Polymer Degradation and Stability 88 (2005) 435e443 www.elsevier.

com/locate/polydegstab

Microstructure changes of extruded ultra high molecular weight polyethylene after gamma irradiation and shelf-aging
lvarezb, E. Go mez-Barrenac, J.A. Pue rtolasa,) F.J. Medela, F. Garc a-A
Department of Materials Science and Technology, Centro Politecnico Superior de Ingenieros, Universidad de Zaragoza, Calle Mara de Luna 3, 50018 Zaragoza, Spain b Department of Orthopaedic Surgery, Hospital Lozano Blesa, Zaragoza, Spain c n Jimenez Daz, Universidad Auto noma de Madrid, Madrid, Spain Department of Orthopaedic Surgery, Hospital Fundacio Received 16 September 2004; accepted 27 November 2004 Available online 2 February 2005
a

Abstract Ultra high molecular weight polyethylene (UHMWPE) components in joint prostheses may undergo oxidative degradation in the long term. This leads to material embrittlement, associated with a subsurface defect named white band that favours surface delamination and premature implant failure. The polymer microstructure, which aects the mechanical properties of the material, is altered by oxidative degradation. In this study, the microstructure of bar-extruded UHMWPE (GUR 1050) was compared before irradiation, one week after 25 kGy gamma irradiation in air (GUR 1050), and 7 years after commercial (25e40 kGy) gamma irradiation in air and shelf-aging (GUR 415). This was performed using transmission electron microscopy (TEM) and dierential scanning calorimetry (DSC) which help to characterize the changes in material microstructure and crystallinity during degradation. Quantitative (lamellar thickness), semiquantitative (lamellar twisting, orientation, lamellar density), and qualitative (lamellar border, stacking) parameters were recorded in the surface, subsurface and inner material. There was an evolution in the degradation in the bar-extruded UHMWPE microstructure during irradiation and shelf-aging. Changes in the in-depth microstructure indicate the progression of an oxidation front. The stacking or high lamellar concentration in the subsurface in shelf-aged samples was coherent with higher crystallinity and with a degradation process controlled by oxygen diusion and free radical distribution after gamma irradiation. 2005 Elsevier Ltd. All rights reserved.
Keywords: UHMWPE; Microstructure; TEM; DSC; Degradation

1. Introduction In the past decades, ultra high molecular weight polyethylene (UHMWPE) has been the preferred interposition material for joint prostheses. Its microstructure, based on the extremely high molecular weight

) Corresponding author. Tel.: C 34 976762521; fax: C 34 976761957. rtolas). E-mail address: japr@unizar.es (J.A. Pue 0141-3910/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymdegradstab.2004.11.015

(4e6 ! 106 g/mol), confers optimal mechanical properties such as high wear and fatigue resistance and high fracture toughness. Those properties, along with a low friction coecient and high biocompatibility, make UHMWPE the perfect candidate to articulate against metallic alloys or ceramics [1]. However, wear and degradation can limit the in vivo service and duration of the reconstructed joint system [1,2]. The UHMWPE microstructure includes an amorphous region, with entangled long molecular chains, and a crystalline region, with thin lamellar crystals or

