Medical Material Choice / Treatment Options for Torsional Strength

[BiPolarMoment]

I have a screwdriver tip that consists of a .117 (3mm) hex with a .063 diameter hole located thru the axis of the driver. Because these are used in a clinical setting our material choices are limited by patient contact (biocompatibility) and corrosion resistance as they are put through repeated steam sterilization. Presently the screw driver is manufactured out of 455 stainless (ASTM A 564/A 564 M) in the H900 condition.

Unfortunately we are having drivers fail thru (primarily ductile) torsional shear at the interface with the screw--also unfortunately we cannot (yet) adequately or accurately quantify a maximum torque the driver will see during use due to multiple variables outside our control. What we do know is that the current use is outside the present capability of ~62 in-lbs (Ultimate strength of tested component) in some (not all) cases.

Finally, and most critically--increasing the drive size is nigh impossible due to the cost and logistical nightmare that would entail replacing (or reworking) all existing screws.

So, because we have no design criteria we're somewhat stuck with throwing ideas at the wall and since there is little we can do to increase the cross-section resisting the shear I'm hoping someone can point me towards a "silver bullet" material that I might not be aware of.

We have tested a 465 stainless (H900) bit and it was proportionally stronger (~68 in-lbs) but it was thought that increase wasn't likely (a guess) to be substantial enough.

The only other materials that have been on the table thus far but not yet tested:
SS440
X15TN (420 Mod)
Co35-Ni20-10Mo (Cold worked/aged) -- some concern on ability to gun-drill the .063 diameter

I know that tungsten carbide is also used in some surgical cutting instruments but I have no experience with its torsional strength.

But I could have overlooked some obvious materials or treatments that would be worth investigating. As previously mentioned, the restrictions are primarily that the material has to have some level of short term biocompatibility and that preferably it doesn't corrode in a steam environment--there's opportunity to create single use pre-sterile parts (TiN Coated Tool Steel?) if they would otherwise corrode but not at all preferable.

Thoughts?

 

[metman]

300M maraging steel will give you about 18% more strength and fair corrosion resistance.

Design for RELIABILITY, manufacturability, and maintainability

 

[Screwman1]

I worked on medical driver bits in the past and custom 455 was the material of choice. You are right that you are not going to get much increase from the available materials. Stay away from the Carbide material, it is very brittle and if you do have a failure you will be faced with the task of trying to track down multiple shards of carbide out of the incision.
Even using tool steel bits, the best that you can get is about Rc 62-63 before brittle failures crop up.
Can you reduce the ID hole at all or increase the AF dimension slightly? Anything that you can do to get a little bit more polar moment will help.

 

[BiPolarMoment]

Apparently I had some erroneous data from a coworker--we actually have tested data for 420 and 440A and 440C prototypes (and 17-4, but I think you'll understand why I didn't mention it before).

Previously I supplied the ultimate torque value but I also have the yield numbers now. What's interesting is that while the yield torque averages trend more-or-less expectedly to the material tensile yield, the ultimate wasn't quite as expected.

Material	Yield Torque	Ultimate Torque
17-4		45.3		52.9
455		56.5		62.3
465		61.5		68.1
420		55.2		83
440A		53.1		75
440C		58.6		76

It appears that 420--in whatever condition it was(I'd have to look it up)--seems to have substantial strain hardening capability that I wouldn't necessarily expect given the "book" typical UTS values.

While I don't know if 83 in-lbs is "enough," the large gap between the yield and ultimate is promising and may allow some "failures" that don't completely break and give us an opportunity to reinspect/replace as needed.

 

[Maui]

What hardness values were these materials each at to generate the values for yield and ultimate torque in your post above?

Maui

http://www.EngineeringMetallurgy.com

 

[EdStainless]

I am very surprised that 420 gave you that value. You need to look at it in more detail. The thermalmechanical history and properties would be very interesting.
The 465 is about as high as you will get.
There are higher strength materials (300M) but you will have much lower corrosion resistance, perhaps not enough for your application.
I have about the same experience as Screwman, at about 60RC or so brittle failure takes over in torsion.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[BiPolarMoment]

I also forgot to mention for all the above numbers we had n=6 for the sampling.

Ed,

I am very skeptical too--if we do consider looking down that path I will be making additional parts to confirm there wasn't some strange mix up in the testing (the parts do look the same after all).


Maui,

I'm not 100% sure, however, when the drawings were made by my coworker they were specced to a min. UTS (for the quench hardening grades) rather than a hardness and the PH steels were all H900 unless I'm mistaken. I'll have to get my coworker to dig up the paperwork when he's back in the office to see what the certs say. I also wonder if we still have the test parts--I could get the hardness checked to make sure it matches up with the certs.

