Heat Treatment - Crystal Weaving Foundation

I re-hted 3 D6 blades with hardness peak at 69-69.25rc and tempered down to 67.5rc. Whittled hardwoods: passed. whittled bone: barely visible micro chips. Opened hole in a can: 2x magnitude compare to niolox above. Tested 8" chopping blade 16dps, 0.020" BET - passed on hardwoods, passed on bone using good chopping technique. Side slap-chop(accidental actually) on bone = loss a big chunk of edge (7/16" wide x 3/16" tall, 0.038" thick).

This is the first time I encounter a carbide this big. Grain diameter looks nifty good.

** updated pic
5i6nAqN.png
 
Last edited:
Whoa! Such a huge carbide!

I was thinking CWF process prevents such grow? More tweak needed?

From your previous calculation sheet, it seems Superblue is close. Is that why some people found it chip prone (I recall Jason B said so)
 
Last edited:
Those are primary carbides, it would take either close-to-melting heat or quite high temp + very long soak to dissolve them. You can see a lot of fine carbides, those are ht controls. I use D6 as easiest way to detect ht properties changes/improvement. D6 ht 2.5 result is good - even when cut a hole in the can. As mentioned - I've a 8" D6 blade (unfortunately partial tang) for impact/chop test, so far is 'too good' for D6. I think, best to leverage D2 & D6 large carbides as buzz-saw edge rather than trying to make uniform carbides (best just buy CPM D2 or use CTS-XHP). Right now, D2 is more balanced than D6.

Whoa! Such a huge carbide!

I was thinking CWF process prevents such grow? More tweak needed?

Just finished re-ht 2.5 3x niolox thin bushcraft - 66rc. Rough sharpening (120g belt + 240 SiC + 1K stone) 14dps, 0.005"BET EG(edge geometry) - cleaned & whittled frozen cooked beef rib bone: some micro ripples and chips (under 30x loupe). This ht 2.5 params for niolox are ready for a large batch of kitchen knives, 2/3 are pre-allocated.
 
Thanks very much for your trust - indeed a heck lot of it, since you didn't even see profile of these knives. I reserved a spot for you, please review knives profile and decide if it/they is/are good match for your need.

7.5" in 3.2 mm and 2.2 mm
5" in 3.2 mm and 2.2 mm
4.25" in 2.2 mm

Edit: found extra 2.2mm blanks 8.5 full tang western handle gyuto, 5" full tang western handle petty, 4.25" full tang paring.

All will grind very thin FFG with distal taper. 7.5" edge might or might not has gradual convex for food-release above sharpening bevel. Stating this because I prefer laser and ok with some stiction.

* sorry about my last batch of s90v kitchen knives - they were hardened and ground but might go from collecting dust to recycle-bin since they are too thin to re-ht 2.5.

UmARuPv.jpg

Pre-orders are possible? Please count me in.
 
Last edited:
What is a stable/great/cool Edge? One's performance on indended tasks with less degradation than projected/anticipated.

Edge stability closedly ties to scope/size of yield at the edge. Yield Scope rough lists in increasing observable sizes:
Crystal blocks, single carbide interface/boundary, cluster carbides interfaces, individual inter-grain, single grain, grain cluster,..., global.

*crystal & interface yield can be either plastic and or fracture.

Rather than writing a thesis, I only discuss one particular instance pic-below. Further discussions/deductions/.. if/when interested.

pwqxw5M.jpg


Resulted edge (smoothly slice and somewhat push-cut phonebook paper) - estimated apex width is about ~3 to 15 microns (crest to trough). Yields occurred at single & cluster carbides and possibily single grain. So 10V 69rc with 15dps for tasks involve moderate lateral and hard compressive cuttings, I think a 15dps 10-15um wide apex would be an excellent stable working edge.

A good stable working edge (given appropriate strength & bevel angle) has radius relative to carbide diameter:
For steels with very hard carbides - 2 to 5x
For steels with moderate (e.g. CrC) carbides - 1 to 3x
For steels with no or easy to shape carbides - 0 to 2x
 
Last edited:
Luong,

D6 chop is crazy! There might be damage, but the 0.5mm carbide is not the culprit. The HT somehow manage to contain it :thumbsup:
The small WA / flat handle (from post#586)looks interesting .. It'll be quite similar to a Victorinox paring (hint hint, 3" blade, 3-3.5" flat handle) ;)
 
Last edited:
Maybe there weren't many 50+um carbides near edge impact areas - idk, nevertheless D6 edge performed very well compare to any steels with similar edge geometry. Its hrc got lowered to 67.5 just as a fast way to go past-peak-hardness ease on matrix stress/dislocation. Where optimal hrc is just a notch below peak hardness - at this point, the matrix is barely into tempered-martensite with minimal precipitation.