436

F.J. Medel et al. / Polymer Degradation and Stability 88 (2005) 435e443

lamellae. Both regions are separated by a narrow border which is considered a part of the lamella. Tie molecules interconnect the regions by crossing the amorphous region from one lamella to another. The tie molecules and entanglements in the amorphous phase are what help to provide the excellent mechanical properties of this material [3,4]. The material has usually been irradiated for sterilization (between 25 and 40 kGy), but this has a negative eect on the entanglement density of long molecular chains, as well as the concentration of tie molecules [5]. Molecular chains undergo scission and free radicals appear in the amorphous region and on the crystal surfaces. The free radicals can take part in the crosslinking of the material or oxidative degradation in the presence of oxygen [5,6]. In the amorphous region, free radicals disappear by recombination or by reacting with oxygen since the molecular chain is more mobile and oxygen easily diuses in this phase. However, in the crystalline regions of UHMWPE, radicals are long lasting [7]. They diuse towards the amorphous region at a slow rate and participate in dierent stages of oxidation. Finally, UHMWPE also contains degradation products such as ketones, acid groups, and esters, as shown by FTIR [8]. Natural aging of UHMWPE on the shelf and in vivo cause maximum subsurface degradation (between 1 and 2 mm under the surface) [9]. The oxidation causes embrittlement of the material and loss of mechanical properties. As a result, research in orthopaedics has focused on the physico-chemical characterization of the degradation. Many studies have been performed using density measurements, calorimetry, and FTIR [10e22]. However, it is still unclear how aging changes the microstructure and how it aects degradation proles. Advances in this area could help to explain how UHMWPE loses its mechanical properties. We analysed the microstructural variations due to irradiation and aging in bar-extruded UHMWPE. Dierential scanning calorimetry (DSC) and transmission electron microscopy (TEM) studies were performed to determine depth-dependent changes in microstructural characteristics of non-irradiated material, freshly gamma irradiated in air, and shelf-aged gamma irradiated in air commercial components. We propose a model of microstructural changes based on our ndings and previous reports in the literature.

bar-extruded Hoechst GUR 415 or 4150 according to the new nomenclature [2]. The experimental samples made from the resins were classied into three dierent groups. The rst group was from three pucks (10 mm thick), machined from a bar of extruded GUR 1050. Three sections were cut from the centre of each puck with a microtome, at 0.1e0.3 mm, 1 mm, and 3 mm from the surface. Samples were prepared for DSC and TEM tests from every 200 mm thick section. The second group consisted of three pucks machined from extrusion bars of GUR 1050 UHMWPE, and then gamma irradiated in air at conventional sterilization dosage (25 KGy; Aragogamma, Spain). Again, three sections were extracted from the centre of each puck (0.1e0.3 mm, 1 mm, and 3 mm from the puck surface). Those sections were used for DSC analysis and TEM, and tested one week after the irradiation process. The third group corresponded to naturally shelf-aged Hoechst GUR 415 material, obtained from three tibial components of knee prostheses that had never been implanted (Natural Knee system, manufactured by Intermedics Orthopedics, Austin, TX, in 1991). They were discarded for clinical use after being stored for 7 years on the shelf after sterilization. All prostheses had a wide subsurface white band in sections. On the back of the component we cut 200 mm sections at dierent depths to obtain a surface group (0.1e0.3 mm deep), a white band or subsurface group (1 mm deep), and an inner material (below white band, 3 mm deep). The samples of each section were divided in two parts, one for DSC measurements and another for TEM. 2.2. Transmission electron microscopy Subsequent sample preparation for TEM included dierent phases. First, 200 mm sections of UHMWPE were stained with 99% chlorosulphonic acid at 60  C for 5 h, which is thought to stabilize the amorphous regions. Samples were washed with acetone (at 0  C) and rinsed with distilled water. Stained samples were dried at 60  C for 1 h and later embedded in epoxy and cured at 60  C for 2 days. Ultrathin sections (w60 nm thick) were cut with a diamond knife and collected on carbon grids. Then the sections were post-stained with uranyl acetate in 1% methanol for 4 min. A Jeol 100CX TEM (operating at 100 kV) was used to produce micrographs at 20,000! and 60,000! magnications. The best 5 TEM preparation samples were chosen from each group and analysed using the Digital Micrograph 3.3.1 software package (Gatan Inc., Pleasanton, CA, USA). Morphological variables were selected by image analysis and evaluated quantitatively or semiquantitatively. These variables assessed lamellar properties (thickness, boundaries, kinking) and interlamellar

2. Materials and methods 2.1. Materials The polyethylene resins were bar-extruded Ticona GUR 1050 processed by a commercial converter (Perplas Medical Ltd.), and components manufactured from

