ADVANCED COMPOSITE MATERIALS TECHNOLOGY FOR ROTORCRAFT
Andrew Makeev*, University of Texas at Arlington, Arlington, Texas, USA Charles Bakis and Eric Strauch, Penn State University, University Park, PA, USA
Mark Chris, Bell Helicopter Textron, Fort Worth, TX, USA
Peter Holemans and Gina Miller, Boeing Rotorcraft Systems, Ridley Park, PA, USA Don Spencer, Kaman Helicopters, Bloomfield, CT, USA
Nicolas Patz, Patz Materials & Technologies, Benicia, CA, USA
Abstract
Composite materials are increasingly used in rotorcraft structures to reduce weight and improve efficiency. The rotorcraft industry is constantly in need of higher-performance materials that offer improved mechanical strength and stiffness at a lower weight. In polymer-matrix composite structures, matrix-dominated failures impose severe limitations on structural performance. The objective of this work is to advance composite material technologies for rotorcraft through the use of nanoadditives to improve structural efficiency. Technical challenges and potential solutions for improving matrix-dominated performance of prepreg composites through nanoparticle reinforcement, are discussed. In particular, a promising technology for improving compression and interlaminar strength and fatigue performance, is identified. The advanced materials technology is based on high weight content loading of approximately 100-nm diameter nanosilica particles in low-viscosity resins. Such technology resulted in compression strength improvement for intermediate-modulus carbon-fiber/epoxy-matrix 250° F curing prepreg composites as recently demonstrated by 3M. This work not only supports the initial findings of 3M regarding the improvement of compression strength performance but also demonstrates improved interlaminar material properties including fatigue performance, and expands the material design space. Fatigue performance is critical to rotorcraft dynamic components as they are subject to extreme oscillatory flight loads that can result in material fatigue failures.
1. INTRODUCTION1
Fiber-reinforced composite materials are increasingly used in rotorcraft primary structures. The industry is constantly in need of higher-performance materials that offer improved structural performance at a lower weight. Rotorcraft dynamic components are among the most challenging composite applications as they are subject to extreme flight loads that are oscillatory in nature and cause material to fail in fatigue. In polymer-matrix composites, matrix-dominated failures such as delamination and low fiber-direction compressive strength compared to the fiber-direction tensile strength, impose significant limitations on structural performance and longevity characteristics. In fiber-reinforced polymer composites, the polymer matrix and the matrix-fiber interface are much weaker than the fibers. The incorporation of nano-sized reinforcement in the matrix may improve the recognized weaknesses of composites, such as interlaminar and compressive strengths. However, the implementation of such advanced materials in
1
Presented at the 39th European Rotorcraft Forum, Moscow, Russia, September 3-6, 2013.
*
Corresponding author, e-mail: makeev@uta.edu.
rotorcraft has been limited by conflicting information in the literature on the best approach for the enhancement of matrix-dominated properties, the lack of suitable material property data, and unproven repeatability and manufacturability at the structural scale. Therefore, a need exists to develop a knowledge base that characterizes advanced material technologies in conjunction with the manufacturing methods employed to process and post-process the advanced materials, establish process controls, and finally demonstrate the feasibility of the advanced materials, with the emphasis on fatigue life and life-cycle costs, versus conventional materials for rotorcraft applications.
The Vertical Lift Consortium, which represents a collaboration of U.S. Government, rotorcraft industry, and academia to develop and transition innovative vertical lift technologies, recently started the Advanced Materials Technology (AMT) Program with a goal to advance material technologies for the improvement of rotorcraft material strength and fatigue behavior. Specific objectives included (a) screen state-of-the-art material technologies; (b) select the most promising materials for improved matrix-dominated performance and acceptable processing and handling qualities; and (c) develop a database of material properties for use in structural design. Thus, a knowledge base
providing a foundation for the insertion of advanced materials in rotorcraft applications, is being developed. The AMT program is a multi-year collaborative effort of U.S. rotorcraft Original Equipment Manufacturers (OEMs) including Bell Helicopter, Boeing Rotorcraft, Sikorsky Aircraft, Kaman Helicopters, and research labs of academic institutions including University of Texas Arlington (UTA) and Pennsylvania State University (PSU). The AMT program also engaged commercial manufacturers of pre-impregnated fiber reinforced polymer composites (prepreg) and offered a unique opportunity for the rotorcraft OEMs and the material manufacturers to work together towards the development of material solutions that improve structural strength and fatigue behavior.
The AMT program started with a screening effort which identified the most promising candidate technologies for improvement of interlaminar fatigue performance and compressive properties. Such improvement was compared to the prepregs (carbon and glass fibers) currently used in rotorcraft structures. Not only structural performance but also manufacturability has been considered. For example, success criteria included improvement of mechanical properties of 350° F curing systems relative to a legacy baseline 350° F cure toughened epoxy system (Hexcel 8552 [1]) without deteriorating processing and handling qualities such as viscosity, tack and working life. Currently, several rotor system applications use 250° F curing resins due to the relatively high viscosity of 350° F curing toughened epoxy systems such as 8552 [1]. Success criteria also included the improvement of mechanical performance of 250° F curing systems relative to legacy baseline 250° F cure toughened epoxy systems such as Cytec 381 [2] and E773 [3], at a lower viscosity compared to 8552. Among many candidates for improving compressive and interlaminar properties of composite materials are nanosilica-loaded matrices. Nanosilica (~100-nm diameter silica particles) is cost-effective; enables high loading (more than 40% weight content in the resin) with minimum impact on viscosity; and can be uniformly dispersible through surface chemistry technology (functional groups) [4, 5]. In 2009, 3M launched 3M™ Matrix Resin 3831, a 36% nanosilica weight content 250° F curing epoxy resin system designed for use in composite prepreg manufacturing processes [5]. Initial implementation of this resin in the sporting goods market has produced carbon-fiber composite fishing rods with 60-90% increased compression-dominated bending failure loads [5]. 3M disclosed the material morphology as well physical and mechanical properties of their resin systems subject to various nanosilica loadings. Also, mechanical properties of carbon/epoxy nanosilica prepregs were published [4, 5]. Unidirectional prepreg tape for each of the resin systems studied in Refs [4,
5] was produced by Patz Materials and Technologies (Benicia, CA) using TR50S carbon fiber (Grafil Inc., Sacramento, CA). In this work, the material characterization is expanded to carbon-fiber and glass-fiber prepregs applicable to rotorcraft structures. In particular, compression and interlaminar material properties including fatigue curves are determined and compared to existing production rotorcraft prepreg material systems. It is worth noting that the previous studies employed standard testing and never questioned their suitability for measuring true material properties. For example, the ASTM D 2344 standard short-beam shear (SBS) test method [6] used in Ref [5] does not capture the interlaminar shear (ILS) strength and modulus material properties. Also, ILS fatigue characterization of carbon-fiber and glass-fiber epoxy nanosilica prepreg composites has not been accomplished before. Furthermore, assessment of the interlaminar tensile (ILT) material properties, including strength, modulus, and S-N curves, has not been attempted in the previous studies. This work not only supports the initial findings of 3M regarding the improvement of compression performance of intermediate-modulus carbon-fiber/epoxy-matrix 250° F curing prepreg composites but also expands the material design space to glass-fiber prepregs, and clarifies the interlaminar properties of nanosilica prepreg composites including fatigue behavior.
