Surface and Coatings Technology 133᎐134 Ž2000. 68᎐77
Fatigue properties of a 4340 steel coated with a Colmonoy 88 deposit applied by high-velocity oxygen fuel a L. Hernandez , F. Oliveiraa , J.A. Berrıos ´ ´ b, C. Villalobos b, A. Pertuz a , E.S. Puchi Cabrerab,U a
School of Mechanical Engineering, Faculty of Engineering, Uni¨ ersidad Central de Venezuela, Apartado Postal 47885, Los Chaguaramos, Caracas 1045, Venezuela b School of Metallurgical Engineering and Materials Science, Faculty of Engineering, Uni¨ ersidad Central de Venezuela, Apartado Postal 47885, Los Chaguaramos, Caracas 1045, Venezuela
Abstract The fatigue behavior of a quenched and tempered AISI 4340 steel has been evaluated in three different conditions: Ža. uncoated; Žb. grit-blasted with alumina; Žc. grit-blasted and coated with a thermal-sprayed Colmonoy 88 deposit 220 m thick, employing a high-velocity oxygen fuel ŽHVOF. gun. The results indicate that grit blasting the base steel can lead to a significant reduction in the fatigue properties of the material. The microscopic observation of the fracture surfaces of the blasted samples indicates that the fatigue processes are initiated at the alumina particles that were retained within the matrix near to the surface of the specimens, giving rise to the presence of stress concentrators that act as nucleation sites of the fatigue cracks. Coating the blasted substrate with this kind of deposit leads to a further reduction of the fatigue properties of the base steel. Such a reduction has been associated with the fracture and partial detachment of the coating from the substrate along the substrate-deposit interface and the reduction in the area of the load-carrying segments of the composite material during fatigue testing. SEM observations of the fracture surfaces of coated samples tested at low alternating stresses, that support this view, have been presented. Under some alternating stresses, the HVOF deposits are believed to contribute to endure the stress applied to the material and therefore the evaluation of the fatigue properties of this kind of coated materials must take into account the thickness of the deposit sprayed unto the substrate. The analysis of the fracture surfaces of the coated specimens revealed that, in this case, the crack nucleation sites are also associated with the presence of Al 2 O 3 particles at the substrate-deposit interface. The fatigue performance of the material under the different conditions analyzed has been quantified by determining the Basquin parameters from the fatigue life curves obtained. 䊚 2000 Elsevier Science S.A. All rights reserved. Keywords: Fatigue; High-velocity oxygen fuel; Steel; Colmony
1. Introduction As has been pointed out by van den Berge w1x, thermal spray technologies involve a number of processes including plasma, wire-arc, flame, high-velocity oxygen fuel ŽHVOF. and detonation-gun, by means of which both metallic and non-metallic materials, U
Corresponding author. Tel.: q58-2-662-8927; fax: q58-2-7539017. E-mail address:
[email protected] ŽE.S. Puchi Cabrera..
which do not sublimate or decompose at temperatures close to their melting point and are available in the form of wire or powder, are converted into a spray of molten and semi-molten particles employing a heat source. Such particles are subsequently deposited onto a substrate in order to improve, among other properties, wear and corrosion resistance, and to restore worn or undersized parts. The sprayed material generally adheres to the substrate through a mechanical bond which can achieve a strength of up to 70 MPa. Regarding HVOF processes w2x, the guns that are employed
0257-8972r00r$ - see front matter 䊚 2000 Elsevier Science B.V. All rights reserved. PII: S 0 2 5 7 - 8 9 7 2 Ž 0 0 . 0 0 8 7 4 - 4
L. Hernandez et al. r Surface and Coatings Technology 133᎐134 (2000) 68᎐77 ´
commercially utilize an internal combustion jet to generate supersonic gas velocities that can achieve approximately 1800 m sy1 . This technology allows the deposition of metal alloys, ceramic᎐metal composites and polymers, whose powders are injected directly into the combustion region of the gun. Thus, the semi-molten particles are accelerated in the high-velocity gas stream, exiting the gun at speeds of approximately 400᎐750 m sy1 , which gives rise to dense coatings, with a porosity of less than 1%, and bond strengths usually in excess of 80 MPa. According to Sartwell and Bretz w2x, current and anticipated environmental regulations, together with concerns over the performance