F.J. Medel et al. / Polymer Degradation and Stability 88 (2005) 435e443

437

features (distance or spacing, chaotic distribution, concentration, stacking). Some were ranged on a semiquantitative scale of one dot (minimum) to three dots (maximum). Lamellar thickness and interlamellar distance in stacked lamellae were measured at least ve times per 60,000! images using the software. 2.3. Dierential scanning calorimetry A PerkineElmer DSC-7 was used for the calorimetric analysis. Samples were heated from 25 to 180  C at a rate of 10  C/min. Their mass ranged from 3e4 mg. Fusion heat was obtained by integrating the area under the endothermic peak from 80  C to 160  C. Melting temperature was registered as the maximum of the endothermic melting peak. In addition to the direct parameters, sample crystallinity was calculated by comparing the fusion heat of the sample to the fusion heat for a fully crystalline polyethylene material (290 J/g) [17]. An estimation of the lamellar thickness, Lc, was obtained from DSC data using the ThomsoneGibbs equation: Tm ZTm0 1 2s=Lc rc DH0 1 m where Tm is the melting point of the polymer, Tm0 is the equilibrium melting point of a perfect crystalline polyethylene, s the specic surface energy, rc the crystallinity phase density, and DH0 m the enthalpy of melting of a perfect crystalline polyethylene. Lamellar thickness was estimated by considering DH0 m Z 290 J/g, Tm0 Z 145.7  C, rc Z 1.005 g/cm3, and taking a constant value for s Z 93 ! 107 J/cm2 [18].

3. Results 3.1. Calorimetric analysis The thermogram of the ram-extruded material was fairly typical, with a strong endothermic peak (Fig. 1a). As seen in Table 1, non-irradiated specimens had a melting temperature of 135.5 G 1.1  C, an enthalpy content of 140 G 12 J/g, a crystallinity percentage of 49 G 4%, and an estimated lamellar thickness (using the ThomsoneGibbs equation) of 27 G 3 nm. In the rst group, there were no signicant dierences between specimens from dierent depths. Gamma-irradiated, one week shelf-aged UHMWPE samples had a higher melting temperature (137.3 G 0.4  C) and crystallinity (53 G 1%; Table 1). Lamellar thickness was also higher (32 G 2 nm) than virgin UHMWPE. As in non-irradiated samples, there were no signicant dierences with depth in the DSC results. A typical thermogram of irradiated UHMWPE is shown in Fig. 1a, along with the thermogram of virgin UHMWPE.

Fig. 1. (a) Thermograms of virgin, non-irradiated, UHMWPE, and gamma-irradiated (25 kGy) plus one week shelf-aging UHMWPE. (b) Thermograms of the three regions (surface, subsurface or white band, and inner or below white band) on gamma-irradiated (25 kGy) plus 7 years of shelf-aging prostheses.

Finally, the thermal values of samples from aged tibial components varied with depth. Fig. 1b shows the typical thermograms of the three regions. Samples from the subsurface (where the white band was present), had the highest crystallinity (71 G 3%), and an increased melting temperature (139.8 G 1.2  C), as well as higher lamellar thickness (48 G 9 nm). In contrast, surface specimens had the lowest enthalpy content, melting temperature (137.1 G 0.1  C), and lamellar thickness (32 G 1 nm). Inner samples, which were already near the wide white bands of the prostheses, had an enthalpy of 190 G 1 J/g, crystallinity percentage of 65 G 1, melting temperature of 140.1 G 1.0  C and lamellar thickness of 50 G 9 nm. The direct parameters obtained from the

438

F.J. Medel et al. / Polymer Degradation and Stability 88 (2005) 435e443

Table 1 Enthalpic contents, % crystallinity, melting temperature data and lamellar thickness estimations based on the ThomsoneGibbs equation of nonirradiated (virgin), gamma-irradiated and one week aged, and gamma-irradiated and 7 years aged extruded UHMWPE UHMWPE Enthalpic contents (J/g) 140 G 12 154 G 4 169 G 5 206 G 8 190 G 1 % Crystallinity (ref. 290 J/g) 49 G 4 53 G 1 57 G 2 71 G 3 65 G 1 Melting temperature (  C) 135.5 G 1.1 137.3 G 0.4 137.1 G 0.1 139.8 G 1.1 140.1 G 1.0 Lamellar thickness (nm)a 27 G 3 32 G 2 32 G 1 48 G 9 50 G 9

Virgin Gamma C 1 week Gamma C 7 years Surface Subsurface Inner region

a

The values for s and rc are mentioned in the text.