2. MATERIALS
As noted in the previous section, nanosilica-loaded prepreg composites are considered in this work. References [4, 5] provide the details of the material structure for 250° F curing carbon/epoxy nanosilica prepregs as well as their matrix-dominated properties measured using standard test methods. As listed in Refs [4, 5], matrix stiffness can be a primary variable affecting composite compression strength in the fiber direction because fiber microbuckling – a major compression failure mechanism – depends on the amount of support provided by the matrix to the fibers. Incorporation of hard particles into polymers increases their modulus and can increase fracture resistance [4, 5, 7]. Micron-scale inorganic fillers have been used to modify cured resin properties, but when processed into fiber-reinforced composite structures, these large particles are filtered out by the reinforcing fibers. Another undesirable effect of conventional micron-size fillers is increased resin viscosity before curing, which can compromise composite processing qualities [4, 5]. 3M attempted to achieve the desirable resin modulus and laminate compression strength improvements through the incorporation of smaller, nano-sized amorphous silica particles into thermoset-matrix resins [4, 5].
In 2009, Patz Materials and Technologies (PMT) began working with 3M to address specific applications where 3M’s nanosilica technology could
yield signific would formu technology apply the na performance is 3M’s abil nanosilica p PMT suppl surface che nanosilica p polymer sy attributes co This also a polymer sys viscosity of agglomerate dispersed th filtration by t The highly particles ena exceeding 5 particles fo loading leve F cured car weight cont functionalize between 7-μ Figure 1. cross sectio of conglome The compo unidirectiona fiber prepre tapes were with 48% w diameter sil weights. Co made. All p The areal w 0.0051 in (0 listed [4, approximate details of th resin system cant benefit fo ulate the app
for the spec anosilica to t e of the nano lity to tailor t particles. For ied, 3M wo emistry for th particles to be ystem, and ould be realiz llows the pa stem thus o f the formu ed compatibil hroughout the the fiber array compatible ables epoxy 50% resin w orm non-agg els. Figure 1 rbon/epoxy c tent of app ed silica μm diameter SEM image on demonstra eration for the osite materia al intermedia g tapes prod made using weight conte ica, diluted t ontrol prepreg prepreg tapes weight of car .129 mm) no 5]. The ely 60%. R e curing cyc ms. or composite plicable therm ific applicatio that polymer osilica in each the surface r each formu ould develop he particles. ecome an in thus the zed in the com articles to flo only minimal ulated produ lized nanosili e composite y. nature of t resins with le weight conten gregated dis shows a rep composite la proximately 1 particles ev carbon fibers e of a carbo ating even dis e nanosilica p als studied ate-modulus duced by PM g 250° F cu ent of appro to 15%, 25% g containing s were 12 in rbon fibers w ominal cured fiber volum References [ le as well as e structures. mosetting pol on and 3M w . The key to h polymer sy chemistry fo lated product the approp This allows ntegral part o desired pa mposite struc w freely with ly increasing uct. The ica can be ev structure wi he functiona evels of nano nt. The nano spersions a presentative aminate with 100-nm diam venly dispe s [4, 5]. n/epoxy lam spersion and particles [4, 5 by 3M incl TR50S car MT. The pre ring epoxy b oximately 100 %, 35%, and no silica was (30.48 cm) w was 145 g/m ply thickness me fraction [4, 5] report s properties o PMT ymer would o the ystem or the t that priate s the of the article cture. h the g the non-venly thout alized osilica osilica at all 250° 36% meter ersed minate d lack ] uded rbon-epreg blend 0-nm 45% s also wide. 2 . A s was was t the of the The prep Inter fiber IM8 (AGY seve sizin diffe the f only betw unid epox of th and thick Base 1.78 calc carb appr to th prep nano prod com work com prod nano PMT the a It is IM8/ prep type fiber emp prep the f mate are (suc rotor mea evol N cu also glas dete corre test In a inclu carb benc rotor AMT progr pregs appl rmediate-mod rs were selec carbon (He Y, Aiken, S eral types o ngs were use erent types of functionality y handling ween the fib directional pre xy resin (250 he carbon and 295 g/m2 knesses were ed on this i 80 for IM7 c ulated nomi bon-fiber an roximately 57 he control pr preg tapes w osilica (appr duced. It is mposites with k, were com mposites. Th
duced for the osilica resi T(20%NS) an appropriate n s noteworthy /PMT carbon pregger (PMT es, and also rs with simila ployed in the pregs are exp
following sec erial properti compared t ch as Cytec r structures w asuring true i
ving and the urves are not presents the s prepregs ermined.) A espond to th condition (70 addition to th udes the ILS bon prepregs chmark in 3 rcraft structu ram also en icable to dulus carbon cted for the i excel, Stamf SC). Glass f commercia ed in this wo f sizing was of sizing for qualities bu bers and th epreg tape w 0°F and 350° d glass fibers , respective e 0.0056 and information a carbon and inal fiber vo nd glass-fib 7% and 52%, repregs cont with up to 40% roximately 1 s worth notin 40% nanosi parable to th he nanosilic AMT program n weight nd PMT(40% nanosilica we y that 250° n prepregs a T) using the s using interm ar properties e 3M studies pected to ex ctions, fiber co es pertinent o the best 381 [2]) curr whenever po nterlaminar m AMT work is t available fo e relevant pr (S2/381 ILS All material he room-tem 0-72°F, 40-60 e 250° F cu fatigue perfo s. Hexcel 8 350° F curi ures and th gaged PMT rotorcraft n and high-st nitial evaluat ford, CT) a fibers were al sizings: 4 ork. The re that unlike c glass fibers ut also che he resin. with two gene
