuous cellulose fibers to a matrix of vinyl chloride polymer can be '~ s~lhstAntiAlly i __~v~ by il~oI~aIdting tberewith a bonding agent which iQ a .- . . -cyclic trimer of toulene diisocyanate. The bonding agent also i __uv~l the - - dispersibilty of the treated fibers into the matrix material. The bonding agent has been found effective at relatively lcw ~nnr~ntrations-as low as 0.1 .
- parts by weight on 100 parts by weight of tbe vinyl chloride polymer in the matrix. Coran and Patel has found that U.S. Patent no. 4,323,625, treated ~-~
fibers comprising d;~rnn~imv~]~ cellulosic fibers of aspect ratio greater than ,. ~ - , :
five, and metylol phenolic -'ifi~ crystalline polymer from alpha olefin -,-- -r having 2-4 carbon atoms, said -'ified crystalline polymer being :
~Ie~ent in an amount sufficient to reduce fiber-to-fiber interactions up to , :
- about 85 parts by weight per 100 parts of fibers by weight, have useful . ~ ,.
properties. Advantageously, the fibers can be oriented to a greater or less degree, providing pl~u~Ls having a greater strength and stiffn~Q~ in the direction of ori~nt~Ation.
.. . .
Gaylord described in U.S. Patent no. 3,645,939 a process for ihili~ing a material con~Aining free hydroxy groups with a polymer which :
is otherwise ir _ tible, by bringing them to~th~r in the LLe~l~e of an ethylenically uusdLurd~ed carboxylic acid, substituted carboxylic acid or carboxylic acid anhydride and a preradical pl`~u~OL- or preradical ~æ.d~ing .~ ~
agent. T~hr~-i~.z et al., in U.S. patent no. 4,107,110 described that cellulose - fibers, coated with a grafted copolymer comprising 1,2-poly~ ~A~i~n~ to which ~ is grafted an acrylate such as butyl '~rylate could be used in reinforcing - - of polyethylene and other plastic ~ _ ti~nQ.
-~: ; 60 ,, :

,'~
;,". ,~

- ~

_.
SUMMARY OF THE INVENTION

According to the present invention, composites are made of chemically pre-treated discontinuous cellulosic fibers with a thermoplastic polymer has superior mechanical properties. More specifically, the thermoplastic composites consists of 90 to 50 weight percent of high density polyethylene (HDPE), the remaining 10 to 50 weight percent being the discontinuous cellulose fibers, pre-coated with a mixture of thermoplastic polymer 0 to 5 weight percent of fiber, and a carboxylic acid anhydride 0 to 6 weight percent of fiber and an activator 0 to 1 weight percent of fiber.
The term thermoplastic polymer includes polyethylene, polypropylene and their copolymers. The term cellulose fiber includes cellulose fiber derived from soft wood, hard wood pulps, e.g. chemical or mechanical or chemimechanical or high-yield of thermo-mechanical or refiner or stone groundwood or chemithermomechanical pulp or explosion pulp. The paper-mill waste pulp and recycled paper pulp can be used. Preferred are wood flour, chemithermomechanical pulp (CTMP). Hard wood fiber, because of its smaller diameter as compared with soft wood fiber is preferred. The fibers having an average fiber aspect ratio (average length to average diameter of the fibers) of 2-200 of which 5 to 150 being preferred. Mixtures of fibers having different average aspect ratios can be usefully employed. The fibers can be pre-treated before incorporation into the polymer matrix to improve the bonding between the fiber and matrix and also to reduce the fiber-to-fiber interactions. Pre-treatment of the fibers is effective in reducing the time and work required to incorporate the fibers in the polymer matrix.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

