A SYSTEMS APPROACH FOR ACHIEVING STRESS FREE PARTS IN HIGH STRENGTH ALUMINUM ALLOYS ©

BY TOM CROUCHER

MARCH 2011

P.O. Box 6437 M Norco, CA 92860 M (888) 502-8488

Email: [email protected] -1-

© COPYRIGHT 2011 by Tom Croucher All rights reserved The use of this document is limited to the party to which it was directly sent. No part of this document may be reproduced, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise for any other purpose without the prior written permission of the copyright owner. However, if this article is to be used for solving residual stress or quenching problems or for training purposes, permission will be readily granted by calling the author at 888-502-8488 or by email at “[email protected]”.

-2-

BACKGROUND AND INTRODUCTION Residual Stresses that can cause problems with dimensional stability in high strength aluminum alloys have been a constant problem in the aerospace industry for many years. This problem reached its peak in the 1960's and early 1970's. The results were: (1) premature failure of parts in service, including both stress corrosion cracking and early fatigue failures and (2) unwanted part movement - both in service and during final machining to meet required dimensional tolerances. Unfortunately, the problem continues today as (1) the aerospace and aluminum industries push the envelope toward bigger aircraft requiring larger and larger high strength aluminum components and (2) extremely tight dimensional stability and tolerances are required in optical components such as space mirrors and telescope parts. This article presents, in detail, a systems approach which we have developed over the past 50 years. We have consistently proven that, when applied correctly, any high strength aluminum alloy part can be produced with minimum residual stresses while at the same time achieving all the structural properties desired. It is based on my 50 years experience in combating these problems first as a senior lead metallurgist at a major aerospace prime, then later directing a commercial aluminum heat treating company whose primary mission was the production of distortion free parts, and finally, during an extensive consulting career, in assisting many who were faced with these same types of distortion problems in high strength aluminum alloys. Through the years, I have had numerous articles published on the subject, but to date have never presented in specific detail, the final systems procedure that we use to accomplish the results that we do. This is the purpose of this paper - to outline in detail the step by step procedure that is necessary to produce stress free, dimensionally stable, high strength aluminum alloy parts. In order to understand the basis for our approach, it is necessary to grasp a few fundamental concepts. 1)

It must be understood that all residual stresses in aluminum alloys are the result of unequal expansion and contraction in a part during processing. This is true whether these stresses result from mechanical or thermal means.

-3-

2)

It must be realized that 90% of our existing dimensional instability problems in high strength aluminum alloys result from residual stresses that are imparted during the quenching operation while being solution heat treated.

3

That the quenching process is a freezing process that attempts to freeze the hardening atoms that have been positioned during the solution heat treating process, in place. To accomplish this freezing process, the cooling rates must be fast enough to prevent significant diffusion during cooling.

4)

During the cooling process, the aluminum is not smart enough to know what it is being cooled by. It does not understand the difference between cold water, hot water, glycol or a spray. It only understands that it is being cooled at some rate which may or may not allow diffusion to take place.

5)

Over the past 50 years, the emergence of less quench sensitive alloys and of newer quenching fluids and methods, mainly the polyalkylene glycols, allow for the first time a truly engineered approach to solving problems of troublesome residual stress.

6)

The proper application of a cryogenic stress relief method of stress relieving “as quenched” parts by a thermo-mechanical technique (commonly called uphill quenching - developed by Alcoa in the late 1950's), allows for an effective stress relief technique to be employed to relieve or reduce these high quenching stresses on finished complex parts.

As I near retirement, it is hoped that my 50 year experience may help those newcomers who are now starting to face their upcoming problems. This current article outlines in detail the current process that we use to produce stress free parts. A previous article which summarizes my experience and provides a technical synopsis of the problem can be viewed and/or printed in pdf format by clicking on the link below to my website. There also appears on my website, at http://www.croucher.us number of other articles on the subject.

-4-

Key Words: aluminum; heat treating; cryogenic; distortion; glycol; quenching; heat treatment; heat treating; liquid nitrogen; polyalkylene glycol; residual stress; solution treating; solution heat treating; stresses; stress relief; stress relieving; machining stresses, uphill quenching; warpage; water quenching; compression; dimensional stability; Ucon Quenchant;

-5-

The following is a step by step outline of our process for producing stress free parts. 1)

Before starting any solution heat treatment, the part must be as free as possible from any prior residual stress, either from prior heat treatment, forming, casting or machining operations. If residual stresses are suspect, the part should be annealed or fully stress relieved adequately, and then check and straightened to bring the part back into dimensional tolerance. Residual stresses present in a part that is loaded into a solution heat treating furnace can be relieved during the heating process, and cause the part to distort in the furnace. This results in a distorted part after quenching, but it needs to be recognized that the distortion was not caused by the quenching process but rather by prior induced stresses. In some instances, this type distortion can scrap a fully machined part. If the part is to be salvaged, it will probably need additional check and straightening operations which will also induce additional stress to the part. If any straightening process to achieve dimensional tolerance has been excessive, the part should be given an additional stress relief and checked a second time.