F cure). The s were appro ely, while n d 0.0090 in., and specific 2.475 for S olume fractio ber compo , respectively taining no na % resin weig 100-nm diam ng that dens lica, characte he correspon ca prepregs m had either content. %NS) are use eight content i F curing IM are made b same resin a mediate-mod to TR50S c s [4, 5]. The xhibit similar b ompression, to rotorcraft 250° F curi rently used i ossible. As th material prop s in progress or IM7/381 ye roperties of 2 S S-N curve properties i mperature am 0%RH). uring prepreg ormance of 3 552 [1] has ing carbon here is sign to produce structures. trength glass ion: IM7 and nd S2-glass e made with 63 and 933 ason for the carbon fibers, provides not emical bond PMT made eral types of areal weight ximately 145 nominal ply respectively. gravities of S2-glass, the ons for the osites were y. In addition anoadditives, ht content of meter) were sities of the erized in this nding legacy which PMT 20% or 40% Notations ed to indicate in the resin. M7/PMT and by the same and nanofiller dulus carbon carbon fibers erefore such behavior. In ILS, and ILT applications ng prepregs in composite he subject of perties is still s (e.g. ILS S-et), this work 250° F curing e has been n this work mbient (RTA) gs, this work 350° F curing s become a prepregs in nificant data e . s d s h 3 e , t d e f t 5 y . f e e e n , f e e s y T % s e d e r n s h n T s s e f l -k g n k ) k g a n a
available in literature. Recent publications document the development of experimental techniques for measuring accurate ILS and ILT material properties, including modulus, strength, and S-N curves, and the application of such techniques to IM7/8552 prepreg [8 – 13]. In this particular situation, not only nanosilica-loaded prepregs but also carbon nanotube (CNT) and graphene additives are considered due to their recent publicity as potential solutions for improving interlaminar properties. PMT manufactured 350° F curing IM7-carbon prepregs with 0.5% CNT, 0.8% graphene nanoplatelets (GR), and a mixture of 40% nanosilica and 0.8% graphene (40%NS + 0.8%GR) by resin weight. A uniform dispersion of CNT and GR in the PMT resin was challenging at higher weight content as the CNT length and the GR in-plane dimensions are not nm but microns. Also, CNT and GR are known to increase resin viscosity at higher weight content.
Experimental data generated in this work are presented in a limited fashion as the AMT program effort to generate reliable material performance characteristics, is in progress, and the availability of data appropriate for publication is limited. However, such limited information indicates potential benefits of the advanced materials technology to emerging rotorcraft platforms and prompts more extensive material qualification.
3. FIBER COMPRESSION
In Refs [4, 5], fiber-direction compression strength of the 250° F curing TR50S-carbon/epoxy nanosilica prepreg composites was assessed based on SACMA SRM 1R-94 [14] RTA testing of tabbed unidirectional 12-ply laminates. Nine specimens were tested for each material configuration. Table I lists compression strength data corresponding to the different nanosilica weight contents in the resin.
Table I. Fiber-direction compression strength of
250°F curing carbon prepreg composites [4, 5] Silica (wt%) Strength (ksi) FV (%)
0 258 62.5
15 267 61.2
25 274 60.3
35 276 60.3
45 287 59.2
The apparent strength changed by 11.2% at 45 wt% nanosilica loading. After the strength values were normalized to a 60% fiber volume (FV), the change from the unfilled to the most highly filled material became 17.4%.
In this work, in-plane fiber compression performance of 250° F curing PMT prepregs was evaluated using the ASTM D6641 combined loading compression (CLC) test method [15] with a 50/50 (50% 0-deg. plies and 50% 90-deg. plies) cross-ply laminate. Use of a cross-ply laminate rather than a unidirectionally reinforced laminate reduces the maximum load applied to the specimen and the CLC fixture splits the load path between face shear and end loading, thereby avoiding premature failure at either the grip entry region or the end of the specimen without the need for tabs. However, the use of a cross-ply specimen requires that classical laminated plate theory [16] be used to back-calculate the stress in the 0-deg. plies at failure. While it is recognized that variations in matrix modulus will affect the relationship between stress at the laminate level and 0-deg. ply level, back-out factors for the 0-deg. stress were held constant throughout this investigation, based on typical unidirectional ply properties: 1.6 for glass/epoxy and 1.9 for carbon/epoxy. Multiplying the laminate stress by these back-out factors provides the stress in the fiber direction of the 0-deg. plies. The fiber-direction compressive strength of a ply determined using a 50/50 cross-ply specimen is considered to be representative of the in situ compressive strength of a unidirectional ply in a wide range of practical laminate configurations used in aircraft structures [17].
Based on ply and laminate properties estimated at the start of this investigation, a 0.5 in. specimen unsupported length, and other specimen design recommendations in ASTM D6641, a [90/0]4s laminate
was selected to achieve 0-deg. ply failure prior to Euler buckling for both the carbon/epoxy and glass/epoxy laminates. The thicknesses of the glass and carbon fiber laminates were approximately 0.144 and 0.091 in., respectively. Uniaxial strain gages of 1/16th in. grid length (Measurements Group
CEA-06-062UW-350) were bonded to both sides of the specimen to record mean and bending strains in the loading direction. Strains and load were recorded throughout the compression tests using a digital data acquisition system. Five to six replicate tests were run for each type of material.
