User login

You are here

Journal Club for December 2016: Dynamic Fracture - when the going gets tough . . .

Leslie Lamberson's picture

Introduction

Nothing happens, and nothing happens, and then everything happens.” - Fay Weldon

Fracture in nominally brittle materials behaves much like the above quote, in that during the build up of stresses at the crack tip, at least from the naked eye, all remains calm – until suddenly and catastrophically, the crack takes off and propagates through a material.  On some level, one could argue that all fracture is dynamic; however it is generally acknowledged in the mechanics community that the dynamic nature of cracks is relevant to failure criteria when the inertia is large enough that it requires the kinetic term to be included in the total energy balance. 

While not a new subject, there remains fundamental open questions in dynamic fracture, including two most relevant to this discussion:

  • How fast can cracks propagate?
  • What are the physics governing the mechanisms of dynamic crack tip instabilities?

crack paths

Figure 1:  Homalite 100 plate with center hole and pre-crack in slight tension (less than 4 MPa) fractured from an out-of-plane high-energy-density impact which acts like a local explosion, where resulting cracks grow left to right in the field of view.  Under certain conditions, cracks grew faster and exhibited branching characteristics, as circled (top), whereas in other less-extreme cases, cracks exhibited slower speeds and an oscillating crack path [1].  

Background  

I feel the need . . . the need for speed.”   - Maverick & Goose, Top Gun (1986)

Even though the study of dynamic fracture in nominally brittle materials has been around for quite some time (often first credited to Shardin who examined dynamic crack growth around in the 1930), there are some puzzling contradictions behind some of the continuum dynamic fracture mechanics theory; particularly in regard to instabilities, or when transient effects come into play. 

Pragmatically speaking, the American Society for Testing and Materials has well-established and detailed standards for static fracture toughness testing, including ASTM E399 for metals, and ASTM C1421 for ceramics [2-4]; yet no generally accepted procedure  to examine the dynamic fracture toughness of these materials exists.  Dynamic fracture characterization methods are generally grouped by types of high-rate bending, high-rate tension, dynamic wedging, and single-point impact (with an extensive body of work on wire explosions for mode II fracture) [4-9]. At the same time, there lies a key discrepancy between static and dynamic initiation fracture toughness values, some materials such as aluminum 7075-T6 have been shown to exhibit little rate dependency, while others such as Ti-6264 and 4340 steel show as much as 1.4 to 1.6 times higher toughness under dynamic loading conditions [10].  For PMMA, the dynamic fracture toughness was shown to be three times higher than it’s static initiation value [11].  Indeed, our research group has shown that a MAX phase (metal-ceramic nanolayered material) Ti3SiC2, exhibited no real rate dependency on its dynamic initiation toughness, whereas tungsten carbides with various levels of nickel and cobalt binders exhibited dynamic toughness increases on the order of 20% from their quasi-static values [12,13].  

From these selected examples, differences appear to be key signatures of the lower length scale initiation mechanisms which govern material inertia or resistance to initiation in the formation of a macrocrack, as well as a function of (given different dynamic loading conditions with various methods of evaluating dynamic fracture) wave interactions in the material, ahead of the moving crack front where the theory used to extract relevant stress intensity values is leveraged.  Regardless of these material differences, those few researching experimental dynamic fracture appear to agree that the history of loading on the crack tip, even in rate-insensitive materials, directly affects resulting crack growth [14].    

 MAX Phase Ti3SiC2

Figure 2:  MAX phase Ti3SiC2 fracturing from the pre-crack and notch on the right hand side, when impacted by a steel rod at 4 m/s on the left hand side.  In this case, DIC is used to measure deformation fields which are then optimized using a least squares overdeterministic analysis to extract the stress intensity factor in mode I (opening).  Displacement in the y-direction in millimeters shown left, and corresponding inertia absorbing microstructural features of kink bands and delaminations in resulting SEM fractography shown right [12].   