DSC experiments and the crystallinity percentages, as well as the estimated Lc values for these three material groups are also presented in Table 1. 3.2. Transmission electron microscopy The structure of extruded, non-irradiated GUR 1050 UHMWPE was a semicrystalline polymer. The crystalline region included lamellae that appeared as narrow white ribbons that were more or less curved, with sharp boundaries (Fig. 2a). Zones among the lamellae denoted the amorphous region. The boundary between lamellae and amorphous region was so thin that it was hard to perceive. Lamellar thickness was between 25e30 nm. All these key variables are shown in Table 2. The lamellae in the TEM images of gamma-irradiated, one week shelf-aged UHMWPE (Fig. 2b) were less contrasted than non-irradiated polymer (Fig. 2a). Lamellar thickness averaged 35 nm, which is clearly thicker than virgin UHMWPE. A slight decrease in the lamellar concentration could be deduced from the 20,000! images. The sharpness of the lamellar boundaries also decreased. There were no signicant changes in the other variables. There were marked dierences between the irradiated, 7-year shelf-aged samples and the non-irradiated material. As shown in Fig. 3, the lamellae were thicker in the irradiated, 7-year shelf-aged group than in the virgin specimens, and the boundaries were broader and clearly visible. On the other hand, there were some dierences in the depth of the shelf-aged samples. In the surface specimens, lamellae were kinked and their distribution was chaotic. In the white band specimens, lamellae appeared straightened and formed stacks in many areas. In the latter, interlamellar distances ranged between 25e30 nm (measured between adjacent lamellae). Lamellar concentration or density in the white band (Fig. 3b) was much higher than at the surface. Finally, some features of the inner samples (Fig. 3c) were similar to the subsurface, but with no signicant stacking. Table 2 includes the results of the quantitative and qualitative microstructural parameters for all the materials.

4. Discussion Many studies have analysed the changes in UHMWPE after irradiation and natural aging [10e22,24e26].

Fig. 2. TEM images (60,000!) of virgin ram-extruded UHMWPE (a), and gamma-irradiated (25 kGy) UHMWPE after one week aging (b).

F.J. Medel et al. / Polymer Degradation and Stability 88 (2005) 435e443 Table 2 The results of a semiquantitative analysis of lamellar variables Key features Virgin Gamma Gamma C 7 years UHMWPE C 1 Surface Subsurface Inner week (white band) region 25e30 No  30e35    No 32e38 Yes  35e40    No 32e38 Yes  45e50    No 30e35 Yes  25e30    Yes 32e38 Yes  35e40    No

439

Thickness (nm) Border/frontier Twisting Spacing (nm) Chaotic appearance Orientation Density of lamellae Stacking

In the qualitative variables the scale ranges from one dot () minimum, to three dots () maximum.