 

[BiPolarMoment]

A little bit of an update. I finally had the opportunity to look into the material certs. It should be noted that apparently the heat treatment was not specified on the drawing, a minimum UTS was, which I think influenced how the thermal processor decided to go about it. Here was the result:


420:
Vacuum Heat @ 1900 deg. F for 1 hour
Temper @ 300 deg. F for 2 Hours in air
Tensile Yield: 195 ksi
UTS: 292 ksi

440A:
Vacuum Heat @ 1900 deg. F for 1 hour
Temper @ 300 deg. F for 2 Hours in air
Tensile Yield: 201 ksi
UTS: 266 ksi

440C:
Vacuum Heat @ 1900 deg. F for 1 hour
Temper @ 300 deg. F for 2 Hours in air
Tensile Yield: 202 ksi
UTS: 274 ksi

Note that in all cases the same 300 deg. F temper was used (presumably trying to maximize the UTS).

So the mechanical info actually does correllate pretty well with our torque data but that raises the question of why these results? Both the 440s seem to be weaker than expected and while the yield for the 420 looks about right, the UTS is off the charts.

Seems to me that if we even consider this route I need to get some more empirical data e.g. different material suppliers & tempers.

Any additional thoughts?

 

[Jplug7009]

Have you sought out higher Yield Strength 465SST, say in the 225ksi range... Made a bunch from that in the past.

 

[EdStainless]

Would you be allowed to use a maraging steel such as C300?
You need to find the chemistry of that 420, and see if you can buy more of it.
I have never seen 420 make those properties.

= = = = = = = = = = = = = = = = = = = =
Plymouth Tube

 

[BiPolarMoment]

Well according to the cert it should be capable of 247 ksi in the H1000 condition (we used H900) but we don't have tensile mechanical testing on any of the PH grades.

One thing I noticed that might be worth at least considering is the extremely thin margin for error on the PH grades yield/ultimate compared to the martensitic. Seems contrary to me that something with a larger elongation % would have a greater margin than the brittle 420/440s but that doesn't seem to be the case empirically.

 

[BiPolarMoment]

oops, I meant "thinner margin"

 

[BiPolarMoment]

Heat No. G11997 from Dunkirk Specialty Steel sold to Fry Steel -- from 2006.

I didn't look into the chemistry until you mentioned it Ed, but now that I do I noticed it has a fairly high carbon content (0.33) for what it's worth. I don't know how to interpret the impact of most of the other elements but the ones required by ASTM A276 are in spec.

I don't know much about maraging steels but we don't have a specific acceptance criteria for corrosion resistance--plenty of our instruments rust after they've been put through alkaline cleaners and sterilized repeatedly; it's unavoidable. What I would probably need to do though is a cytotoxicity test since it's not a commonly used material in invasive medical applications. It looks like the aging temperature of C300 is high enough to PVD coat it--can C300 be PVD coated? That would address both issues for the most part.

 

[BiPolarMoment]

Slight update: I finally received some test samples in 465 (H950) and two different treatments of x15tn. The preliminary results for the slightly more exotic heat treat on one x15 sample indicate a yield torque (2 degree offset) of ~75 in-lbs and ultimate at ~87 in-lbs. The less exotic heat treat on the x15 yielded slightly lower ultimate values but notably lower yield (62 & 85 in-lbs). Probably more importantly, like the 420 in the previous testing, the gaps between the yield and the ultimate are notable and compare favorably to the 465 which had nearly identical yield/ultimate torsion loading of ~66 in-lbs (the stress/strain curve is remarkably flat after yield).

So, I'm not sure if there is some aspect of these alloys that work harden very favorably in torsion--it's very interesting and gives some credence to that original 420 testing that netted values those "unusual" values of 55/83 in-lbs.

Nothing beats empirical testing.

 

[tbuelna]

BiPolarMoment-

In your OP you mentioned the 3mm hex screwdriver tip was failing at the point of contact with the socket in the screw head, due to what appeared to be excessive torsional shear stress. Besides the change in material/heat treat you are considering, I would also suggest looking at changes to the screwdriver itself. One suggestion would be to change the cross section of the tip and screwdriver shaft to eliminate any abrupt changes in torsional stiffness that would create stress concentrations. Another suggestion would be to add a slight lengthwise profile crown (around .003"-.005" of crown drop) to the flanks of the hex that contact the screw socket faces. This crowning will help mitigate the effect of misalignment between the screwdriver tip and screw socket during installation. One final suggestion would be to add some form of torque limiting device to the screwdriver to prevent a shear failure in the tip.

Good luck to you.
Terry