The CLC test results for 0-deg. ply compressive ultimate strength are normalized to the fiber-direction compressive strength of 250° F curing IM7/381 production prepreg composite, measured based on the SACMA testing of unidirectional laminate, which is 215 ksi (COV 3%) [2]. The CLC cross-ply laminate modulus results are normalized to the calculated cross-ply modulus of IM7/381 (11.4 Msi). These results are plotted in Figure 2. The coefficient of variation was typically less than 5% for the strength and modulus data. Such results are consistent with Refs [4, 5]. An additional set of data for the IM7/PMT(40%NS) cross-ply laminate, obtained using the SACMA test method, shows a higher 0-deg.
compressive using the CL Figure 2. carbon com All the PMT with or wit nanosilica im IM8 compo modulus of data for IM7 the SACMA strength tha test method test method the baseline Figure 3. C glass compo for PMT pre to AGY sizin Next, the fi compared fo Figure 3 s strength dat unidirectiona compressive The CLC co to a calcula S2/E773 an the strength The S2/PM nanosilica p exceeding t 30% fiber-d and modulu stronger th e strength th LC test metho Compressio posites with a T systems ex thout the na mproves the osite (with 4 the baseline 7 fiber and 40 A test met an similar m d. Cross-ply ds. The RTA e, in this case Compression osites with an epregs starts ng) iber-direction or the 250° F shows the ta normalized al S2/E773 e properties t ompression ated cross-p nd S2/381. T and modulu MT prepreg performs the the baseline direction com us. The 9 han the 463
han the same od. on results fo and without 4 xceed the ba anofiller. T strength the 40% nanosili system. The 0% nanosilic thod, shows aterial tested laminates w A strength im e, was as high n results for 2 nd without na with a numb n compressiv F curing glas compressio d to the SACM 3 which to S2/381 tap modulus dat ply modulus The coefficie s data was le with 933 s best of all th in terms of mpressive st 933 sizing is 3 sizing at e laminate te or 250° F c 40% nanosilic aseline streng The use of e most. Only ca) exceeds e additional s ca, obtained u s slightly h d using the were used in mprovement h as 45%. 250° F curing anosilica (not ber correspon ve properties ss/epoxy syst n modulus MA test resul exhibits si pe (185 ksi) [ ta are norma of 4.45 Ms nt of variatio ess than 5%. sizing with he PMT syst strength (al trength incre s always sli t any nano ested curing ca gth – 40% y the s the set of using igher CLC both over g S2-tation nding s are tems. and ts for imilar [2, 3]. alized si for on for 40% tems, most ease) ightly osilica conc glas with of al 4. In R 250° com RTA Sam mea span stren weig Tab AST silica the h It is capt mod mate SBS 10, to fa the stan A fe this coup appr long is m carb coup nose the S centration inv s and carbo the 40% na l the PMT sy INTERLAMI Ref [5], short ° F curing TR mposites was A testing o mples of 10 asuring 0.25 n of 4t (0.5 i ngth data cor ght content in ble II. ASTM c Silica (wt 0 15 25 35 45 TM D 2344 S a content. A highest nano known that ture ILS mate dulus [6]. In erial propertie S test method 11]. The mo atigue loadin first publ ndardization p ew details spe work must b pons to mea roximately 0 g. The span modified from bon/epoxy a pons to avoid e under static SBS test setu vestigated. O n PMT prep nosilica mate ystems shown NAR SHEAR t-beam shea R50S-carbon/ assessed ba of unidirecti SBS specim x 0.75 in. ( n.) was used rresponding t n the resin sys D 2344 SBS carbon comp %) SBS St SBS strength An increase o silica concen ASTM D 234 erial strength this work, IL es were mea dology recen odified SBS t g conditions lication aim process for th ecific to the t e mentioned asure the ILS
.25-in. thick is 1.2 in. Th the ASTM D nd to 2 in d compressiv c and fatigue up [10, 11] us Overall, the 2 regs exhibit erials perform n in this comp R (ILS) ar (SBS) stre /epoxy nanos ased on ASTM onal 24-ply mens were t 2t x 6t, t=th d. Table II l to the differe stem. S strength of 2 posites [5] trength (ksi) 13.5 14.9 15.5 16.8 17.3 improved wit of 27% was ntration [5]. 44 test meth and cannot LS strength a asured using ntly develope test method [10]. Refer med at in he new metho test configura . The unidire S material pr and wide; he loading no D 2344 0.25-i n. for glass/ ve damage a loading. Fig sed in this wo 250° F curing similar trend ming the best
parison. ength of the silica prepreg M D 2344 [6] y laminates. tested, each hickness). A ists the SBS ent nanosilica 250°F curing th increasing measured at hod does not measure ILS and modulus the modified d at UTA [8, is applicable rence [11] is nitiating the od. ation used in ectional SBS roperties are and 1.75-in. ose diameter n. to 4 in. for /epoxy SBS at the loading gure 4 shows ork. g d t e g ] . h A S a g t t S s d , e s e n S e . r r S g s
Figure 4. M modulus, str Five to six strength and The digita deformation capture the strain mea geometric s modulus [8 approximatio engineering 6,000 με wa ILS stress-s glass prepre 1% enginee modulus va IM7/381 an unidirectiona Figure 5. IL curing carbo The modifie values of standard S manufacture approximatio composite is 13.3 ksi (C Similarly, th unidirectiona while Cytec strength [2] Section Are stress value 0% 20% 40% 60% 80% 100% 120% 140% 0% N or m al izad IL S S tr en gth Modified SBS rength, and fa specimens w d modulus fo al image measureme surface stra asurement w stress appro 8, 10, 11]. on correspo stress-strain as used in thi strain curves eg composite ering shear alues were 0 nd 0.567 Ms al tape. LS strength a on composite ed SBS met ILS strength SBS tests ers. For on for IM s 12.0 ksi (C COV 3.1%) he ILS streng al tape comp c lists 12.3 ks . The sam a) expression es in both cas 50% Normalized ILS S test setup f atigue perfor were tested or