Luong,

D6 chop is crazy! There might be damage, but the 0.5mm carbide is not the culprit. The HT somehow manage to contain it :thumbsup:
The small WA / flat handle (from post#586)looks interesting .. It'll be quite similar to a Victorinox paring (hint hint, 3" blade, 3-3.5" flat handle) ;)

** I updated a few micrographs with corrected scale bar for a few previous posts. **
 
Last edited:
I took 2 days break from making niolox knives, to test HT 2.5 for W2, 52100 & Cfv. With current ht params, 52100 and cfv max hrc around 65.5. I will micrograph these test blades later...

Cut a hole on the side of a 14oz curry can:
W2 66.5rc = rolled, ripples and micro chips.
52100 64.5rc = rolled and 1 micro chip. This edge chopped dried beef rib bone fine (no visible or by touch damage).

Compare this result with steels with high alloy (this and last page), it would be apparent - low alloy steels have higher edge plasticity/toughness in 64.5-67rc hardness range.

8O4rNA2.jpg
 
Compare this result with steels with high alloy (this and last page), it would be apparent - low alloy steels have higher edge plasticity/toughness in 64.5-67rc hardness range.
This bears the question, which is better for general continued use without sharpening, and which would be faster to repair back to an optimal edge with best sharpener.
 
At this time, I think - niolox is well balance in combination of excellent: wear resistance, toughness, ease of sharpening and afaic stainless. You can sharpen/shape Niolox edge with regular alumina waterstone to produce edge from coarse aggressive to dry shaving keenness.

Dig a little deeper...

Your question covers a wide scope/applications. Let's remove a couple variables:
  1. Blade has cross section similar to box cutter - keep sharpening bevel constant.
  2. Sharpening abrade rate is steady - perhaps, using a full line of waterstone with high density diamond or cbn abrasive. Also along with adaptive pressure to keep abrading rate almost constant.
Scenario#1: edge loss by dulling (slow wear - emperical assumption) with 18-20dps edge geometry.
= apex width of 10v, 15v will narrower than steels with less carbide shielding. Hence quicker to restore edge.

Scenario#2: edge loss by dulling, with 15-17dps edge geometry.
= m4/cruwear/m2/niolox

Scenario#3: edge loss by dulling, with 10-14dps edge geometry.
= w2/52100/cfv/m2/niolox

Scenario#4: edge loss by rolled/rippled/chipped, with 18-20dps edge geometry.
= m4/cruwear/m2/niolox - further assumption magnitude of each step of loss is about the size of carbide diameter.

Scenario#5: edge loss by rolled/rippled/chipped, with 10-17dps edge geometry.
= w2/52100 small cementite (~200-4o0nm dia) translate to smaller failure unit (combination of: carbide interface, grain, grain cluster).

This bears the question, which is better for general continued use without sharpening, and which would be faster to repair back to an optimal edge with best sharpener.
 
Let's take all the steels you've dealt with, all ht's, sharpened 15-17dps to about 400-600 diamond grit (medium-fine). Do this with them.
Cut a hole on the side of a 14oz curry can
.
1) Which ones would still have a useful edge where the cutting took place.
2) which ones would be trashed and require sharpening before further use. (I consider rolled and flattened edges useless).
3) Which ones suffer the deepest edge damage, whether roll, chip, or flattening.
 
Inline answers

I did a few more cuts with w2/52100/cfv. Looking straight down on apex afterward - 52100's apex width is over 3 times wider than w2 & cfv. otoh, 52100 micro chip cross section is about 75% and 60% thick relative to w2 and cfv.

Let's take all the steels you've dealt with, all ht's, sharpened 15-17dps to about 400-600 diamond grit (medium-fine). Do this with them.
.
1) Which ones would still have a useful edge where the cutting took place.
W2@66rc, Cfv@65rc, M2@65rc, M4@65rc, Cruwear@65rc, Niolox@64rc, Aebl@64rc, ...

2) which ones would be trashed and require sharpening before further use. (I consider rolled and flattened edges useless).
I agree with you - any edge that rolled or flattened, which essentially turn edge into a thick wire would indeed ineffective for cutting. Roll/flatten at least quadruple the apex radius plus smoothing from bend - this condition is very poor for cutting/slicing. As long as 50% of edge isn't chipped, it is usable even when chips are large. Depth of rolled/flattened (damage to thickness) usually 30-60% less than chip, so it is a trade between blade consumption vs continue in service at greater cost. Unless rolled/flattened is a spec, I prefer blades (any steel) with hardness 63+rc.