of hard chrome deposits, widely used in the military industry, have led to the qualification of different types of coatings that could be employed as an alternative. In this regard, HVOF thermally sprayed coatings seem to represent the best available technology for the replacement of such a plating in the manufacture and maintenance operations on aircraft vehicles and ships, ground vehicles and machinery. For this purpose, a number of research projects have been conducted in order to evaluate the fatigue, corrosion and wear performance of different substrates materials coated with HVOF deposits and to compare such a performance with that observed after hard chrome plating. In the past few years there has been a considerable interest in the study of the fatigue properties of HVOF thermally sprayed materials w3᎐11x. For example, Brandt w12x conducted a study on the fatigue strength of both an Al-3% Mg alloy and mild steel, previously grit-blasted with Al 2 O 3 particles of different sizes in the range of 180᎐1180 m, at an air-flow pressure of 0.25 MPa. The substrates were subsequently coated with WCrCo 83:17, WCrCorCr 86:10:4 and Cr3 C 2rNiCr 75:25 HVOF deposits and tested in the alternating stress range of approximately 110᎐290 MPa. The results of this work allowed to conclude that the carbide coatings produced by this process had porosity levels of less than 1% and behaved like an homogenous material. Also, it was determined that grit blasting the aluminum samples with Al 2 O 3 particles of a size in the range of 600᎐1180 m, gave rise to a high surface roughness and notches that brought about a reduction in fatigue strength in comparison with particles of a smaller size, particularly at low alternating stresses. Also, it was concluded that steel or aluminum substrates coated with such deposits do not exhibit any change in fatigue strength and that machine parts with HVOF carbide coatings can be safely designed by ensuring that the maximum stress applied do not exceed the endurance limit of the substrate material. More recently, Watanabe and co-workers w13x have carried out an investigation in order to evaluate the fatigue behavior of a medium carbon steel Ž0.43 C, 0.21 Si, 0.66 Mn, 0.081 P, 0.008 S, 0.13 Cu, 0.09 Ni and 0.20
69
Cr, wt.%. previously grit-blasted with Al 2 O 3 and subsequently HVOF thermally sprayed with WC-Co coatings of 50 " 10 and 100 " 10 m thickness. For the deposition process two types of WC-12 wt.% Co powders were employed: Ža. cast and crushed; and Žb. agglomerated and sintered. The fatigue tests were conducted under rotating bending conditions at 36 Hz, at different stress amplitudes in the range of 250᎐400 MPa, calculated using the substrate specimen diameter. In this work, it was reported that the fatigue limits of the grit-blasted samples and the coated specimens were similar, but higher in comparison with that determined for the annealed substrate. In general, it was also observed that the fatigue behavior of the specimens with a thinner coating was better than that of the thicker ones, although it was particularly noticeable for the deposits obtained with the cast and crushed powder. This work allowed to conclude that the morphology of fatigue cracks within the coatings was dominated by the bonding state at the interface of the deposited particles and at the substrate ᎐deposit interface. Sartwell and Bretz w2x have reported the results of a number of low cycle fatigue ŽLCF. and high cycle fatigue ŽHCF. tests conducted with smooth bars of 4340 steel in different conditions, including uncoated and coated with hard chrome, HVOF WC-Co Ž17% Co. and HVOF triballoy 400, a Co᎐Mo᎐Cr alloy. The tests were carried out at alternating stresses in the range of approximately 130᎐230 MPa. The S vs. N results indicated a substantial reduction in fatigue life for the hard chrome plated specimens, whereas the data for both of the HVOF coatings were observed to be within the statistical uncertainty of that of the uncoated samples, which implied essentially no loss of fatigue strength. Thus, the present investigation has been conducted in order to study the fatigue behavior of an AISI 4340 steel which has been oil quenched and tempered prior to grit blasting with Al 2 O 3 and coating industrially by HVOF thermal spray with a Colmonoy 88 Ža Ni-W-CrFe-Si alloy. deposit of approximately 220 m thickness and to compare the results obtained with those of the uncoated material in order to quantify the change in fatigue strength both after blasting, and blasting and coating.