reportedly around 50% [27], depending on the processing route, resin, molecular weight, etc. There is no detectable oxidation and the density is about 0.93 g/cm3 [1,14]. According to the DSC thermograms, crystallinity was 49 G 4% and melting temperature was 135.5 G 1.1  C, as found in previous studies [1,2]. Kurtz and Schmitt point out that the crystalline phase of UHMWPE consists of folded long molecules, added to thin orthorhombic crystals (10e50 nm thick, and 10e50 mm long). Lin and Argon [33] also report similar features for polyethylene single crystals obtained from a solution, with an orthorhombic structure (5e25 nm thick, and 1e50 mm long). Our estimated Lc, using the melting temperature data and the ThomsoneGibbs equation was 27 G 3 nm, in agreement with measured Lc values on TEM images of virgin UHMWPE (compare Tables 1 and 2). 4.2. Second stage (low degradation) At low doses, gamma or electron beam irradiation causes the radiolytic scission of molecular chains by cleavage of CeC and CeH bonds, producing dierent free radicals. Since irradiation aects both crystalline and amorphous regions, the mobility of the free radicals is dierent. Depending on their location and mobility, they can undergo recombination (leading to crosslinking, branching, and the formation of double bonds), stabilize, or react with incoming oxygen [5]. Irradiating in the absence of oxygen will increase cross-linking and decrease crystallinity [10,18,26]. It also decreases above 100 Mrad since the crystals are seriously damaged [24]. However, at low and medium doses in air, crystallinity increases after gamma [10,25,26] or electron beam [16,22,24] irradiation. In those cases, the degradation process (via oxygen) is more probable, favouring oxidative chain scission as oxygen molecules react with free radicals. Recrystallization processes are favoured by oxidative reactions, and crystallinity will increase. In the proposed microstructure for this stage, the crystallinity and lamellar thickness are higher than in the virgin samples. In addition, lamellar boundaries are less sharp, darker and wider in TEM images (see Fig. 4b). Our DSC and TEM results on gamma-irradiated UHMWPE conrm the increase in crystallinity. One week after irradiation, the estimation for UHMWPE crystallinity, melting temperature and lamellar thickness increased by 8%, 2  C, and 5 nm, respectively (Table 1). According to the TEM images, the lamellae were thicker (32e38 nm), than virgin UHMWPE. Although our thermodynamic results agree with other reports, there was an apparent contradiction in the lamellar size. Bhateja et al. [16] and Premnath et al. [22] found that in the rst hours after beta irradiation, UHMWPE crystallinity and melting temperature increased but there was no change in lamellar thickness, as

However, although variation in-depth proles in terms of physico-chemical properties (density, oxidation index, crystallinity, etc.) are well documented, little is known about microstructural depth changes and its relationship with physico-chemical changes. On the other hand, DSC thermograms and TEM images suggest an in-depth material distribution in the 7-year shelf-aged UHMWPE but not in the freshly irradiated UHMWPE or virgin material. In view of these ndings, a global model of structural evolution after gamma irradiation in air can be proposed to discuss our results in relation with the literature. This model would rely on three consecutive stages.

4.1. First stage (virgin UHMWPE) This stage relates to virgin UHMWPE. Its microstructure can be considered a semicrystalline polymer with crystalline and amorphous phases. The crystalline phase consists of folded long molecules added to thin orthorhombic crystals, named lamellae. Those lamellae are randomly oriented and separated by highly entangled polymer molecules, which form the amorphous phase. Some tie molecules connect dierent lamellae through the amorphous phase and improve the mechanical properties of UHMWPE quite remarkably [4,25,28]. These well-known features are represented in Fig. 4a. In the TEM images, typical white ribbons are easily identied as randomly oriented lamellae, immersed within a dark region corresponding to the amorphous phase. Lamellar boundaries appeared to be sharp, probably because there was only primary crystallization during the processing of the extruded bars, and fold surfaces of the lamellae lie almost parallel to the electron beam of the TEM [29]. Several studies support our results [4,23,25,29e32]. Regarding the physico-chemical properties at this stage, crystallinity values are