each mate correlation ent techniqu ain compone was combin oximation for Linear onding to th n curves bet is work. It is for unidirect es become h strain [8 – 0.709 Msi ( si (COV 4.6 and modulus es with and w hod typically h compared provided b example, th M7/381 uni COV 1.77%) apparent S gth approxim posite is 10.6 si (COV 4.1% me 3/4*(Appli n was used t ses. 100% 150% Modulus for measuring mance [10, 1 to determine erial configura (DIC) full ue was use ents. DIC b ned with si r measuring elastic mod he slope of tween 1,000 worth noting tional carbon highly nonline – 11]. The COV 2.38% 60%) for S2 results for 25 without nanosi y results in l to results by the pre he ILS stre directional while Cytec SBS strength mation for S2 ksi (COV 2.0 %) apparent ed Force)/(C to calculate s % IM7/381 IM7/PMT(Cont IM7/PMT(40% IM8/PMT(40% g ILS 11] e ILS ation. l-field ed to based imple ILS dulus f ILS 0 and g that n and ear at e ILS %) for 2/381 50° F ilica ower from epreg ength tape c lists h [2]. 2/381 02%) SBS Cross shear Figu data norm with stren and IM7/ incre with Next curin witho base fatig The cons The serv cell cond mon was the 2 Figu com To g shea was failu visua curv The Figu ILS f PMT as mea evol for IM trol) NS) NS) 1 1 N o rm a liz e d P e a k IL S S tre ss ure 5 shows a for 250° F malized to th 40% nanos ngth compar up to 20% /381 product ease compar the 3M findin t, ILS S-N c ng carbon P out nanosilic ed on consta ue tests run custom S sistent ILS fa SBS cou vohydraulic lo capacity. T dition and an nitor coupon observed. F 250° F curing ure 6. ILS S mposites with generate the ar stress and plotted agai re. The fai ally-detectab ves were det 10 million cy ure 8, were no fatigue chara T prepreg co the ILS str asuring true i ving and a tr M7/381. Bas 0% 20% 40% 60% 80% 00% 20% 2 3 the ILS mo curing carb he IM7/381 b silica exhibit red to the co higher ILS s ion prepreg c red to the con
ngs [5]. curves are de
PMT prepre ca. The S-N ant load amp
at 0.1 load ra SBS test c ilure mode [1 pons were oad frame w he tests wer n infrared th temperature. Figure 6 show g carbon com S-N curves f and without n ILS S-N curv d the baselin nst the log o lure was de ble delamina termined bas ycle runouts, ot included in acteristics for omposites ex rength prope nterlaminar m rue ILS S-N sed on the tre
R2 = 0.628 R2= 0.88 R2= 3 4 Log ( IM7/PMT(40 IM8/PMT(40 IM7/PMT(Co odulus and s bon prepreg baseline. PM up to 30% ontrol prepreg strength com composite. T ntrol prepreg etermined fo g composite N curves ar plitude unidire atio and 10 H onfiguration 10]. tested in with 5.5 kip ( re conducted hermometer w . No heating ws the ILS S-mposites. for 250° F c nanosilica ves, the ratio ne (IM7/381) of the number efined as the ation. Powe sed on linea , indicated w n the trend ap the 250° F c hibit similar i erties. The material prop curve is not end informati 881 = 0.9695 5 6 (Cycles) 0%NS) 0%NS) ontrol) strength test composites, MT prepregs % higher ILS g composite; pared to the The strength is consistent r the 250° F es with and re generated ectional SBS Hz frequency. ensures a a uniaxial (25 kN) load d at the RTA was used to g of coupons -N curves for uring carbon o of the peak ILS strength r of cycles to e onset of a er law S-N r regression. with arrows in pproximation. curing carbon improvement e subject of perties is still yet available ion, the PMT 7 8 t , s S ; e h t F d d S . a l d A o s r n k h o a N . n . n t f l e T
prepreg composites with 40% nanosilica show more than a factor of ten increase in fatigue life compared to the control laminate. The ILS fatigue data for IM7/PMT(40%NS) show larger scatter compared to the other composites. About half of the IM7/PMT(40%NS) SBS fatigue coupons were mistakenly tested with a 2-in diameter loading nose instead of the 4-in diameter. Compression damage was detected in some of the carbon SBS coupons fatigue tested with the 2-in diameter loading nose. The 250° F curing S2-glass prepreg composites with nanosilica and the appropriate fiber sizing also show significant improvement in the ILS characteristics. Figure 7 shows the ILS modulus and strength test data normalized to the S2/E773 baseline composite with a 10.0 ksi (COV 2.19%) ILS strength and a 0.609 Msi (COV 1.98%) ILS modulus.
Figure 7. ILS strength and modulus results for 250° F
curing S2-glass composites
Figure 8. ILS S-N curves for 250° F curing S2-glass composites with and without nanosilica
Figure 8 shows the ILS fatigue data with the peak stress values normalized to the ILS strength of the S2/E773 composite. S2/PMT had 933 sizing based on the best ILS strength behavior. ILS fatigue data for the S2/381 composite slightly outperformed S2/E773 (not included in the figure).
Similar to the carbon ILS fatigue data, the S-N curves were generated based on constant load amplitude unidirectional SBS fatigue tests run at 0.1 load ratio and 10 Hz frequency. The loading nose diameter was two inches. All SBS coupons exhibited shear delamination failure. The tests were conducted at the RTA condition and the infrared thermometer was used to monitor coupon lateral surface temperature. No appreciable increase of the surface temperature was detected. Figure 8 shows that nanosilica improves ILS fatigue performance of 250° F curing S2-glass PMT prepreg composites with 933 sizing compared to the legacy system.
The final sets of unidirectional SBS test data represent ILS material properties of 350°F curing carbon/epoxy composite systems. Figure 9 shows the ILS modulus and strength test data normalized to the IM7/8552 baseline composite with a 16.0 ksi (COV 2.70%) ILS strength and a 0.742 Msi (COV 2.26%) ILS modulus. The IM7/8552 SBS coupons were not manufactured (cured and machined) by the same laboratory that manufactured the IM7/PMT laminates.