2. Experimental techniques The present investigation has been carried out with samples of an AISI 4340 steel with the following composition Žwt.%.: 0.41 C, 0.69 Mn, 0.24 Si, 0.25 Cu, 0.79 Cr, 0.23 Mo and 1.73 Ni. This material is widely used in the production of automotive crankshafts and rear axle shafts, aircraft crankshafts, connecting rods, propeller hubs, gears, drive shafts, landing gear parts and heavy
70
L. Hernandez et al. r Surface and Coatings Technology 133᎐134 (2000) 68᎐77 ´
duty parts of rock drills. The alloy was provided as bars of approximately 16 mm diameter and 6 m length. Such bars were cut to pieces of approximately 120 mm length for machining tensile specimens and of 90-mm length for machining the fatigue samples. Both type of specimens had a gauge diameter of 6.35 mm and shoulder diameter of 12.7 mm. The tensile samples had a gauge length of 32 mm and fillet radius of 5 mm, according to the ASTM standards A 370-97 w14x. The fatigue specimens had a curved gauge length of 38.1 mm along the cord, machined following a continuous radius of 58.73 mm. The material was already provided in the quenched and tempered condition from which 91 samples were machined for both type of tests. The substrate specimens were subsequently ground with successive SiC papers grit 100᎐1200 and polished mechanically. Fifty fatigue and six tensile samples were cleaned, pre-heated and grit-blasted with Al 2 O 3 particles grit 24, at a pressure of 621 kPa, at a distance of 30 cm normal to the surface of the specimens. Three tensile specimens and 25 fatigue samples previously blasted were subsequently thermal sprayed industrially at Plamatec Ingenieros C. A. ŽGuarenas, Venezuela., employing a HVOF JP-5000 gun under the following conditions: fuel pressure Žkerosene.: 1.17 MPa, oxygen flux: 11.75 l sy1 , nitrogen flux: 0.23 l sy1 , spraying distance: 330 mm, fuel flux: 0.0063 l sy1 and powder ŽColmonoy 88. feeding rate: 1.5 g sy1 . The deposit had a thickness of approximately 220 m and its chemical analysis was determined by means of SEM techniques ŽHitachi S-2400, Japan. with EDS facilities. The observations were conducted at a constant potential of 20 kV. Tensile tests were carried out on a computer-controlled servohydraulic machine ŽInstron 8502, Canton, USA. at a cross-head speed of 3 mm miny1 . At least three samples were employed for characterizing the monotonic mechanical properties of both the coated and uncoated substrate. Fatigue tests were carried out under rotating bending conditions ŽFatigue Dynamics, RBF-200, Walled Lake, USA. at a frequency of 50 Hz and alternating stresses of 612, 650, 688 and 726 MPa, for the uncoated substrate, which corresponds to 59, 63, 66 and 70% of the tensile strength of the unplated substrate, respectively. For the grit-blasted samples, the tests were conducted at 517, 539, 566, 589 and 634 MPa, corresponding to approximately 45, 47, 50, 52 and 56% of the tensile strength of the base steel. The blasted and thermal sprayed samples were tested at alternating stresses of 463, 482, 500, 518 and 542 MPa, which corresponds to 50, 53, 54, 56 and 59% of the tensile strength of the substrate, respectively. The alternating stresses applied to the coated samples were calculated taking into consideration the thickness of the deposit. A total of 24 samples were employed for evaluating the fatigue properties of the uncoated substrate, 25 for the grit-blasted material and
25 for the blasted and coated samples, which exceeds the minimum number of specimens required in S᎐N testing for reliability data according to the ASTM standard 739 Ž12᎐24 samples.. Thus, the testing procedure followed in the present work allowed a replication greater than 80%. The samples corresponding to the substrate material were mechanically prepared in order to have similar polished, mirror-like surfaces before testing. On the contrary, the grit-blasted and blasted and coated specimens were tested without further preparation. The fracture surfaces of the blasted and blasted and coated samples that failed at a number of cycles close to the mean, at the lowest and highest alternating stresses, were examined by means of SEM techniques, particularly in relation to the initiation sites of the fatigue cracks, morphology of the fracture surfaces and the different stages of their propagation.
3. Results and discussion 3.1. Characteristics of the deposit Fig. 1 illustrates a SEM photomicrograph of the Colmonoy 88 deposit and the interface between the coating and the substrate previous to fatigue testing. As can be appreciated, the deposit is quite smooth whereas the interface is observed to be somewhat irregular, free of cracks but with the presence of large second phase particles. Such a distortion is believed to be associated with the large plastic strain induced by grit blasting. On the other hand, Fig. 2 shows a typical EDS analysis conducted on the deposit which allowed to determine its approximate chemical composition Žwt.%.: 44.52 Ni, 23.60 Fe, 15.68 W, 12.85 Cr, 2.06 Si and 1.29 Al. The hardness of the deposit was found to vary from approximately 650 " 85 HVN300 , near to the surface of the coating, to approximately 700 " 85 HVN300 , at the substrate-deposit interface. 3.2. E¨ aluation of mechanical properties The influence of the deposit on the monotonic mechanical properties of the composite coating᎐substrate material was evaluated by means of tensile tests. Thus, a number of samples were tested both in the uncoated and grit-blasted and coated conditions in order to compare their behavior. The Colmonoy 88 deposits plated onto the substrate employed in the present study were observed to give rise to a slight reduction both in yield stress and ultimate tensile strength ŽUTS. in comparison with the uncoated steel. The mean yield stress of the plated samples was found to be approximately 872 " 7 MPa, in comparison with 1037 " 43 MPa for the uncoated material, whereas the UTS was found to
L. Hernandez et al. r Surface and Coatings Technology 133᎐134 (2000) 68᎐77 ´
Fig. 1. SEM photomicrograph illustrating the Colmonoy 88 deposit ŽD. and the interface between the coating and the substrate ŽS. previous to fatigue testing. The deposit is quite smooth whereas the interface is somewhat irregular, although free of cracks. Large second phase particles ŽP. are also observed.