3) Which ones suffer the deepest edge damage, whether roll, chip, or flattening.
D6 at any hardness - expect big chip at 50+rc; roll&chip below 50rc. Rex121/15V @70-72rc. Chips cross section can be in 200-500um. Whereas compare to more ductile steel+ht roll/flatten - damage cross section mostly be less than 150um.
 
m2 and m4 no surprise, been using this class of knife steel since mid 80's.
Cruwear no surprise, it's equivalent to my zwear blade from you.
Niolox interesting composition, no real shocker.
I can somewhat understand cfv and w2. My w2 from you performs well but figured it might flatten/dull a little due to low alloy.

What I can't understand is aebl, lowest c, highest cr and that's about it except for low mn and s. Seems like the simplest of stainless steels. I understand alloy isn't everything. What % cr is carbide and free? What is it that puts it in the same category as the others with this test... being so different.

imo, cutting that can tests adhesive wear, strength, and toughness all in one whack, and one of the few tests I can do at home to determin how it will do at work.
 
* a quick clarification of 'flatten' term: rapid edge/apex broadening beyond normal wear. Similar to 'smush' condition but less severe.

The BCMW zwear blade, you own, has early experimental ht (around ver 0.5 relative to current versioning). It has good strength and grain size but matrix lacked of elasticity (thus poor lateral toughness). Your 65rc W2 blade has ht 1.0 (iirc), where edge would micro ripple & macro chip. In compare to 67rc w2 with ht 2.5 with similar edge behaviors - perf different is quite large because last few percent, toward 99%, can and most likely contribute 20-50% gain in performance. There few percent are weakest links of the microstructure chain - I think, ht 2.5 strengthened these weak links however they are still weaker than the rest.

Aebl composition: 0.67%C, 13.00%Cr, ...
* Please feel free and I would appreciate any correction/augment/dispute/comment at my calculated figures relative to my ht.

0.57%C use by mart matrix; 1.04%Cr binds in Cr carbide form (in 95% | 5% ratio between M7C3 and M23C6. 11.96% free Cr);

btw - using same calculation hardened niolox has 12.7% of free Cr, so it is a tad more corrosion resistance than hardened aebl.

Aebl (and other high cr% with less than 0.7%C) has carbide with diameter pointy-bell-curve around 300nm, which translate to smaller unit of failure when yielded. 64rc matrix would micro ripple and macro chip - leaving edge usable after cutting a round hole in a can. 65rc would works too - likely be smaller ripple and macro chip OR just ripple because of extra strength. At hardness 62r or lower, edge would bend into roll, then subject to greater cutting pressure (became equiv to cut using a round wire) lead to much larger rolled = no longer functional edge.

* my perspective: edge behavior preference is subjective on individual's usage+style+perception... the best one is the one works well/best for you!

In survival situations (outdoor, combat, ..) - a rolled edge may fail to cut and slice functions (sure - no problem with stab/dig/...) vs ripple+micro-chip edge would retained most of cut&slice functions. Cheaper to grab/buy another knife(after X numbers of recurred adventures) than spending 9 lives :p

m2 and m4 no surprise, been using this class of knife steel since mid 80's.
Cruwear no surprise, it's equivalent to my zwear blade from you.
Niolox interesting composition, no real shocker.
I can somewhat understand cfv and w2. My w2 from you performs well but figured it might flatten/dull a little due to low alloy.

What I can't understand is aebl, lowest c, highest cr and that's about it except for low mn and s. Seems like the simplest of stainless steels. I understand alloy isn't everything. What % cr is carbide and free? What is it that puts it in the same category as the others with this test... being so different.

imo, cutting that can tests adhesive wear, strength, and toughness all in one whack, and one of the few tests I can do at home to determin how it will do at work.
 
Last edited:
Scenario#1: edge loss by dulling (slow wear - emperical assumption) with 18-20dps edge geometry.
= apex width of 10v, 15v will narrower than steels with less carbide shielding. Hence quicker to restore edge.

Scenario#2: edge loss by dulling, with 15-17dps edge geometry.
= m4/cruwear/m2/niolox
Looking at this a different way - can I infer that 10v would ideally be sharped to about 18 dps (or not less than 18dps)?
And it seems the low carbide steels are able to handle much lower edge geometry? So If I like to go below 15dps, I should be looking for W2 or CFV?

I also made a mental note that niolox shows up on those lists a lot:thumbsup:. Starting to see why you like it so much.
 
Excellent views/takes!

In general, higher alloy = higher compressive and bending strength but compromises on less elasticity & drastically lowered plasticity range. For 10v - yes, ideal cutting edge needs steel volume from 18+dps angle. Due to difficulty and amt of effort of shaping carbides on apex, best result by shape micro 18+dps first then carefully thin behind it with main bevel. For a long while now, I think - most people (including me in 95% occasions due to laziness and lacked of need) done this non-optimal (backward) way: doing main bevel, then add micro. Think, when apexed 15dps first, less stiffness to shape carbides on apex, therefore 18+dps micro applies to a weaker apex, than 18+dps first where it flex-away less because stiffer.