2)

Have a part produced to as close as possible to final dimension before solution heat treatment. The thicker the part, the higher the stress that will be imparted upon quenching. As residual stresses are almost exponentially a function of thickness during the quench, the lowest stress levels are obtained if the section thickness is reduced. However, with parts of large section transitions, sometimes it is better to leave some excess material on the thinner sections to attempt to achieve equal cooling in all areas of the part and machine if off later.

3)

Avoid imparting machining stresses into the part. Be sure that machining is performed in a manner to eliminate induced stresses. Adequate cooling and proper selection of cutters, feeds and speeds is necessary to avoid any heat build up in the part while machining. High speed machining can remove a lot of material but if not done properly, can impart huge stresses into the material.

4)

Do not put stress in the parts during the quench in the first place. This process involves an understanding of the heat -6-

treating process and the effects of different quenching procedures on developing adequate properties in alloys of given chemical compositions, and different thickness. Most of this information is already available for common alloys, although for some of the newer alloys, particularly with those of zirconium additions, quench sensitivity data needs to be developed. If it is necessary to develop a new quench sensitivity curve, use the five or seven bar tensile test procedure shown in References [1] and [2]. 5)

Define the alloy, i.e. 7075 or 7050.

6)

Determine the minimum tensile strength that is required in the part.

7)

Determine the maximum thickness of the part during solution heat treating and quenching. If possible, follow step (2) and machine the part to as close to final dimension as possible.

8)

Obtain the alloy’s quench sensitivity curve. These have been published for most alloys and many of them that we use are shown in Appendix #1 which is attached. For newer alloys, a simple method shown in References [1] and [2] involving the testing of as little as 5-7 tensile test bars can be used to develop an adequate curve.

9)

From the quench sensitivity curve, determine the quenching rate that is necessary to achieve the desired strength properties from the quench sensitivity curve.

10)

Before solution heat treating, a detailed racking plan and immersion procedure needs to be developed and put into place that will ensure even cooling in all areas of the part during quenching. How a part is racked and supported is critical to achieving a uniform distribution of stress throughout the part during cooling. The part must be allowed to expand and contract freely and be quenched effectively without any restrictions. An inadequate racking procedure, an improper orientation or an incorrect immersion rate of the part into the fluid both can result in unequal stresses developed in different areas of the part. Immersing the part at an incorrect angle or immersion rate can result in differential stresses from the first area that hit the fluid -7-

from the last area being immersed, particularly when using faster quenching rates. 11)

When solution heat treating, select a quenchant that will achieve low residual stresses. Ideally, you want to use a product that will cool as slow as possible to achieve low stresses, but fast enough to achieve all required properties. Using available charts, find a quenchant that will repeatably produce that quench rate for the part thickness determined in Step (7). Avoid water, prefer glycol or spray if possible. Boiling water quench is sometimes acceptable if desired mechanical properties can be met. Some of the charts we use are shown in Appendix B.

12)

Solution heat treat and quench the parts.

13)

If there is to be any delay after the quenching operation before stress analysis or check and straightening operations, the part must be refrigerated immediately after quenching at -10°F or lower. Allowing the part to naturally age for any length of time can reduce the affect of any stress relieving operation and will induce additional stresses if a later straightening operation is required.

14)

After quenching, measure the residual stresses by an effective technique (we use x-rays) to determine the level of stresses developed by the quenching process are acceptable and if additional stress relief is warranted.

15)

If necessary stress relieve after quenching either by a. mechanical if effective on raw material b. uphill quench if necessary on parts

16)

Apply the stress relieving process if required. In many instances, if a slow enough quenchant is used, low residual stresses will result and additional stress relieving will not be necessary. In fact, applying an arbitrary mechanical or cryogenic stress relieving process to a part that has developed a low stress level as a result of a proper quenching process can actually induce unwanted stresses. If stress relieving is required, consider uphilling with high velocity steam. Avoid using hot water as the uphill media, as it is only partially effective. If it has been proven that hot water will reduce the stresses sufficient for -8-

an individual case, the process might be acceptable. 17)

Measure the resulting stresses after the stress relieving process. If the level of stress is acceptable, go directly to aging the part. If the stress level is unacceptable, uphill one more time, again measure stresses and then go to age. Lately we have found, that in contrast to earlier work, that sometimes more than one uphill procedure may be necessary. This is especially true with extremely thick parts.