be approximately 919 " 6 MPa, in contrast to 1143 " 29 MPa for the base steel. Both parameters were calculated taking into consideration the thickness of the coating. These results can be explained in terms of the relative difference in mechanical properties between the Colmonoy deposit and the substrate. During testing of the coated samples, the deposits were observed to fracture at an early stage, leaving the substrate as the only load-carrying element throughout the test. These observations are in agreement with the results previously reported by Brandt w12x regarding the determination of the maximum tensile strength of a number of HVOF coatings by means of three-point bending tests conducted on aluminum substrates. According to this author, the maximum value of this property was found for WC-Co 83r17 and Cr3 C 2-Ni 83r17 deposits, of
71
approximately 380 MPa, although it was also acknowledged that such a property can be influenced by the residual stress conditions caused by dissimilar metal coatings and different surface shapes. In relation to the fatigue tests conducted in order to evaluate the fatigue life of the substrate, grit-blasted and blasted and coated samples, the previous determination of the monotonic mechanical properties of the material allowed to establish a stress amplitude range of 612᎐726 MPa for the substrate, which corresponded to a fraction of the UTS of approximately 0.59᎐0.70. The grit-blasted specimens were tested in the stress range of 517᎐634 MPa, i.e. 0.45᎐0.56 of the UTS. The blasted and coated samples were tested at alternating stresses in the range of 463᎐542 MPa, which corresponded to a fraction of the UTS of approximately 0.50᎐0.59. The data showing the mean number of cycles prior to fracture Ž Nf . in terms of the alternating stress applied to the material ŽS. for the uncoated, grit-blasted and HVOF-sprayed specimens, are presented in Tables 1᎐3. The results obtained have been plotted in Fig. 3 in which it can be observed that for the uncoated substrate, six tests were conducted at four different alternating stresses, whereas for both the grit-blasted samples and the HVOF-coated specimens five tests were carried out at five different alternating stresses. These testing conditions allowed to fulfill the reliability conditions for fatigue testing prescribed in the ASTM standard E 739 w15x. One of the most important aspect highlighted in this figure is the fact that both grit blasting the uncoated substrate under the present conditions, as well as the deposition of a Colmonoy 88 coating by HVOF of these characteristics onto the grit-blasted substrate steel, gave rise to a decrease of the fatigue life of the material in relation to the uncoated substrate. As shown in Fig. 4, in the grit-blasted condition, at elevated alternating stress levels Ž663 MPa., curve Žb. in Fig. 3, the reduction in fatigue life, in comparison with the Table 1 Number of cycles to failure Ž Nf . vs. stress amplitude ŽS. for the as-polished substrate specimens Stress ŽMPa.
Fig. 2. Typical EDS spectrum for the Colmonoy 88 deposits involved in the present work.
Mean S.D.