Figure 9. ILS strength and modulus results for 350° F
curing carbon composites
The IM7/PMT prepreg composite with 40% nanosilica shows a 12% higher ILS strength compared to the control material and a 14% higher ILS strength compared to the IM7/8552 system. It is worth noting that IM7/PMT(40%NS) also outperformed the prepreg composites with CNT and graphene. In fact, adding only 0.8% graphene to the resin with 40% nanosilica reduced the ILT strength by 26% compared to the single 40% nanosilica in the resin. ILS strength values for the composites with 0.8% graphene only and with 0.8% graphene and 40% nanosilica were too low to support their fatigue performance evaluation. Figure 10 compares ILS fatigue data for 350°F curing carbon composites selected based on their ILS strength behavior. Constant load amplitude unidirectional SBS fatigue tests were run at 0.1 load ratio and 10 Hz frequency. The peak stress values were normalized with respect to the mean ILS strength value for the IM7/8552 composite to plot the S-N data. All SBS coupons exhibited shear delamination failure. 80% 85% 90% 95% 100% 105% 110% 115% 120% 0% 50% 100% 150%
Normalized ILS Modulus
N o rm a liz a d I L S S tre n g th S2/E773 S2/381 S2/PMT(Control) S2/PMT(20%NS) S2/PMT(40%NS) 463 Sizing 933 Sizing R2 = 0.9579 R2 = 0.7941 R2 = 0.9281 R2 = 0.6656 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 1 2 3 4 5 6 7 8 Log (Cycles) N o rm a liz e d Pe a k I L S St re ss S2/381 S2/PMT(Control) S2/PMT(20%NS) S2/PMT(40%NS) 80% 85% 90% 95% 100% 105% 110% 115% 120% 0% 50% 100% 150%
Normalized ILS Modulus
N or m al iz ad IL S S tr engt h IM7/8552 IM7/PMT(Control) IM7/PMT(20%NS) IM7/PMT(40%NS) IM7/PMT(0.8%GR) IM7/PMT(40%NS+0.8%GR) IM7/PMT(CNT)
Figure 10. composites Figure 11. 350°F curing IM7/PMT(40 ILS fatigue material and R 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2 N o rm aliz ed P eak I L S S tr e ss
IM7/PM
IM7/PM
IM7/PM
ILS S-N cur with and withSEM imag g carbon pre 0%NS) show performanc d the legacy s R R2 = 0.9245 3 IM7/85 IM7/PM IM7/PM IM7/PM IM7/PM
MT(Cont
MT(40%
MT(40%
rves for 350° hout nanofille ges of ILS fa preg compos wed significan ce compared system. R2 = 0.8514 4 5 Log (Cycles) 552 MT(40%NS) MT(20%NS) MT(CNT) MT(Control)trol)
NS)
NS)
° F curing ca ers ailure surface sites (1,000 x nt improveme d to the co R2 = 0.8948 6 7 arbon es of x) ent in ontrol Figu fatig witho Figu 350° The surfa witho mec 5. AST was Sam wide 6415 lami and was auto appr the Thes surfa 7IM
IM
ures 11 and ue failure su out nanosilica ure 12. SE °F curing car nanosilica-lo ace (i.e., sh out nanosilic chanism. INTERLAMI TM D 6415 utilized in th mples of five e unidirection 5 specificatio nate was laia matching placed on t oclave. The roximately 0. female brack se two brac aces and co
M7/PMT(
M7/PMT(
12 compa rfaces of car a. M images o bon prepreg oaded resin ear hackles) ca, confirming NAR TENSIO curved-beam his work for to six 0.25 – nal CB coup ons to determ d on a male female tool top of the la outer corner .25 in. radius ket had appr ckets forme orners of the(Control)
(40%NS
re SEM ima rbon compos of ILS failure composites ( shows a rou ) compared g a different ON (ILT) m (CB) test measuring I – 0.26 in. thi ons tested p mine ILT stre e tool (an an(another an aminate and r of the male s and the inn roximately 0. ed the inner e curved-bea
)
)
ages of ILS ites with andsurfaces of (5,000 x) ugher failure to the resin shear failure method [18] ILT strength. ick and 1 in. per ASTM D ength. A CB gle bracket), ngle-bracket) cured in an bracket had ner corner of .5 in. radius. r and outer am laminate. S d f e n e ] . . D B , ) n d f . r .
After the la diamond saw Figure 13 sh a typical t specimens displacemen The failure the CB spe thirds of the the bend, c location, and The DIC tec ASTM D64 modulus [12 Figure 13. delamination Figure 14 s modulus d composites, strength and the baseline loaded pr improvemen prepreg, in t Figure 14. F curing car However, th strength dat of the fiber COV was IM7/PMT(40 prepreg com data was 6 0% 20% 40% 60% 80% 100% 120% 140% 0% N o rm aliz ad ILT S tr e n g th aminate was w into CB co hows the AS tensile delam were tested nt rate, and mode was te ecimen radius e thickness in corresponding d quickly pro chnique was 415 stress c 2]. ASTM D n failure of a shows norm data for th , normalized d a 1.77 ksi e IM7/381 com repreg com nt in the mea this case, as ILT strength rbon compos he coefficient ta was much compression 17.6% for I 0%NS); and mposites. T 6.66% for I 50% Normalized ILT s cured, it upons. STM D6415 C mination fai at the stan d exhibited ensile delam s area, typic nward from th g to the max opagating thro s used in con calculation to 6415 test s unidirectiona malized CB s he 250° F to a 6.5 ksi (COV 4.15% mposite. The mposites s n CB strengt high as 62% and modulu ites t of variation higher comp n and the ILS M7/PMT(Co 4.31% for IM The COV of M7/PMT(Con 100% 150 T Modulus was cut wi CB test setup lure. All s ndard 0.05 in similar beha ination starti cally at about he outer radi ximum ILT s ough the flan njunction with o determine setup and te al CB specime strength and curing ca (COV 15.1% %) ILT modul e 40% nanos show signif th over the co . us results for (COV) of the pared to the S properties. ntrol); 8.41% M8/PMT(40% the ILT mod ntrol); 5.57% 0% IM7/381 IM7/PMT(Con IM7/PMT(40% IM8/PMT(40% ith a p and static n/min avior. ng in t two us of stress nges. h the e ILT ensile en d ILT arbon %) CB us of silica-ficant ontrol 250° e CB COV The % for %NS) dulus % for IM7/ mate Altho nano base and the C prep is co poro com large teste was enco on th Refe in th defe IM7/ 0.12 area A m on mea elem new defe stren corre was D 64 spec AST man muc It is 6415 IM8/ sizes requ to m Figu carb unid ratio appr resp com coup ntrol) %NS) %NS) /PMT(40%NS erial systems ough the me osilica-loaded eline IM7/38 20% for the CB strength preg composi onsistent with osity detects mposite) unidi er samples o ed and the C as high as ountered in t he ASTM D 6 erences [12, he IM7/8552 ects. Althoug /8552 CB ar 2%), small in a reduced ILT method for ac three-dime asurement of ment stress a w method was ects in the IM ngth and fa esponding to remarkably 415 AVG CB cimens. It ha TM D 6415 C nufacturing qu ch as ILT mat worth noting 5 CB stren /PMT(40%NS s than five to uired for a rel manufacturing ure 15 show bon compos directional CB o and 5 Hz f roximate str pect to the me mposite to p pons exhibite S); and 4.0% s.