Classical Dynamic Fracture Mechanics  

Aye, there’s the Rub!”   - Shakespeare, from Hamlet's Soliloquy

In classical dynamic fracture mechanics, experimental evaluation leverages derived physical quantities that require adequate theory [15].  Problems dealing with dynamically loaded stationary cracks are relatively well understood, although not straightforwardly computed in terms of linear elastodynamics.  However, crack initiation and propagation under transient conditions exhibits a rapidly changing stress field at the crack tip.  While the typical square root singularity exists, the field initially is present only in an asymptotically small zone around the crack tip that spreads as a function of the wave speed, influenced by the geometry of the sample [16].  As a result, this domain has been shown to take a characteristically long time to develop, generally recognized to depend on the difference between the shear wave speed and crack propagation velocity.   Thus, the faster the crack propagates (or more dynamic the case), the longer it takes for this domain to develop ahead of the moving crack front.  Often, series expansion descriptions of the stress fields are employed, yet this approach implies steady-state conditions and consequently raises questions on the validity in accurately evaluating dynamic crack tip energetics by these means.  At the same time, it is argued and reasonably shown in literature that at least to some degree, this condition is met and the results are still useful forms of predictive fracture criterion [15].  In fact, you will see a lot of the experimental dynamic fracture literature employing various means such as long bars and low impedance boundaries with clay to help ensure this condition is met (and the validity of their solutions held); but in doing so, are essentially mitigating the interesting and under-explored, yet more complex transient regimes where crack speeds and complex wave interactions are increased, unstable crack fronts are instigated, and branching often ensues.  It should be noted that the most useful mathematical solutions considered infinite domains with asymptotic behavior of the deformation (and stress) fields near the moving crack tip, and were originally supplied by Freund and Kostrov [17,18]. 

Measuring Dynamic Stress Intensity Factors (SIF)

“Measure what is measurable, and make measurable what is not so.”  - Galileo Galilei

Characterizing fracture behavior under dynamic conditions is a fundamentally more challenging problem than its quasi-static counterpart.  Classically, the methods of photoelasticity and caustics are used to evaluate crack tip stresses, and extract stress intensity factor(s) (SIFs) by combining high-speed imaging with coherent light and optics.  The former method relies on the interference of light, while the later relies on the reflection/refraction of light.  Caustics is related to simple geometric measurements of the shadow spot at the crack tip to determine a stress intensity value, whereas photoelasticity determines the SIFs non-trivially based on the isochromatic fringe pattern that is generated when the material undergoes stress. Both methods have drawbacks of requiring a transparent or reflective material, or in the case of photoelasticity, a birefringent material [19]. Photoelasticity can be used with a greater range of materials through the application of a birefringent coating, yet issues can arise with the thickness and adherence of the coating to the true behavior of the material being investigated.  In addition, while the crack tip location is known with photoelasticity, the domain at the crack tip is generally not well resolved.  Caustics has the issue of being a purely local measurement, and there is still some ambiguity of the true crack tip location and sizing of the ‘z’ or out of plane focus in regards to encompassing the process zone [15].  These drawbacks are mentioned for discussion only, and not to understate the fact that there has been a large body of impressive work using these techniques spanning over 40 years.  