The composites of the invention comprise pre-treated discontinuous cellulose fibers mixed with the thermoplastic polymer. The pre-treatment of fibers can be carried out by coating with a bonding agent, an unsaturated 1 33~.sa7 carboxylic anhydride in the ~rese--ce of an activator. Maleic anhydride or phthAlic anhydride is the preferred bon~ing agent. Suitable activators are dicumyl peroxide, benzoyl peroxide and di-t-butyl peroxide. A polymer can be ~Pl~rtPd which i8 m~lten at the mixing t~ _ d~U~ employed and which acts to coat the fiber and to prevent fiber-to-fiher interactions, e.g. polyethylene.
The bonding agent of the invention include those agents which have been found to be effective in enhancing A~h~Qion with c~ Qir materials, for example, an ethyl~nPriAlly ~ 1 r~l~ carboxylic acid, ~lhstitl~d carhoxylic acid or carboxylic acid anhydride. Generally, an amount of 0.5 to 10 parts by weight in 100 parts by weight of the fiber is ~lfficiPn~ to optimize A~hP~ion of the fiber to the matrix. The bonding process can be explained as follQws;
maleic anhydride, for example, reacts with OH groups of cellulose in the presence of an activator and a polymer, which acts as a binder, to form a cellulose maleate half ester. The half ester subsequently reacts with ; polyethylene in the presence of a free radical initiator such as dicumyl- 30 : ; ~ peroxide, the unreacted peroxide in the pre-treated fiber acts as a means for generating free radicals on the the polymer, thus the polyethylene and cellulose are linked togP~h~r by means of maleic anhydride forming a bridge - between the normally ~_ _ 'ihil~ cellulose and polyethylene.
The pre-treating step can usually be performed by using co..v~i.tional masticating ~ li, ', for example, rubber mills, Brabender mixers, Banbury - ~ mixers, or twin screw continll~u~ mixer extruder. The BrAhPn~r mixer is particularly effective for this purpose in the laboratory. The materials, fiber, bonding agent, activator, polyethylene can all be charged ;nitiAlly.
The order of addition of materials to the mixer is not critical. The temperature of the mixing should be s~lff;~i~n~ly high to melt the bonding -~ agent and polyethylene to pl`O~uCe a 1 _ mixture with the fiber.
Usually, about 160-C, is s~lff;~;~nt to obtain the treated fibers in the form of clusters lightly held t~J~ . m e time of mixing will usually ni ~e~
and will depend upon number of factors, such as the type of mixer, the ; 60 proportion~ of the ingredients, and the size and t~ , d~Ule of the batch.

.~. ", ~
. i :
.: ~
.' .~ ' ~, ` ':.
.
~: .

1 33~7 The proportions of the ingredients will usually be dictated by the ~- properties descriked in the tre~ted fibers. Generally, the ratio of fibers to the other ingredients will be as high as possible in order to r -xi i7e ~ pro~ n of the treated fibers. The amLunt of polymer u~el will be at least - ~lffi~j~n~ to prevent fiber-to-fiber interactions, usually at least 5 parts of ` polyethylene by weight per 100 parts by weight of cellulose fibers. The ; ~ preferred level of bonding agent in the comeosite of invention is from 0.5 to 10 parts per 100 parts of c~ o~ fiber by weight. The amount of activator ~ , . i8 usually 0 to 1 parts by weight of rJPlllllose fiber. In most in~t~n~, it - ~ will be re convenient to include all of the bonding agent in the treat~
fibers, since no further additions of this ingredient need be added in making .,, the final ~ _ te. Since the tl. step~ coats the surface of fibers to certain extent, the polymer present in the coating will be in a po~iti~n to be bonded to the fibers. It appears also, that some ~iti~n~l bonding of the fiber to polyethylene is achieved during the ~ te f- ion~

The mixing of treated fibers and polymer to fonm a ~ _ t~ is usually . .
performed in an internal mixer, or on a roll mill. A Miti~n~l ingredients, ~ ~ such as fillers, color~n~ and st~hili7~rs can be added at this point. This is - ~
followed by compression molding to plo~u~e a desired article. Injection -` molding techniques can also be used for the fabrication of different - ~ articles.

;' ~ E5~PMP[E 1 ~ -~ In order to compare the effects of various hon~ing agent on - - ~ 50 properties of the 1 _- t~4, a series of qreci - were prepared cont~ining the bonding agent of invention as well as other aromatic anhydride and control.
- ,;,.~ ~
~- Maleic anhydrideO(Aldrich) was u~ed as the preferred bonding agent.
C ~
~:', 11 o C ~ _ / 5 ";':
~.