612 Cycles to fracture
650 Cycles to fracture
688 Cycles to fracture
726 Cycles to fracture
216 900 233 900 283 300 398 400 424 100 732 000
77 600 84 800 104 100 136 400 142 000 177 500
68 800 69 600 72 300 110 300 120 500 130 200
43 000 44 200 46 800 57 500 70 300 70 700
381 433 174 868
120 400 34 995
95 283 25 722
55 417 11 645
L. Hernandez et al. r Surface and Coatings Technology 133᎐134 (2000) 68᎐77 ´
72
Table 2 Number of cycles to failure Ž Nf . vs. stress amplitude ŽS. for the grit-blasted specimens Stress ŽMPa. 517 Cycles to fracture
539 566 589 Cycles Cycles Cycles to fracture to fracture to fracture
634 Cycles to fracture
358 600 488 100 662 400 888 800 1 002 200
207 900 209 000 257 400 278 500 335 500
105 000 108 700 125 700 128 300 161 600
66 500 73 000 78 100 86 700 129 900
29 600 30 700 31 400 32 400 38 600
Mean 680 020 S.D. 239 935
257 660 47 618
125 860 20 065
86 840 22 520
32 540 3165
uncoated substrate, is approximately 89.4%, whereas at low stresses Ž463 MPa. the samples undergo a reduction of approximately 79.7%. However, for the blasted and coated specimens the situation is even worse since, as shown by curve Ža. in Fig. 3 and also in Fig. 4, at 663 MPa the fatigue life is reduced by 97.4%, whereas at 463 MPa it is reduced by 95.8%. These results show an opposite trend to those previously reported by Brandt w12x for both an Al-3% Mg alloy and mild steel, and also those reported by Watanabe et al. w13x for a medium carbon steel and by Sartwell and Bretz w2x for a 4340 steel, who have reported that HVOF coatings either do not modify or give rise to an increase in the fatigue strength of the substrate material. In relation to the grit-blasted samples, in spite of the potential compressive residual stress pattern that is likely to be induced, the combined effects of blasting pressure and particle size employed in the present work have apparently given rise to a number of stress concentrators which act as nucleation sites for fatigue cracks. A similar effect was reported by Brandt w12x in Al-3% Mg alloy specimens grit-blasted with Al 2 O 3 particles of sizes in the range of 600᎐1180 m at a pressure of 0.25 MPa, which were found to
Fig. 3. Number of cycles prior to fracture Ž Nf . as function of the alternating stress applied to the material ŽS. for the substrate, grit-blasted and blasted and coated specimens.
give rise to a high surface roughness and notches that decreased the fatigue strength of the composite material. As discussed later, in the present work it was observed that during the grit blasting of the samples, some of the Al 2 O 3 particles penetrated the substrate steel at the surface and therefore became the nucleation site of the cracks whose propagation gave rise to the final fracture of the specimens. Figs. 3 and 4 also illustrate that, in contrast to the results of previous studies, the deposition of a HVOF Colmonoy 88 coating on the grit-blasted specimens, reduces even further the fatigue strength, a fact that could be partially interpreted in terms of the tests conditions. The analysis of the work conducted earlier
Table 3 Number of cycles to failure Ž Nf . vs. stress amplitude ŽS. for the grit-blasted and coated specimens Stress ŽMPa. 463 Cycles to fracture
482 500 518 542 Cycles Cycles Cycles Cycles to fracture to fracture to fracture to fracture
393 800 722 700 2 000 000 2 000 000 2 000 000
154 800 160 100 176 800 265 100 281 900
246 600 255 000 257 300 280 300 369 900
125 700 167 300 175 900 187 300 224 600
63 000 63 700 76 400 88 200 91 500
Mean 1 423 300 S.D. 713 927
208 340 55 083
281 820 45 434
176 160 31 923
76 560 11 899
Fig. 4. Percent reduction in the fatigue strength of the grit-blasted and blasted and coated specimens, in comparison with the as-polished substrate, as a function of the alternating stress.
L. Hernandez et al. r Surface and Coatings Technology 133᎐134 (2000) 68᎐77 ´
by Brandt w12x and more recently by Watanabe and co-workers w13x and also by Sartwell and Bretz w2x indicates that the stress amplitude range employed in these investigations was lower than that used in the present study. For example, the tests conducted by Brandt w12x were carried out at stresses of the order of approximately 110᎐290 MPa. However, Watanabe and co-workers w13x tested their samples at alternating stresses in the range of approximately 250᎐400 MPa, whereas the results presented by Sartwell and Bretz w2x spanned the stress range of approximately 130᎐230 MPa. As mentioned earlier, in the present work the coated samples were tested at alternating stresses in the range of 463᎐542 MPa. Taking into consideration that the maximum bending stress of most HVOFsprayed carbide coatings is in the range of 340᎐380 MPa, w12x, it would be expected that at low alternating stresses, such deposits maintained their integrity and remained uncracked and well adhered to the substrate in the tensile part of the stress cycle, and acted together with it to endure the stress applied. Even more, under these conditions, the calculation of the alternating stress on the basis of the diameter of the uncoated samples, as in the work conducted by Watanabe et al. w13x, would also lead to an apparent increase in the fatigue strength of the coated material, since the actual load-carrying area of the specimen would be larger than that considered in the computation. In order to determine whether the further reduction in fatigue strength observed in the coated specimens was associated with the fracture of the deposit and its detachment from the substrate by cracks running along the substrate ᎐deposit interface, the fatigue curve corresponding to such samples was corrected for diameter effects and re-computed assuming the nominal diameter of the uncoated material Ž6.35 mm instead of 6.79 mm approx... This re-calculation is shown as the dotted line of Fig. 3, where it is clearly observed that the corrected line falls close under the fatigue curve that corresponds to the uncoated and unblasted specimens, instead of approaching the curve that corresponds to the grit-blasted samples, as should be expected. Thus, it can be concluded that the decrease in fatigue strength observed in the HVOF-sprayed samples cannot be entirely attributed to the complete fracture of the deposit and its separation from the substrate along the interface, leaving the substrate alone to endure the load applied to the specimen. Therefore, it could be speculated that the reduction in fatigue strength brought about by coating the grit-blasted specimens, could be due to the partial fracture and detachment of the deposit from the substrate rather than its complete failure. A crude estimate of the extent to which the crosssection of the deposit should have fractured, in order that the fatigue curve of the coated samples corrected