ean ILT stren d carbon 1 is 10% fo e IM8/PMT(4 data for the ites is worth h recent asse in IM7/8552 irectional CB of the IM7/85 COV in the AS 26.5% [12]. he S-N fatigu 6415 ILT stre 13] showed CB test res gh the poros rticles was e ndividual void T strength a curate ILT st ensional c the critical d analysis was s able to capt M7/8552 CB atigue behav o the pristine higher (62% B strength ap as been dete CB strength a uality to prod terial properti g that althoug ngth of the S) was relat o six coupon liable assess g defects inclu ws CB fatigue sites. Co B fatigue tes frequency. T ress values ean CB stren lot the S-N ed tensile dela % for IM8/P ngth improve composites or the IM7/P 40%NS), larg baseline and noting. The essment of t 2 (350° F cu B test specim 552 CB spec STM D 6415 Large scat ue data gene ess approxima that such a sults was due
ity volume co extremely low ds present i and caused la trength calcu computed defects coupl developed [1 ture the effec
specimens vior. The I e (porosity-fre ) compared t proximation f ermined that s approximation duce the CB es. gh scatter in IM7/PMT(4 tively low, la ns tested in t sment of the uding porosity e data for 2 onstant load sts were run The peak AS were norm ngth value for data. All amination fai PMT(40%NS) ement for the over the PMT(40%NS) ge scatter in d the control large scatter the effects of uring prepreg mens. Much cimens were CB strength tter was also erated based ation [13]. large scatter e to porosity ontent in the w (less than n the radius arge scatter. ulation based tomography ed with finite 12, 13]. The cts of porosity on their ILT ILT strength ee) condition to the ASTM for IM7/8552 scatter in the n reflects the laminates as the ASTM D 0%NS) and arger sample this work are susceptibility y. 250°F curing d amplitude at 0.1 load STM D 6415 malized with r the IM7/381 CB fatigue lure. ) e e ) n l r f g h e h o d r y e n s . d y e e y T h n M 2 e e s D d e e y g e d 5 h e
Figure 15. ILT fatigue data for 250° F curing carbon
composites with and without nanosilica
Large scatter in the ILT fatigue test data is evident. IM7/PMT(Control) and the IM7/381 CB coupons had similar ASTM D 6415 stress values at 1,000 cycles to failure and 10,000,000 cycle runouts. It does not necessarily mean that S-N curve is “flat” as a higher stress level might result in the same trend if large number of coupons is tested. IM8/PMT(40%NS) also had similar stress levels at much different lifetimes, from thousands of cycles to a runout in some cases. Due to limited recourses, the CB fatigue sample sizes were 10 coupons – too small for any reliable assessment of the S-N fatigue curves. And scatter in the CB fatigue data for glass composites was even worse than for the carbon composites characterized in this work.
A more thorough follow up assessment of the manufacturing defects and their effects on the ILT strength and fatigue performance is required as suggested in [12, 13]. Fidelity of the non-destructive inspection needed to characterize the critical defects becomes extremely important. The susceptibility of the CB radius area to delamination limits the nondestructive inspections to the radius area and makes such specimens strong candidates to study the effects of manufacturing defects.
6. CONCLUDING REMARKS
This work shows that incorporating nano-sized silica reinforcement in the matrix may improve the well known weaknesses of carbon-fiber and glass-fiber thermoset-matrix prepreg composites, including compressive strength in the fiber direction as well as interlaminar strength and fatigue performance characteristics. The implementation of such advanced materials technology in rotorcraft has been limited by conflicting information in the literature, the lack of reliable material property data, and unproven
repeatability and manufacturability at the structural scale. A need exists to collaboratively develop a knowledge base that characterizes advanced material technologies in conjunction with the manufacturing methods employed to process and post-process the advanced materials, establish process controls, and finally demonstrate the feasibility of the advanced materials, with the emphasis on fatigue life and life-cycle costs, versus legacy composite materials for rotorcraft applications.
Test results indicate that prepreg composites with 40% nanosilica weight content in the matrix demonstrate improved fiber-direction compressive strength and the interlaminar strength and fatigue performance, and maintain comparable density and lower resin viscosity compared to the legacy systems. As high as 45% improvement in compressive strength; 20% improvement in the interlaminar strength and more than a factor of 10 increase in fatigue life are demonstrated. Functionalized nanosilica particles are cost-effective and well-integrated in the resins used in this work. Small diameter (100 nm) of the nanosilica particles, compared to the fibers, enables uniform dispersion in the composite. On the other hand, micron-scale length of CNT and in-plane dimensions of graphene platelets cause filtration of such fillers by the fibers, and result in poor interlaminar performance of prepreg composite material systems.
Physical mechanisms governing the improvement in matrix-dominated performance must be further investigated. For example, nanosilica increases ILS and ILT stiffness of the composite system. It is expected that increased matrix-dominated stiffness provides better support to the fibers and therefore it might improve compressive strength in the fiber direction. But matrix stiffness is far from being the only characteristic driving fiber compression strength. Bond of the fibers to the resin and the nanosilica to the resin must also be strong. Appropriate sizing is critical to enabling good chemical bond of the fibers to the resin in the glass-fiber prepreg composites. Also, the SEM assessment showing a rougher ILS failure surface in the nanosilica-loaded resin compared to the resin without nanosilica, confirms different shear failure mechanism yet to be understood. Better understanding of the failure mechanisms is required for engineering optimum material reinforcement.