More recently, the proliferation of advanced computing equipment and development of newer full-field experimental techniques (namely shearing interferometry methods such as coherent gradient sensing, coherent digital sensing, and digital image correlation) has given rise to new methods to extract fracture behavior [20-22]. Tippur’s and Yoneyama’s laboratories, in recent times, developed a method for extracting SIFs from DIC data taken during fracture by using the relationship between SIFs and displacement around the crack tip [23,24]. This method has been applied to SENBs of tests for several materials. Yoneyama first developed the technique based on earlier work by Sanford mapping deformation fields around a crack tip to fracture properties [25]. Yoneyama’s experiments focused on an isotropic material loaded quasi-statically [24]. Kirugulige and Tippur examined mixed-mode loading in syntactic foams under dynamic loading, expanding the method to work with higher loading rates [26]. Lee, Tippur, and Bogert studied graphite/epoxy composites both prior to and during crack propagation, further broadening the applicability of the technique to anisotropic material under dynamic conditions [27].  While the spatial resolution is dramatically increased using a DIC based approach (increasing the amount of data available), the major drawback remains that the precise location of the crack tip is an unknown, and must be optimized in a nonlinear formulation to numerically determine.  My reserach group has used this increased resolution that full-field displacement mapping provides to explore uncertainty quantification in extracting SIFs; namely the influence of data size and location in regards to maintaining K-dominance, as well as quantifying error in the location of the crack tip [28]. These newer experimental dynamic fracture investigations - with increased resolution - usher in a new era of hybrid numerical-experimental methods to deal with the sheer volume of measured field data available in a meaningful way.  

caustics and interferometry

Figure 3:  Example of photoelasticity and caustics in Homalite 100 plate (FOV 150 mm) in slight tension fracturing from hypervelocity impact at 6 km/s [1]. 

Considerations for Dynamic Crack Behavior

 "When the going gets tough, the tough get going" – English proverb

So why do these differences between quasi-static and dynamic toughness exist?  How do we begin and/or continue to understand material inertia, transient effects, and their dependence on crack tip velocity and instabilities? Kalthoff in 1986 first suggested an incubation time for dynamic initiation, during which the SIF increases rapidly above the quasi-static value [29].  Aoki and Kimura in 1993 also suggested a delay time, but further surmised that it was a pragmatic result of detecting a surface crack via a 2D optical technique, where in reality the SIF at the mid-thickness of the specimen is higher than that on the outer surface, and the time is due to the propagation of the intensity to the outer surfaces (creating a dwell time) [30].  Others still have suggested that the higher dynamic toughness values from quasi-static can be attributed to the time needed to establish the singular crack-tip field as compared to the actual fracture time [31,32].  

Another train of thought for a physical explanation of this phenomenon is that the actual crack front shape (the area being created by the moving crack) might have some roughness or lower length scale irregularities- which have since been visualized by post-mortem fractography [33,34]. Theoretically, this microstructural insight has been explored by Gao and Rice in 1989 as heterogeneity along the crack tip, in addition to a weak geometric perturbation (transient effects), illustrating a dynamic toughening effect [35].   Broadly speaking, crack front perturbations and influence of local microstructural variation along the crack path have been addressed from a statistical physics view via crack pinning and depinning due to local obstacles [36].  It has been pointed out that the energy dissipation increases with crack velocity, a fact that is reasoned by the increasing process zone size by Broberg [37]. 

A simpleminded way to explore this is in terms of energy release rate, G, and assume the critical energy release rate correlates to fracture toughness of the material.  When considering the potential energy of the material with A as the crack area (where you may be used to seeing crack length a), by definition, the energy release rate is a measure of energy available for crack extension [38].  For quasi-static regimes, this is simply determined as the change in two energy states 1 and 2 (in equilibrium) as a function of change in crack area with known elastic fields U1 and U2, with the same prescribed loads for ease, so only the crack area and crack front shape are different between the two states.  State 1 is initial crack condition, and state 2 is final crack shape and area after an incremental propagation taken in the limit.

G

For the dynamic case, there is a term to account for inertial effects in kinetic energy, Ek (where the dynamic path-independent J –integral is equal to G [39]):

G(t)

where F is the work done by external forces.  For the sake of discussion, if we further assume (for simplicity) that the forces prescribed are identical, the dynamic energy release rate can be approximated (for the case of general crack advance for triangual and flat crack fronts) as:

G*(t)

So using this simple means, we can demonstrate that fundamentally the kinetic term leads to a toughening effect, as was first suggested by Chandra and Krauthammer in 1995 [40]. Further, depending on microstructural considerations which influence material resistance to fracture on a local scale (i.e. crack front shape), along with transient effects creating a gradient of forces rapidly changing across the moving crack front, one would modify the work term that was neglected here and create interplay in the overall energy balance available for crack extension.  This competition is what creates interesting and useful understanding of true dynamic fracture, and perhaps sheds light on the reported rate-sensitivity for a given type of loading, geometry and material microstructure.