1 3329~7 . ~, Also inr~ forOconparison was a phthAli~ anhydride (Aldrich) ~ Icl~

e activato~ ~sed was dicunyl per xide.
c~ O--o--c~
,. 10 ~

- m e samples had the following formulation:

,. . ; .
, ; TABLE I

Material Percentage by weight :.' ' ~
t~ ir~l pulp ~CTMP) of aspen 0 to 40 , Polyethylene 100 to 60 ,, ~
Maleic anhydride 0 to 4 Dicumyl peroxide 0 to 1 ..::
. .
Mixing was done in a Rr~ roll mill at 165-C. In each case the sample - ~-- was remixed a 'n' of five times to produce a better dispersion of fiber in ` ~ the polymer matrix. The above mixture was ground to mesh ~ize 20 and , , e~Yion molded at 160-C ~ ule 2.7 MPa) for 10 minutes. Dog bone shaped - tensile ~peri ~ were nbt~in~d after cooling with the p~e~ure intAin~
- '- during the process.
-- Tensile tests were done using an Instron Model 4201 at 23-C and 50% RH. The ~; ~ cross-head speed was 5 mm/min. Tensile modulus was measured at 0.1%
~ nn9A~-inn. Tensile properties were ~ at peak load, break and at proof - -~ stress point (the point where the non ~ro~ortional strain deviates by a pre~t~ inPd amount) The results were Allt~ ' iCAl ly calculated by HP86B
.

r using the Instron 2412005 General Tensile Test Program.
. . 60 ,. ,.,. ~ .
", ~ .
. . . .
.:. ~
, ...
.. . . ..
:~- :
, ..... . ~ . ;, . . . .. . . ..

:

1 3~2q87 TAeLE II

FiberTensile strength El~n~ti~n Tensile modulus - (MPa) ~%) ~Gæa) ~% wt.) 10 20 30 40 10 20 30 40 10 20 30 40 a 26.0 9.9 1.06 HDPE
b 18.4 3.1 -~ HDPE+ a26.5 24.7 24.1 21.3 6.1 4.8 3.22.0 1.31 1.42 1.58 1.83 CTMP
aspen b19.7 19.4 18.7 18.5 2.4 2.3 2.32.1 - - - -, ~ 20 HDPE+ a29.7 30.1 35.6 32.1 8.7 8.4 6.85.6 1.27 1.37 1.53 1.76 - -~ treated ~ ~ fiberl b20.5 21.3 27.5 26.3 2.6 2.5 2.7 2.6 . , HDPE+ a29.6 29.3 31.3 30.1 9.4 9.1 7.8 6.0 1.28 1.34 1.49 1.70 ~ treated -~ fiber2 b20.3 20.7 22.2 23.1 2.6 2.5 2.6 2.6 ' a - Maximum load b - Proof stress point Treated fiberl: CTMP aspen ~100 parts)+Maleic anhydride ~2.0 parts)+
polyethylene ~5.0 parts)+dicumyl peroxide ~1.0 parts) Treated fiber2: CTMP aspen ~100 parts)+Maleic anhydride (4.0 parts)' polyethylene (5.0 parts~+dicumyl peroxide (1.0 parts) The results in Table II show an i ~ in tensile properties when treated fibers were used in the ~ _ te~. Tensile strength, at 30.0% filler level, increased from 24.1 MPa of untreated fiber ~ _~ t~Q to 35.6 MPa in the case of treated fiberl ~ _ ~ t~. Higher el~n~tion values were ob~L~d, with the increase in filler conr~ntration, in treated fiber I _ - t~Q than - untreated fiber ~ _ t~Q. Tensile modulus i~ a~ed steadily with the filler iti~n and was not ~uch ~fffrted by fiber ~L~_' ' .

, ~ 332987 EX~MPLE II

-The c~ tes were prepared and evaluated as in Exa~ple I, but in this case wood flour (aspen) was used instead of CTMP aspen pulp. The comçarison of - tensile properties with untreated fibers, plæ~G ~ in Table III, in~in~t~
that i ,_~v~ tensile properties result when treated fibers were used. Tensile strength increased steadily with the increase in filler con~ntration in treated fiber co~posites. At 40.0% filler level in treated fiberl ~ 8, tensile strength increased to 34.7 MPa comeared to 19.6 MPa of untreated fiber composites. At proof stress point, higher tensile strength values were observed in treated fiber composites. The increase in maleic anhydride ~nn~n~ration in treated fibers do not appear to give further i ,_~. ~.

TABLE III

Fiber Tensile strength ~lon~ti~n Tensile modulus ~MPa) ~%) ~GPa) ~% wt.) 10 20 30 40 10 20 3040 10 20 30 40 a 26.0 9.9 1.06 - HDPE
; b 18.4 3.1 ;
,:.