73
for diameter effects, coincided with the curve corresponding to the grit-blasted specimens, indicates that approximately 64% of such a section should have been swept out by the main crack on the plane of fracture. This would leave in the specimen an effective diameter to endure the tensile stresses applied during bending of approximately 6.51 mm. Even more, by assuming that the maximum tensile strength of the coating is approximately 380 MPa, as suggested by Brandt w12x for other HVOF deposits, the calculations conducted considering an alternating stress of 463 MPa indicate that a Colmonoy coating of this thickness would induce a compressive residual stress in the substrate of the order of 70᎐75 MPa. However, it would be reasonable to expect that the partial fracture of the deposit, particularly at low alternating stresses, would also be associated with the partial detachment of the coating from the substrate. As discussed later, the analysis of the fracture surfaces of the coated samples revealed the localized detachment of the deposit at the substrate ᎐deposit interface, which would support the speculation that has been advanced in order to explain the present results. The plot of the number of cycles to failure as a function of the alternating stress in a double logarithmic scale can be represented by a linear relationship which indicates the validity of the simple parametric expression similar to the one earlier proposed by Basquin w16x for the description of this type of data, of the form: S s ANfym where A and m represent constants that depend on both material properties and testing conditions. A represents the fatigue strength coefficient of the material and m the fatigue exponent. Table 4 summarizes the values of the parameters A and m for the three sets of data represented in Fig. 3. The appropriate determination of the parameters A and m, particularly for the composite coating᎐substrate material, is important for the evaluation of the fatigue performance of any component made of this steel that could be HVOF sprayed either for improving some of its properties, such as corrosion and abrasive wear resistance, or achieving the required dimensions in order that worn, undersized or ground parts could fulfill properly their role in service. 3.3. E¨ aluation of the fracture surfaces of the samples The fracture surfaces of some of the grit-blasted and blasted and coated samples were examined by means of SEM techniques in order to study more closely the sites of crack initiation and their microstructural features, as well as the general morphology of such sur-
74
L. Hernandez et al. r Surface and Coatings Technology 133᎐134 (2000) 68᎐77 ´
Table 4 Parameters involved in the Basquin relationship for the conditions tested Condition
A ŽMPa.
m
Substrate Grit-blasted Grit-blasted and coated
1605.2 1245.0 1152.6
0.075 0.066 0.068
faces after the propagation of the cracks. In the coated specimens, particular attention was paid to the extent of the detachment of the deposit at the substrate-coating interface due to the propagation of secondary cracks. In relation to the grit-blasted specimens, Fig. 5a illustrates a typical fracture surface of a specimen tested at 517 MPa in which it is clearly visible that the final fracture of the sample occurred due to the propagation of several merging cracks. The origin of two of such cracks have been pointed out on the photomicrograph as A and B. Fig. 5b shows a detailed view of site A, where it is clearly observed that the nucleation of the fatigue crack is associated with the Al 2 O 3 particles that penetrated into the substrate due to the elevated blasting pressure applied. Fig. 5c represents a magnified view of the detail designated as A⬘ in Fig. 5b where the radial propagation lines indicate clearly the origin of the main crack. Finally, Fig. 5d illustrates a detailed view of site B in Fig. 5a, where again it is observed the presence of several Al 2 O 3 particles associated with the crack that emerged from this area. Thus, the microscopic evidence that has been presented accounts for the reduction in the fatigue strength of the grit-blasted specimens in comparison with the as-polished samples of the substrate, in spite of the compressive residual stresses produced by the blasting process on the surface of the material. The Al 2 O 3 particles attached to the matrix act as interfacial defects that induce the nucleation of fatigue cracks which subsequently propagate throughout the cross-section of the specimen. In relation to the blasted and HVOF-sprayed samples, Fig. 6a illustrates the general fracture surface of a specimen tested at 463 MPa where it can be observed that, after fracture, the deposit is not completely detached from the substrate but that a partial separation has occurred along some areas of the interface. Such evidence would support our view in the sense that the reduction in fatigue strength that takes place in the coated specimens could be related to the partial fracture and detachment of the coating from the substrate, leaving it and just part of the deposit as the load-carrying elements of the composite material. Again, as observed in the grit-blasted specimens, the final fracture of the samples occurred as a consequence of the action of several cracks, although the main one has been designated by A in the photomicrograph. Thus,
Fig. 6b,c shows two detailed views of the site initiation of the main crack, where it can be observed that the nucleation of such a crack is associated with the presence of Al 2 O 3 particles. Both pictures also show some secondary cracking along the substrate ᎐deposit inter-
Fig. 5. Ža. Typical fracture surface of a grit-blasted specimen tested at 517 MPa. The final fracture of the sample occurred due to the propagation of several merging cracks. The origin of two of such cracks ŽA and B. have been pointed out. Žb. Detailed view of site A. The nucleation of the fatigue crack is associated with the Al 2 O 3 particles ŽP. that penetrated into the substrate. Žc. Magnified view of detail A⬘ in Žb.. Radial propagation lines indicate the origin of the main crack. Žd. Detailed view of site B in Ža.. Several Al 2 O 3 particles ŽP. are associated with the crack that emerged from that area.