Next-Generation Dynamic Fracture Mechanics?

"If we knew what we were doing, it wouldn't be called research "  - Albert Einstein

With the increase in spatial and temporal resolution of high-speed and ultra high-speed imaging (see a nice chart of the current specs on cameras by Dr. Phillip Reu here: http://photodyn.org/tools/ultra-high-speed-camera), as well as increasing resolution of full-field optical techniques, the future holds the possibility of accurately measuring not only displacements, but also acceleration fields ahead of a moving crack under dynamic loading.  This is a very powerful metric on many fronts, not only in general dynamic behavior reserach, but also specifically for dynamic fracture.  Namely, that ability would lead to a direct means of assessing the kinetic term that gives rise to the aforementioned dynamic toughening effect, and allows for more quantitative tracking of the transient nature of a moving crack front under conditions with high rates of loading and complicated wave interactions.  In some regard, tracking spatially and temporally evolving acceleration fields suggests a shift away from the classical dynamic fracture mechanics of square root singularity, or K dominant zone, and opens the door to directly extracting the energetic balance terms directly and straightforwardly (albeit with a new, yet less stringent set of assumptions to reckon with).  It should be noted that at present noise in experimental resolution of accelerations, due to the fact that with full-field methods like DIC or grid methods you are taking a derivative of the measured deformation field data not once, but twice, which magnifies noise, remains somewhat prohibitively significant (to the authors chagrin); yet again this notion demonstrates promise in moving dynamic fracture mechanics forward in regards to its fundamental understanding, remaining open questions, and associated experimental techniques, leaving restrictions of singular domains and asymptotically fitted solutions (in infinite spaces), or complications of a direct J-integral leveraging in its wake (excuse the pun).  To this end, we are presently demonstrating the full-field image-based energy method to extract dynamic fracture behavior virtually via FE modeling, examining the implication of some of the assumptions of classical dynamic fracture mechanics and correlations to the J-Integral.  One can think of this approach akin to a control volume problem, where now we are celebrating transient fields instead of avoiding them as in more classical methods [41].  It’s an exciting future for dynamic fracture mechanics- keeping in mind that the information we extract from such experiments provides essential failure criteria for relevant dynamic modeling.  Regardless, who doesn’t like to break stuff for a living?!  (See https://sem.org/dynamic-behavior-of-materials-technical-division/ for proof.)

fracture energy approach

Figure 4: Example of virtual dynamic fracture experiment in Mode I on PMMA using Abaqus explicit (with cohesive surfaces).  An impulse akin to a striker is introduced on the left hand side, and the wave propagates to the right hand side, opening up the notch in tension.  The top images show the acceleration field in the x direction, where waves are interacting and reloading the crack tip during crack extension.  The bottom plot illustrates the total energy balance (calculated from field quantities of displacement and accelerations extracted in the model, as like what would be done experimentally), with green being the input energy, red the kinetic, and navy the strain.  Note that the energy lost is going to fracture, as shown by the light blue line.  The black line shows the total energy balance of the system is maintained.  

 

REFERENCES

[1]  Lamberson, L., V. Eliasson, and A. J. Rosakis. "In situ optical investigations of hypervelocity impact induced dynamic fracture." Experimental Mechanics 52.2 (2012): 161-170.

[2] ASTM-C1327, 1999. “Standard Tests Method for Vickers Indentation Hardness of Advanced Ceramics.”

[3] ASTM-C1421, 1999. "Standard Test Methods for Determination of Fracture of Advanced Ceramics at Ambient Temperature.”