HDPE+ a 30.0 32.5 33.1 34.7 7.9 7.7 6.9 6.2 1.35 1.48 1.62 1.80 treated fiberl b 22.9 23.8 24.4 26.6 2.7 2.7 2.6 2.5 HDPE+ a 28.4 31.0 33.4 34.5 8.1 7.2 6.2 5.9 1.15 1.46 1.59 1.84 treated fiber2 b 22.2 23.3 24.6 25.5 3.0 2.8 2.6 2.6 . ' - ~ 60 . .

; .~
-- - 1 332~87 ~: .
a - Maximum load b - Proof stress point Treated fiberl: Wbod flour (100 pHrts)+Maleic anhydride ~2.0 parts)+
polyethylene ~5.0 parts)+dicumyl peroxide ~1.0 parts) Treated fiber2: Wbod flour ~100 parts)+Maleic anhydride ~4.0 parts)+
~, '10 polyethylene ~5.0 parts)+dicumyl peroxide ~1.0 parts) ~ ; The treated fiber ~ ~ sites also p~udueed better elongation values at ,~ higher filler addition when compared to untreated fiber ~ t~. A sharp - - ~ increase in modulus was oL~e.v~ with the A~;ti~n of filler. The increase in ; 20 m~dulus was not much Aff~r~ed by fiber ~r~ ' '.
: ' ~
,~.
E52YD~LE III
-:
~ ~ Using the process of Example I, ~ , t~ are prepared as shown in Table IV. The samples had the fol l~ting fo l~tinn:
, ~

." ;,., . -" , Ingredients Percentage by weight . . . ~
t~ ~1 pulp ~CTMP) of aspen 0 to 40 Polyethylene 100 to 60 Phthalic anhydride 0 to 4 - Dicumyl peroxide 0 to 1 .~ :
- ~ 50 ~- Treated fibers were prepared as in Example I and the c ,- ~4 were made under the same e~ 1 con~iti~n~ as ~ ~P~ in Exa~ple I. The results are given in Table V. The tensile test results in~i~te that treated fibers ~; give very good ~h~inn. Even at a relatively low level of bonding agent ~2.0 parts of p~th~lir anhydride per 100 parts of fiber) a ~ignifir~nt increase in `', ;', ~

..'' :
:. . .

~, 1 3 ~2987 tensile strength was observed. The treated fibers also have better modulus values compared to l]nfillf~ HDPE.

TAeLE V

Fiber Tensile ~c~Elon~tjon Tensile modulus (MPa) t%) ~GPa) ~ ~ (% wt.) 10 20 30 40 10 20 30 40 10 20 30 40 -;
a 26.0 9.9 1.06 HDPE
b 18.4 3.1 HDPE+ a 26.5 24.7 24.1 21.3 6.1 4.8 3.2 2.0 1.31 1.42 1.58 1.83 CTMP
aspen b 19.7 19.4 18.7 18.5 2.4 2.3 2.3 2.1 HDPE+ a 28.4 30.9 32.1 34.9 8.7 8.1 7.1 6.0 1.22 1.43 1.58 1.87 treated - ~; fiber3 b 19.9 20.9 23.4 26.2 2.8 2.5 2.5 2.4 ~ 30 .. . ~
.~ HDPE+ a 28.0 30.3 33.0 34.3 8.9 6.7 6.1 5.5 1.21 1.37 1.53 1.79 treated : . , fiber4 b 19.8 22.9 24.8 26.0 2.8 2.7 2.6 2.7 - - - -.-;:'.. :
i C~`~ a - Maximum load b - Proof stress point ~ 40 :.
Treated fiber3: CTMP aspen (100 parts)+Phthalic anhydride (2.0 parts)+
polyethylene (5.0 parts)+dicumyl peroxide (1.0 parts) Treated fiber4: CTMP aspen (100 parts)~Phthalic anhydride (4.0 parts)+
polyethylene (S.O parts)+dicumyl peroxide (1.0 parts) . .
- EX~MPLE IV
'.~,'-~; The ~ _ tes were prepared and evaluated as in Example III, but in this i case wood flour (aspen) was used instead of CTMP aspen pulp. It is evident ` ~ from the results of tensile properties (Table VI) that addition of treated ;: ` 10 -~ ~
' -, r ::

: I' ~. : ' :

" - 1 3 ~2987 fibers in polyethylene causes a significAnt i _ ~ ' in bonding between the - fiber and polymer matrix. The highest tensile strength is recorded with -~ treated fiber3, where the strength in~ledse~ from 19.6 MPa of untreated fiber composites to 37.2 MPa at 40.0 % filler ~OI~ 1rd~iOn in the ~ _ t~. A
significant i __~ ' in strength was also ~e~v~a at Proof stress point.
` 10 Higher addition of b~n~ing agent (4.0 parts of phtahlic anhydride per 100 - part~ of wood fiber~ do not produce much increase in tensile strength.