440

F.J. Medel et al. / Polymer Degradation and Stability 88 (2005) 435e443

detected by SAXS measurements. In our samples, lamellar thickness increased one week after gamma irradiation in air. These ndings can be explained by the Thomsone Gibbs equation (Eq. (1)) and the relation provided by Bershtein and Egorov [34] for the surface enthalpy of the lamellae qi, (Eq. (2)), which relates this thermodynamic property to lamellar thickness and total melting enthalpy, DHm: q i Z DH 0 2 m DHm Lc =2:5 where DH0 m is again the melting enthalpy of a perfect crystalline polyethylene. The surface enthalpy, s, and qi are related by the expression s Z qi Tmsi, where Tm is the melting point of the polymer and si the entropy of the polymer [34]. In the rst hours after irradiation, the increase of DHm in the virgin sample does not correspond to an increase in the degree of crystallinity, since the lamellar size remains constant. According to Eq. (2), the increase is associated with a decrease in surface enthalpy, qi. On the other hand, the positive shift of the melting temperature can only be related to a decrease in the free energy surface, s. Both changes are attributed to the radiolytic scission of tie molecules, which subsequently rearrange into the crystals improving the surface of the original lamellae. However, the increase in melting temperature, enthalpy, and lamellar size could be related to the beginning of a secondary crystallization at the surface. An indication of this process is the change in the lamellar boundary. As seen in the TEM images, the boundaries are less sharp. After the secondary crystallization, the surface folds of the lamellae are less parallel to the TEM electron beam than in the virgin samples [13]. The low degradation level associated with this stage can be explained within the general framework of UHMWPE degradation, which is controlled by the diusion of oxygen and free radicals after irradiation. The oxidation process requires that oxygen and radicals generated during irradiation coincide in space and time. The theoretical simulation of oxygen diusion into UHMWPE, along with the time evolution of the spatial radical concentration [35] implies that the surface degradation of up to 0.5e1 mm deep remains basically constant (not time dependent). However, the subsurface is strongly degraded with physical and chemical changes after more than 5 years post-irradiation [13,19,23,35]. This is why the thermodynamic parameters and the microscopic features in the 7-year old samples were similar to the ones after one week of natural aging.
Fig. 3. TEM image (60,000!) of gamma-irradiated (25 kGy) plus 7 years of shelf-aging prostheses. Surface zone (a), subsurface or strong degradation zone corresponding to the white band (b), and inner or medium degradation zone below the white band (c).

F.J. Medel et al. / Polymer Degradation and Stability 88 (2005) 435e443

441

Fig. 4. Schematic drawings of UHMWPE microstructure evolution after irradiation and shelf-aging: (a) rst stage in virgin UHMWPE, (b) freshly irradiated UHMWPE, (c) second stage in the surface region after 7 years shelf-aging UHMWPE, and (d) third stage in the white band area.

Both situations can be considered as two images corresponding to a stage of low degradation. There was an apparent increase in lamellar tortuosity at the surface after long shelf-aging (see Fig. 4c), which is mainly due to oxygen accumulation at the lamellar boundaries [32]. 4.3. Third stage (strong degradation) The microstructure in the third stage is linked to the oxidative degradation process that generates the white band defect. This introduces important changes in material properties such as density, molecular weight, crystallinity degree or toughness. As widely accepted, the changes only occur after irradiation in air followed by an incubation period of more than 4e5 years. Our microscopic model for this stage includes lamellar thickening. The concentration is high with predominant stacking of lamellae that almost lie straight and parallel (see Fig. 4d). There are also small crystallites within the amorphous region. The model is based on the results from DSC and TEM from highly degraded material (see Fig. 1b and Fig. 3b) and data from the literature. The DSC results in the white band region imply an increase in crystallinity and melting temperature after

7 years of shelf-aging. The shoulder in the DSC thermograms (Fig. 1b) may reect the presence of new crystallites produced by chain scission processes that generate shorter molecular chains. As a result, new crystallization can occur in the amorphous region with crystals that are shorter than the original lamellae. Generalized chain scission also produces an overall decrease in tie molecules, allowing for lamellar rapprochement and reordering [31], which increases mechanical brittleness [21,36] and melting temperature, as shown by DSC (see Fig. 1b). This small crystallite production is supported by previous ndings of a decrease in the interlamellar space at SAXS and associated with the shoulder of the DSC thermograms [22]. Partial recrystallization in the amorphous region has also been proposed [17]. Other studies of crystallinity and melting temperatures conrm an increase in fusion heat after some months of aging, mostly with high radiation dosage [22]. The conventional dosage of 25 kGy produces a mild increase in fusion heat in the early months, which proceeds at a lower rate with aging time [10e11]. Later on, fusion heat continues to increase, almost inadvertently in the rst year [10,31,32], but quite noticeably after 5 years [9e11,21].