[4] ASTM-E3. 1990, 1999. “Test Method for Plane-Strain Fracture Toughness of Metallic Materials.”

[5] Shukla, A., Ed. Dynamic Fracture Mechanics. World Scientific Publishing Company Incorporated, 2006.

[6] Fengchun, J. and K. Vecchio. "Hopkinson bar loaded fracture experimental technique: a critical review of dynamic fracture toughness tests." Applied Mechanics Reviews 62.6 (2009): 060802.

[7] Rittel, D. Dynamic Crack Initiation Toughness. World Scientific, Singapore, 2006.

[8] Rittel, D., and A.J. Rosakis. "Dynamic fracture of berylium-bearing bulk metallic glass systems: A cross-technique comparison." Engineering Fracture Mechanics 72.12 (2005): 1905-1919.

[9]  Rosakis, A. J., O. Samudrala, and D. Coker. "Cracks faster than the shear wave speed." Science 284.5418 (1999): 1337-1340.

[10] Yokoyama, T. "Determination of dynamic fracture-initiation toughness using a novel impact bend test procedure." Journal of Pressure Vessel Technology 115.4 (1993): 389-397.

[11] Rittel, D., and H. Maigre. "An investigation of dynamic crack initiation in PMMA." Mechanics of Materials 23.3 (1996): 229-239.

[12] Shannahan, L., M. W. Barsoum, and L. Lamberson. "Dynamic Fracture Behavior of MAX Phase Ti3SiC2." Engineering Fracture Mechanics (2016).

[13] Jewell, P. et al. “Rate and Microstructure Influence on the Fracture Behavior of Cemented Carbides WC-Co/Ni.” Journal of Dynamic Behavior of Materials, submitting December 2016. 

[14] Belenky, Alexander, I. Bar-On, and D. Rittel. "Static and dynamic fracture of transparent nanograined alumina." Journal of the Mechanics and Physics of Solids 58.4 (2010): 484-501.

[15] Knauss, W. G., and K. Ravi-Chandar. "Fundamental considerations in dynamic fracture." Engineering Fracture Mechanics 23.1 (1986): 9-20.

[16] Ravi-Chandar, K. Dynamic Fracture. Elsevier, 2004.

[17] Freund, L. B., and R. J. Clifton. "On the uniqueness of plane elastodynamic solutions for running cracks." Journal of Elasticity 4.4 (1974): 293-299.

[18] Kostrov, B. V. "On the crack propagation with variable velocity." International Journal of Fracture 11.1 (1975): 47-56.

[19] Zehnder, A. T., and A. J. Rosakis. "Dynamic fracture initiation and propagation in 4340 steel under impact loading." International Journal of Fracture 43.4 (1990): 271-285.

[20] Tippur, H. V., S. Krishnaswamy, and A. J. Rosakis. "A coherent gradient sensor for crack tip deformation measurements: analysis and experimental results." International Journal of Fracture 48.3 (1991): 193-204.

[21] Periasamy, C., and H. V. Tippur. "Full-field digital gradient sensing method for evaluating stress gradients in transparent solids." Applied Optics 51.12 (2012): 2088-2097.

[22] Schreier, H., J-J. Orteu, and M. A. Sutton. Image correlation for shape, motion and deformation measurements. Springer US, 2009.

[23] Kirugulige, M., H. Tippur, and T. Denney. "Investigation of mixed-mode dynamic fracture in syntactic foams using digital image correlation method and high-speed photography." Proc. 2007 SEM Annu. Conf. Exposition Exp. Appl. Mech. 2007.

[24] Yokoyama, T. "Determination of dynamic fracture-initiation toughness using a novel impact bend test procedure." Journal of Pressure Vessel Technology 115.4 (1993): 389-397.

[25] Sanford, R. J. "Application of the least-squares method to photoelastic analysis." Experimental Mechanics 20.6 (1980): 192-197.