L. Hernandez et al. r Surface and Coatings Technology 133᎐134 (2000) 68᎐77 ´
75
Fig. 6. Ža. General fracture surface of a blasted and coated specimen tested at 463 MPa. After fracture the deposit displays partial separation from the substrate along some areas of the interface. The final fracture of the sample occurred due to several cracks. The main crack is shown as A. Žb,c. Detailed views of the site initiation of crack A. The nucleation of such a crack is associated with the presence of Al 2 O 3 particles ŽP.. Secondary cracking ŽSC. along the substrate-deposit interface is also observed. The coating ŽD. has maintained its integrity in this area.
face, although the coating has maintained its integrity over this area. However, Fig. 7a,b illustrates a detailed view of the site initiation designated as B in Fig. 6a, where the fracture and complete detachment of part of the deposit is clearly observed. A large Al 2 O 3 particle
associated with the crack nucleation site is clearly visible as well as, what appear to be, fatigue striations. Finally, in relation to the coated samples tested at 542 MPa, Fig. 8a shows a general fracture surface which indicates that under elevated alternating stresses, the coating has been almost completely detached from the substrate along the interface. Also, it can be seen that the fracture process occurred due to the propagation of a number of cracks, which has given rise to the presence of a number of fracture steps. Fig. 8b,c illustrates a detailed view of the sites designated as A and B, respectively, in Fig. 8a, where the fracture of the coating and its detachment along the interface are clearly visible. Also, the radial markings on the substrate fracture surface indicate the transcrystalline propagation of the fatigue cracks from such sites.
4. Conclusions
Fig. 7. Ža,b. Detailed view of the site B in Fig. 6a showing the fracture and complete detachment of part of the deposit ŽD.. A large Al 2 O 3 particle ŽP. associated with crack nucleation is clearly visible. Fatigue markings ŽFM. can also be seen.
Grit blasting of a quenched and tempered AISI 4340 steel with Al 2 O 3 particles of grit 24, at a pressure of 621 kPa, gives rise to a significant reduction in the fatigue life of the material. Such a reduction, at a stress amplitude of 463 MPa, can achieve up to approximately 83%. The decrease in the fatigue strength of the blasted material is associated with the presence of Al 2 O 3 particles which, during blasting, penetrated the substrate steel and became stress concentrators that enhanced
L. Hernandez et al. r Surface and Coatings Technology 133᎐134 (2000) 68᎐77 ´
76
Fig. 8. Ža. General fracture surface of a blasted and coated specimen tested at 542 MPa. The coating appears to be almost completely detached from the substrate along the interface. The fracture process occurred due to the propagation of a number of cracks, which has given rise to the presence of a number of fracture steps ŽFST.. Žb,c. Detailed view of the sites designated as A and B, respectively, in Ža.. Fracture of the coating and its detachment along the interface are clearly visible. Radial markings ŽRM. on the substrate fracture surface indicate the transcrystalline propagation of the fatigue cracks.