I don't think, there is a big edge cutting effort differences between (dps/micro) 12, 12/15, 15, 15/18, 15/20 when micro is facet length is less than 30-40um wide. Choices would depend on amt of elasticity and failure-unit you prefer. W2/52100 carbide/cementite is very fine, correspond to finer granularity of failure unit.

Currently niolox is on top of my list because what I mentioned before. Multilayer matrix protection/shield is its key/fundamental reason for choosing it. Many other steels also have multi-layer carbides to protect the matrix however traded on wear | corrosion | tensile | ...

Niolox
0.9%V (carbide less than 1um) provides shield against large wear unit (e.g. sand, iron, ceramic, etc..), 0.7%Nb(sub 200nm) shields smaller wear unit (e.g. silicates, iron, ceramic, etc..) and at the same time slow down matrix loss around Vanadium Carbide footing.

Worth repeating on carbide count/number by compare niolox vs 3v (with 0.8%C and similar total alloy volume). For 1(count) 3V 2 micron diameter carbide, niolox has 4/count 1um vc and 500/count 0.2um NbC. So when facing 0.01-0.05um silicates, niolox has order of magnitude more shielding than 3V. This is why niolox has effective wear resistant as D2.

Niolox won't reach wear level when compare at working/sustain apex width of maxamet. otoh, maxamet wear will be far short of niolox working/sustain apex width.

Envision sand blasting/flowing directly and diff angles at the edge is a good way to understand role of carbide size + shielding. Easy to see when grit is less than 100nm and direct flow is into edge (push cut | chop), steels with coarser (2+um) carbides have low % of shielding at various depth or continuous erosion because carbide underfooting (matrix) worn away by grits. For given same percent of exposed (unprotected) matrix, loss rate is much faster for coarse carbide steels since each carbide fall off = 2-10 magnitude of finer carbide steels. An exception when shielding exceed certain %. e.g. WC x%Co (x less than 20) grit flows may not wear away enough footing of WC grain to undermine its binding in Co matrix.

Looking at this a different way - can I infer that 10v would ideally be sharped to about 18 dps (or not less than 18dps)?
And it seems the low carbide steels are able to handle much lower edge geometry? So If I like to go below 15dps, I should be looking for W2 or CFV?

I also made a mental note that niolox shows up on those lists a lot:thumbsup:. Starting to see why you like it so much.
 
How's about shielding gaps in matrix (grain boundaries, between mart blocks, carbide interfaces)?

Carbide provides this 3rd matrix shielding layer is super small. My ht goes after eta-carbide, see pg 82
https://www.phase-trans.msm.cam.ac.uk/2004/z/3750-003.pdf

I simply refer to hexagonal structure carbides as hcp reflects its structure.

It required combination of sophisticated instruments for detect/verify hcp presents in matrix, thus beyond my access. Nevertheless, I can infer (indirect method) by varying the cryogenic dip vs long cryo soak vs BEC. In summary, test coupons with BEC (simulated approaching to Zero Kelvin quench) are about 0.5-0.7rc higher than other cryo steps = thereby 'Inference' gaps in BEC matrix are less porous and gaps boundaries contain less retained austenite (more ra to mart conversion). Oh well, this inference could turn out be a wag or good result by wrong reason - either way, I am :cool: with that.

And when tempering surpassed activation point, carbon detached from martensite matrix - it precipitates in various transitional structures for range of temperatures, much depend on dislocation and alloy elements. ferrites are byproduct of tempering process. Size of ferrite unit depend of carbon movement and whether still tether to some Fe or latch onto alloy. These ferrites provide ductitility however it look like a quite lop-sided trade of ductility while increase soft/weak points in the matrix. So, question is - what percent of iron on apex of your blade, you are willing to accept? Worse yet - majority of ferrites will be locate in grain boundaries and carbide interfaces. Carbides in excess tempered matrix - some may look like #2 then turn into #3 with wear :p

http://i1.wp.com/www.mcomiedentistr...iodontitis-stages-mcomie-family-dentistry.jpg

Many of you probably well aware of keen edge low performance when matrix in these condition - how's about going down hill starting around 62rc.
 
Compiled and calculated Peak HRC of BCMW HT 2.5 for a few steels. Coarse carbide network could negatively increase/decrease HRC. Increase hcp volume would positively increase HRC.

* uElem = estimated unbound/free elements volume (V,W,Mo,Si,P, etc.). Balance elems are locked in mart matrix; ferrite; carbide and ra.

AlSNxbY.png
 
Back
Top