[26] Kirugulige, M. S., and H. V. Tippur. "Measurement of Fracture Parameters for a Mixed‐Mode Crack Driven by Stress Waves using Image Correlation Technique and High‐Speed Digital Photography." Strain 45.2 (2009): 108-122.

[27] Lee, D., H. Tippur, and P. Bogert. "Quasi-static and dynamic fracture of graphite/epoxy composites: An optical study of loading-rate effects." Composites Part B: Engineering 41.6 (2010): 462-474.

[28]  Shannahan, L., Fletcher, F. Pierron, L. Lamberson. "Assessment of Error and Sensitivity in DIC-Based Dynamic Fracture, Strain, submitting December 2016.

[29] Kalthoff, J. F. "Fracture behavior under high rates of loading." Engineering fracture mechanics 23.1 (1986): 289-298.

[30]  Aoki, Shigeru, and Tadashi Kimura. "Finite element study on the optical method of caustic for measuring impact fracture toughness." Journal of the Mechanics and Physics of Solids 41.3 (1993): 413-425.

[31]  Liu, C., W. G. Knauss, and A. J. Rosakis. "Loading rates and the dynamic initiation toughness in brittle solids." International Journal of Fracture 90.1-2 (1998): 103-118.

[32]  Zhang, Ch, and D. Gross. "Pulse shape effects on the dynamic stress intensity factor." International Journal of Fracture 58.1 (1992): 55-75.

[33]  Rittel, D., N. Frage, and M. P. Dariel. "Dynamic mechanical and fracture properties of an infiltrated TiC-1080 steel cermet." International Journal of Solids and Structures 42.2 (2005): 697-715.

[34] Rittel, D., and A. J. Rosakis. "Dynamic fracture of berylium-bearing bulk metallic glass systems: A cross-technique comparison." Engineering Fracture Mechanics 72.12 (2005): 1905-1919.

[35] Gao, H., and J. R. Rice. "A first-order perturbation analysis of crack trapping by arrays of obstacles." Journal of Applied Mechanics 56.4 (1989): 828-836.

[36] Ponson, L., 2009. Depinning transition in failure of inhomogeneous brittle materials. Physical Review Letters 103 (5), 055501

[37] Broberg, K. B. "Constant velocity crack propagation––dependence on remote load." International Journal of Solids and Structures 39.26 (2002): 6403-6410.

[38] Gdoutos, E. E. "Fracture Mechanics: An Introduction, 2nd Edn. (Vol. 123 of Solid Mechanics and Its Applications)." (2005).

[39] Anderson, Ted L., and T. L. Anderson. Fracture Mechanics: Fundamentals and Applications. CRC press, 2005.

[40] Chandra, D., and T. Krauthammer. "Strength enhancement in particulate solids under high loading rates." Earthquake Engineering & Structural Dynamics 24.12 (1995): 1609-1622.

[41]  S. Pagano, L. Fletcher, F. Pierron, L. Lamberson, “Image Based Transient Dynamic Fracture: Virtual Experiments and Practical Considerations,” International Journal of Fracture, in preparation (2017).  

 

 

 

Comments

D.Rittel's picture

Very nice and thorough review. Thanks, Leslie.

Leslie Lamberson's picture

Thank you so much Prof. Rittel, obviously your contributions in the field are seminal to this entry.  I look forward to seeing what dynamic fracture experiments and analysis will come in 2017!  

alifahem's picture

It's helpful analysis, Thank you  Dr. Leslie

D.Rittel's picture

This movie illustrates the subject of the blog....Dynamic fracture of a transparent ceramic bar....Enjoy! We have many more.

https://youtu.be/-xn4-EHQSxw

Subscribe to Comments for "Journal Club for December 2016:  Dynamic Fracture - when the going gets tough . . . "

Recent comments

More comments

Syndicate

Subscribe to Syndicate