the multiple nucleation of fatigue cracks. Coating the blasted specimens with a Colmonoy 88 deposit applied by HVOF, leads to a further reduction in the fatigue strength of the material, which at an alternating stress of 463 MPa can achieve up to approximately 96%. It has been speculated that the effect of the HVOF sprayed deposit on the fatigue performance could be due to the partial fracture and detachment of the coating from the substrate along the substrate ᎐deposit interface. The localized fracture of the deposit would reduce the area of the load-carrying segments of the composite material during fatigue testing. This speculation has been supported by SEM observations of the fracture surfaces of coated samples tested at low alternating stresses. Therefore, it can be concluded that under some stress amplitudes, the HVOF deposits can contribute to endure the stress applied to the material and that the evaluation of the fatigue properties of this kind of coated materials must take into account the thickness of the deposit sprayed unto the substrate. As expected, the analysis of the fracture surfaces of the coated specimens revealed that, in this case, the crack nucleation sites are also associated with the presence of Al 2 O 3 particles at the surface of the substrate.
Acknowledgements This investigation has been conducted with the fi-
nancial support of the Venezuelan National Council for Scientific and Technological Research ŽCONICIT. through the project LAB-97000644 and the Scientific and Humanistic Development Council of the Central University of Venezuela ŽCDCH-UCV. through the project 08-17-4595-2000. The authors are grateful to Dr. B. Sartwell and Dr. K. Legg for the provision of valuable information. J.A. Berrıos ´ is also grateful to the School of Mechanical Engineering, Faculty of Engineering and Architecture of the University of El Salvador. References w1x M.J. van der Berge, Adv. Mater. Process. 12 Ž1998. 31᎐34. w2x B.D. Sartwell, P.E. Bretz, Adv. Mater. Process. 156 Ž2. Ž1999. 25᎐28. w3x H.D. Steffens, R. Dammer, U. Fischer, Advances in Thermal Spraying, The Welding Institute of Canada, 1986, pp. 417᎐425. w4x O. Knotek, R. Elsing, H.R. Heintz, Advances in Thermal Spraying, The Welding Institute of Canada, 1986, pp. 427᎐433. w5x E. Lugscheider, R. Mathesius, G. Spur, A. Kranz, in: C.C. Berndt, T.F. Bernecki ŽEds.., Thermal Spray CoatingsrResearch, Design and Applications, ASM International, 1993, pp. 569᎐573. w6x J. Wigren, L. Pejryd, D.J. Greving, J.R. Shadleyand, E.F. Rybicki, in: A. Ohmori ŽEd.., Thermal Spraying: Current Status and Future Trends, High Temperature Society of Japan, Osaka, 1995, pp. 113᎐118. w7x M. Sugano, H. Masaki, J. Kishimoto, Y. Nasu, T. Satake, in: A. Ohmori ŽEd.., Thermal Spraying: Current Status and Future Trends, High Temperature Society of Japan, Osaka, 1995, pp. 145᎐150.
L. Hernandez et al. r Surface and Coatings Technology 133᎐134 (2000) 68᎐77 ´ w8x J.U. Hwangm, T. Ogawa, K. Tokaji, in: A. Ohmori ŽEd.., Thermal Spraying: Current Status and Future Trends, High Temperature Society of Japan, Osaka, 1995, pp. 767᎐772. w9x T. Shiraishi, H. Ogiyama, H. Tshkuda, in: A. Ohmori ŽEd.., Thermal Spraying: Current Status and Future Trends, High Temperature Society of Japan, Osaka, 1995, pp. 845᎐850. w10x M. Hadfield, R. Ahmed, S. Tobe, in: A. Ohmori ŽEd.., Thermal Spraying: Current Status and Future Trends, High Temperature Society of Japan, Osaka, 1995, pp. 1097᎐1102. w11x K. Tokaji, T Ogawa, J.U. Hwang, Y. Kobayashi, Y. Harada, J. Thermal Spray Technol. 5 Ž3. Ž1996. 269᎐276.
77
w12x O.C. Brant, J. Thermal Spray Technol. 4 Ž2. Ž1995. 147᎐152. w13x S. Watanabe, T. Tajiri, N. Sokoda, J. Amano, J. Thermal Spray Technol. 7 Ž1. Ž1998. 93᎐96. w14x Standard Methods and Definitions for Mechanical Testing of Steel Products, A 370, Annual Book of ASTM Standards, 03, ASTM, Philadelphia, 1984, pp. 1᎐56. w15x ASTM E 739, Standard Practice for Statistical Analysis of Linear or Linearized Stress-Life ŽS-N. and Strain Life Ž -N. Fatigue Data, ASM Handbook, Mechanical Testing, 8, ASTM, Philadelphia, 1995, pp. 363᎐426. w16x O.H. Basquin, Proc. ASTM 10 Ž2. Ž1910. 625.