442

F.J. Medel et al. / Polymer Degradation and Stability 88 (2005) 435e443 [5] Premnath V, Harris WH, Jasty M, Merril EW. Gamma sterilization of UHMWPE articular implants: an analysis of the oxidation problem. Biomaterials 1996;17(18):174153. [6] Coote CF, Hamilton JV, McGimpsey WG, Thompson RW. Oxidation of gamma-irradiated ultrahigh molecular weight polyethylene. J Appl Polym Sci 2000;77:252542. [7] ONeill P, Birkinshaw C, Leahy JJ, Barklie R. The role of long lived free radicals in the ageing of irradiated UHMWPE. Polym Degrad Stab 1999;63(1):319. [8] Costa L, Luda MP, Trossarelli L, Brach del Prever EM, Crova M, Gallinaro P. Oxidation in orthopaedic UHMWPE sterilized by gamma-irradiation and ethylene oxide. Biomaterials 1998;19(7e9):65968. [9] Currier BH, Currier JH, Collier JP, Mayor MB, Scott RD. Shelf life and in vivo duration. Impacts on performance on tibial bearings. Clin Orthop Rel Res 1997;342:11122. [10] Deng M, Shalaby SW. Long-term g-irradiation eects on ultrahigh molecular weight polyethylene. J Biomed Mater Res 2001;54:42835. [11] Yeom B, Yu YJ, McKellop A, Salovey R. Prole of oxidation in irradiated polyethylene. J Polym Sci: Part A. Polym Chem 1998;3:32939. [12] Yu YJ, Shen FW, Mckellop HA, Salovey R. Hydroperoxide formation in irradiated polyethylene. J Polym Sci: Part A. Polym Chem 1999;37:330916. [13] Daly BM, Yin J. Subsurface oxidation of polyethylene. J Biomed Mater Res 1998;42:5239. [14] Goldman M, Lee M, Gronsky R, Long GG, Pruitt L. Oxidation of UHMWPE characterized by Fourier transform infrared spectrometry. J Biomed Mater Res 1997;37:4350. [15] Kurtz SM, Rimnac CM, Bartel DL. Degradation rate of UHMWPE. J Orthop Res 1997;15:5761. [16] Bhateja SK, Duerst RW, Martens JA, Andrews EH. Radiationinduced enhancement of crystallinity in polymers. J Macromol Sci RMC 1995;4:581659. [17] Buchanan FJ, White JR, Sim B, Downes S. The inuence of gamma irradiation and aging on degradation mechanisms of UHMWPE. J Mater Sci Mater Med 2001;12:2937. [18] Andjelic S, Richard RE. Crystallization behavior of ultrahigh molecular weight polyethylene as a function of in vacuo g-irradiation. Macromolecules 2001;34:896906. [19] Shen FW, McKellop H. Interaction of oxidation and crosslinking in gamma-irradiated ultra-high molecular weight polyethylene. J Biomed Mater Res 2002;61:4309. [20] Jacob RJ, Pienkowski D, Lee KY, Hamilton DM, Schroeder D, Higgins J. Time- and depth-dependent changes in crosslinking and oxidation of shelf-aged polyethylene acetabular liners. J Biomed Mater Res 2001;56(2):16876. [21] Edidin AA, Jewett CW, Kalinowski A, Kwartang K, Kurtz SM. Degradation of mechanical behavior in UHMWPE after natural and accelerated aging. Biomaterials 2000;21:145160. [22] Premnath V, Bellare A, Merril EW, Jasty M, Harris WH. Molecular rearrangements in UHMWPE after irradiation and long-term storage in air. Polymer 1999;40(9):221529. [23] Kurtz SM, Pruitt LA, Crane DJ, Edidin AA. Evolution of morphology in UHMWPE following accelerated aging: the eect of heating rates. J Biomed Mater Res 1999;46:11220. [24] Ikada Y, Nakamura K, Ogata S, Makino K, Tajima K, Endoh N, et al. Characterization of ultrahigh molecular weight polyethylene irradiated with g-rays and electron beams to high doses. J Polym Sci: Part A. Polym Chem 1999;37:15968. [25] Kang PH, Nho YC. The eect of g-irradiation on ultra high molecular weight polyethylene recrystallized under dierent cooling conditions. Rad Phys Chem 2001;60:7987. [26] Shinde A, Salovey R. Irradiation of ultrahigh-molecularweight polyethylene. J Polym Sci: Polym Phys Ed 1985;23: 16819.