ACKNOWLEDGEMENTS
This work is sponsored by the Vertical Lift Consortium and the National Rotorcraft Technology Center, US Army Aviation and Missile Research, Development and Engineering Center under Technology Investment Agreement W911W6-06-2-0002, entitled National Rotorcraft Technology Center Research Program. ILS and ILT Material characterization for prepreg composites currently used in production is sponsored by the US Office of Naval Research under a Grant 0% 20% 40% 60% 80% 100% 120% 140% 2 3 4 5 6 7 Log (Cycles) N o rm aliz ed P e ak I L T S tr e ss IM7/PMT(Control) IM7/PMT(40%NS) IM8/PMT(40%NS) IM7/381
Award N00014-11-1-0916 to the University of Texas Arlington. This support is gratefully acknowledged. The views and conclusions contained in this article should not be interpreted as representing the official policies, either expressed or implied, of the US Government. The authors are also grateful to Mr. Brian Shonkwiler, Research Associate and Ms. Paige Carpentier, Ph.D. student, University of Texas Arlington, for their assistance with running the material characterization tests.
COPYRIGHT STATEMENT
The authors confirm that they, and/or their company or organization, hold copyright on all of the original material included in this paper. The authors also confirm that they have obtained permission, from the copyright holder of any third party material included in this paper, to publish it as part of their paper. The authors confirm that they give permission, or have obtained permission from the copyright holder of this paper, for the publication and distribution of this paper as part of the ERF2013 proceedings or as individual offprints from the proceedings and for inclusion in a freely accessible web-based repository.
REFERENCES
[1] Hexcel Corporation. 2012. HexPly 8552 Epoxy Matrix, Product Data, available online at http://www.hexcel.com/Resources/DataSheets/Pr
epreg-Data-Sheets/8552_us.pdf, last accessed on
July 30, 2013.
[2] Cytec. 2012. CYCOM 381 Epoxy Prepreg, Technical Data Sheet, available online at http://www.cemselectorguide.com/pdf/CYCOM_3 81_032012.pdf, last accessed on July 30, 2013. [3] Cytec. 2011. CYCOM E773 Epoxy Prepreg,
Technical Data Sheet, available online at http://www.cemselectorguide.com/pdf/CYCOM_E 773_032112.pdf, last accessed on July 30, 2013. [4] Hackett, S.C., Nelson, J.M., Hine, A.M.,
Sedgwick, P., Lowe, R.H., Goetz, D.P., Schultz, W.J. 2010. The Effect of Nanosilica Concentration on the Enhancement of Epoxy Matrix Resins for Prepreg Composites, SAMPE, available online at http://multimedia.3m.com/mws/mediawebserver? mwsId=tttttvW9lEgUmy7VpzAVpy7_2XW62EW9i Xut2Xut2tttttt--&fn=Prepreg_WhitePaper704.pdf, last accessed on July 30, 2013.
[5] Hackett, S.C., Nelson, J.M., Hine, A.M., Sedgwick, P., Lowe, R.H., Goetz, D.P., Schultz, W.J. 2010. Improved Carbon Fiber Composite Compression Strength and Shear Stiffness through Matrix Modification with Nanosilica,
American Society for Composites Twenty-Fifth Technical Conference, 2010.
[6] American Society for Testing and Materials. 2006. Standard Test Method for Short-Beam Strength of Polymer Matrix Composite Materials and Their Laminates. ASTM Standard D 2344/D 2344M, ASTM International.
[7] Kinloch, A.J., and R.J. Young. 1983. Fracture
Behaviour of Polymers. Elsevier Applied Science Publishers Ltd.
[8] Makeev, A., He, Y., Carpentier, P., Shonkwiler, B. 2012. A Method for Measurement of Multiple Constitutive Properties for Composite Materials.
Composites: Part A, 43 (12): 2199–2210.
[9] He, Y., Makeev, A., Shonkwiler, B. 2012. Characterization of Nonlinear Shear Properties for Composite Materials Using Digital Image Correlation and Finite Element Analysis,
Composites Science and Technology, 73 (2012):
64–71.
[10] Makeev, A. 2013. Interlaminar Shear Fatigue Behavior of Glass/Epoxy and Carbon/Epoxy
Composites, Composites Science and
Technology, 80 (2013): 93–100.
[11] Makeev, A., He, Y., Short-Beam Shear Method for Assessment of Stress-Strain Curves for Fiber-Reinforced Polymer-Matrix Composite Materials, Accepted for Publication, Strain, 2013.
[12] Makeev, A., Nikishkov, Y., Seon, G., Armanios, E. 2013. Methods for Assessing Interlaminar Tensile Properties in Composite Materials, Proceedings of
American Society for Composites 28th Technical Conference, University Park, Pennsylvania.
[13] Makeev, A., Nikishkov, Y., Seon, G., Armanios, E. 2013. Effects of Defects of Interlaminar Performance of Composites, Proceedings of the
39th European Rotorcraft Forum, Moscow, Russia.
[14] SACMA SRM 1R-94, “Test Method for
Compressive Properties of Orientated Fiber-Resin Composites,” Suppliers of Advanced Composite Materials Association, Arlington, VA, 1994.
[15] ASTM D6641-09, Standard Test Method for Compressive Properties of Polymer Matrix Composite Materials Using a Combined Loading Compression (CLC) Test Fixture, ASTM International, West Conshohocken, PA, 2009. [16] Jones, R.M., Mechanics of Composite Materials,
2nd Edition, Taylor & Francis, New York, NY, 1998.
[17] Adams, D. F., and Welsh, J. S. 1997. The Wyoming Combined Loading Compression (CLC) Test Method, Journal of Composites Technology
and Research, 19 (3): 123-133.
[18] American Society for Testing and Materials. 2006. Standard Test Method for Measuring the Curved Beam Strength of a Fiber-Reinforced Polymer-Matrix Composite. ASTM Standard D 6415/D 6415M, ASTM International.