- TAeLE Vl Fiber Tensile ~L~CI~I Elon~t;nn Tensile (MPa) (%) (GPa) (% wt.) 10 20 30 40 10 20 30 40 10 20 30 40 a 26.0 9.9 1.06 HDPE
b 18.4 3.1 HDPE+ a25.8 23.5 21.2 19.6 8.3 5.8 3.5 2.3 1.27 1.47 1.63 1.82 sawdust aspen b18.5 19.2 19.6 19.4 2.4 2.3 2.2 2.1 HDPE+ a29.4 30.0 32.2 37.2 8.5 6.7 5.9 5.6 1.22 1.40 1.65 1.81 treated fiber3 b 21.0 23.1 25.4 27.8 2.6 2.7 2.6 2.6 HDPE+ a28.2 30.4 32.7 34.9 8.7 7.5 7.1 6.2 1.16 1.36 1.59 1.85 treated fiber4 b 20.3 22.4 25.0 26.2 2.8 2.7 2.8 2.7 : ~
a - Maximum load b - Proof stress point ` 50 -- Treated fiber3: Wood flour (100 pdrts)+Phthalic anhydride (2.0 pdrts)+
polyethylene (5.0 pdrts)+dic~nnyl peroxide (1.0 parts) ; Treated fiber4: Wood flour (100 parts)+Phthalic anhydride t4.0 parts)+
- polyethylene (5.0 pdrts)+dicumyl peroxide tl.0 pQrts) ~A

''~ ' '` ` ~
' `' ~` ' ' ~ '~ ' 1 3 ~29~7 "
EX~DPLE V
.

The Izod-impact test results in Table VII inAir~te that HDPE filled with treated fibers produce higher impact X~elly~l values, even at relatively low level of bonding agent ~2.0 parts per 100 parts of cellulose fiber) compared to untreated fiber ~ , _ites. Also the better impact strength of treated fiber2 and treated fiber3 ~ than untreated fiber . t~ in~ir~te go~d a~h~ n of bonded fibers with the polymer matrix.
~. :

D~LE VII

C ~ 'te Izod-Impact strength (KJ/mZ) - (un-notched) ~Eiber wt. %) 10 20 30 40 HDPE 2907+
- ~ sawdust untreated 26.9 24.9 23.8 22.2 -~ 30 HDPE 2907+
~ treated fiberl29.8 28.2 26.4 24.9 -j~- HDPE 2907+
; ~ treated fiber231.5 28.6 26.9 23.6 ~' HDPE 2907+
r treated fiber331.2 29.6 26.1 24.4 HDPE 2907~
treated fiber429.1 28.0 23.7 22.8 ~unfilled) 30.8 . ~

~ Although the foregoing invention has been described in some detail by way .~, ; - 50of examples, it is not limited thereto. ~h~ng~ and modifications of the . .
examples of the invention herein chosen for purpose of disclosures can be made which do not ron~titl~ departures from the sprit and scope of the invention.

: !
~- 12 :
`~

Claims

1. A composite consisting essentially of 10-40 weight percent discontinuous cellulose fibers and 90-60 weight percent of high density polyethylene, prepared by a two-step process in which the cellulose fibers are pre-treated with ethylene polymer, carboxylic acid anhydride, and a free radical initiator in a first-step, and in a second-step the pretreated fibers are compounded with high density polyethylene.

2. The composite of claim 1, wherein the anhydride is maleic anhydride or phthalic anhydride.

3. The composite of claim 1, wherein the free radical initiator is a dicumyl peroxide in an amount of 0.1 to 1 weight percent of cellulose fibers.

4. The composite of claim 1, wherein said cellulose fibers, said ethylene polymer, said anhydride and said free radical initiator are mechanically mixed together under conditions of high shear.

5. The composite of claim 1, wherein the cellulose fibers are chemithermo-mechanical wood pulp of hardwood and/or softwood or wood flour derived from hardwood and/or softwood.

6. The composite of claim 1, wherein the cellulose fibers have an average aspect ratio of from 2 to 200.

7. Articles made from composites of claim 1 by compression or injection molding.