Finally, the ndings in the subsurface region (related to high degradation), were basically similar to the deeper region, with some dierences. Whereas the melting transition occurred at the same temperature (140  C), neither the enthalpy content nor the TEM images were similar. According to the micrographs, the lamellae were more tortuous, less oriented, and not clustered. On the other hand, the incipient shoulder in the DSC thermograms and the lower value of crystallinity could indicate the lack of new crystallization, which would have produced the small crystals in the subsurface region. However, both thermal aspects suggest that the inner region is in a medium degradation stage. This behavior is supported by Blanchet and Burroughs [35], who found that the combination of oxygen diusion and radicals concentration explains the presence of carbonyl groups in the inner region (close to the subsurface). Their density is higher than at the surface but lower than at the white band, which corresponds to the highest degradation region. 5. Conclusions Both irradiation and post-irradiative degradation of UHMWPE cause thermal and microstructural changes. There was an increase in crystallinity in UHMWPE just after gamma irradiation in air. As expected, lamellar crystals appeared to thicken soon after gamma irradiation. Lamellar thickness seemed to remain almost constant after irradiation and during shelf-aging, as shown by TEM. As degradation proceeded, crystallinity increased at the subsurface region, whereas the surface regions had relatively low crystallinity contents due to the oxygen diusion front and the specic spatial radical concentration at the surface. Subsurface areas had high crystallinity contents and lamellar stacking. In addition, shoulders in thermograms of subsurface regions denote the crystallization of new small crystals in highly degraded zones. The lower enthalpy content and the absence of stacking at the inner region indicate that this region is immersed in a low degradation stage, compared to the subsurface.

References
[1] Li S, Burnstein AH. Ultra high molecular weight polyethylene. The material and its use in total joint implants. J Bone Joint Surg 1994;76-A(7):108090. [2] Kurtz SM, Muratoglu OK, Evans M, Edidin AA. Advances in the processing, sterilization, and crosslinking of UHMWPE for total joint arthroplasty. Biomaterials 1999;20(18):165988. [3] Farrar DF, Brain AA. The microstructure of ultra-high molecular weight polyethylene used in total joint replacements. Biomaterials 1997;18(24):167785. [4] Kurtz SM, Crane DJ, Pruitt LA. An Atlas of UHMWPE microscopy. www.uhmwpe.org.

F.J. Medel et al. / Polymer Degradation and Stability 88 (2005) 435e443 [27] Turell ME, Bellare A. Eect of crystallization temperature on the morphology of high pressure crystallized ultra-high molecular weight polyethylene. Polym Mater Sci Eng 2002;87:471. [28] Schmitt S. TEM analysis of microstructural morphology in ultra high molecular weight polyethylene. Scripta Mater 2000;43:5238. [29] Gedde UW. Polymer physics. Chapman and Hall; 1995. [30] Goldman M, Gronsky R, Ranganathan R, Pruitt L. The eects of gamma radiation sterilization and ageing on the structure and morphology of medical grade UHMWPE. Polymer 1996;37(14):290913. [31] Goldman M, Gronsky R, Pruitt L. The inuence of sterilization technique and ageing on the structure and morphology of medical-grade ultrahigh molecular weight polyethylene. J Mater Sci: Mater Med 1998;9:20712.

443

[32] Goldman M, Gronsky R, Long GG, Pruitt L. The eects of hydrogen peroxide and sterilization on the structure of UHMWPE. Polym Degrad Stab 1998;62(1):97104. [33] Lin L, Argon AS. Review structure and plastic deformation of polyethylene. J Mater Sci 1994;29:294323. [34] Bershtein VA, Egorov VM. Dierential scanning calorimetry of polymers. Ellis Horwood; 1994. [35] Blanchet TA, Burroughs BR. Numerical oxidation model for gamma radiation-sterilized UHMWPE: consideration of dosedepth prole. In: 47th Annual Meeting of the Orthopaedic Research Society; 2001. p. 1022. rtolas JA, Larrea A, Go mez-Barrena E. Fracture behavior of [36] Pue UHMWPE in non-implanted, shelf-aged knee prostheses after gamma irradiation in air. Biomaterials 2001;2:210714.