18)

Lock in the process in detail with a fixed plan, specification or drawing note. The plan must be extremely detailed. Just specifying a normal heat treating or quenching requirements is not adequate for achieving repeatable low stress parts. Quenchants types, temperature tolerances, concentration level, racking method and quench orientation, possible immersion rates must be specified.

19)

Once the process is working and is locked into place, do not allow anyone to make even the slightest change without a complete review by knowledgeable personnel especially purchasing agents who are trying save a little money.

Summary: Understanding the primary causes for residual stress in high strength aluminum alloys is paramount to achieving success in producing stress free parts. This includes an understanding of the quenching process and the means by which parts can be effectively quenched to produce low stress levels. It also includes an understanding of the different methods of stress relieving, mechanical, cryogenic, and their characteristics, advantages and disadvantages. With this understanding and a proper application of the principles advocated here, any aluminum part can be easily produced essentially stress free while at the same time meeting all required properties.

-9-

References:

[1] T.R. Croucher; "Critical Parameters for Evaluating Polymer Quenching of Aluminum," Heat Treating, Vol XIX, No. 12, December 1987; pp 21-25. [2] Ed Blalock and Tom Croucher; A Proposed Method For Selecting Polymer Quenchant for New Aluminum Alloy Applications, Tom Croucher and Associates; Report AMR 87-202. October 1987.

-10-

Appendix A Compilation of Quench Sensitivity Curves Collected From the Literature

-11-

Figure A1 - Quench Sensitivity Curves According to Hunsicker [1]

Figure A2 - Three Point Quench Sensitivity Curves From Vruggink, Alcoa Data; Unknown Publication. [2]

-12-

Figure A3 - Quench Sensitivity Curves Published by Mackenzie [3]

Figure A4 - Quench Sensitivity Curves For Different Alloys and Tempers [4]

-13-

Figure A5 - 7075-T6 Quench Sensitivity Curve From Croucher Using Different Glycol Concentrations [5]

Figure A6 - Croucher Schematic Showing the Relationship of Quench Sensitivity in Two Areas. -14-

FigureA7 - Quench Sensitivity Curves for 2014 and 7075 Alloys Showing Effects on Tensile Strength and Corrosion Resistance According to Hunsicker [1] References [1] H.Y. Hunsicker, "The Metallurgy of Heat Treatment"; Aluminum, Vol I, p135; Published by American Society for Metals, 1967. [2] J.E. Vruggink, "Quenching Rate Effects on Mechanical Properties of Heat Treatable Aluminum Alloys," Aluminum Company of America Report, New Kensington, Pa. [3] Scott MacKenzie; “Design of Quench Systems for Aluminum Heat Treating Part I, Quenchant Selection;” Industrial Heating; June 14, 2006. [4] P. Kavalco, L Canale and G Totten; Quenching of Aluminum Alloys: Property Prediction by Quench Factor Analysis; Heat Treating Progress; May/June 2009. [5] Unpublished data by Tom Croucher

-15-

Appendix B Compilation of Cooling Rate Charts of Different Quenchants Collected From the Literature

-16-

Figure B1- Cooling Rate Chart According to Hunsicker [1]

Figure B2 - Effect of Different Quenchants on the Cooling Rates of .040-Sheet - After Lauderdale [2]

-17-

Figure B3 - Compilation of Lauchner’s Cooling Rate Data For Glycol Quenching of Sheet Metal [3].

Figure B4 - Croucher Re-Plot of Lauchner’s Data From Figure B-3. -18-

Figure B5 - Croucher Plots of Cooling Rate Data For AMS 3025 Required Thicknesses.

Figure B6 - Original Croucher Plots - Effect of Glycol on Cooling Rates of Thin Plate Material.

-19-

Figure B7 - Original Croucher Cooling Rate Data For Plates Plotted According to Hunsicker Method. [4]

Figure B8 - Original Croucher Rate Data For Plates Up to 3-inches Thick. [5]

Glycol Cooling

-20-

Figure B9 - Cooling Performance of Water at Four Different Temperatures Compared to Two Glycol Concentrations Plotted According to Hunsicker Method [6].

Figure B10 - Croucher Semi-Log Plots of Glycol Cooling Rate Data Up to 6-inch Thick.

-21-

References: [1] H. Y. Hunsicker, "Chapter 5, The Metallurgy of Heat Treatment" Aluminum Vol 1; American Society for Metals; 1967, pp 136-137. [2] Lauderdale, R.H., "A New Quenchant for Thin Gage Aluminum;" Metal Progress, December 1967. [3] Lauchner, E. A. and Smith, B. O., "Evaluation of Ucon® Quenching," NOR 69-65, Northrop Corporation, May 1969. [4] Tom Croucher; "Applying Synthetic Quenchants to High Strength Alloy Heat Treatment" Presented October 1970 at the ASM National Metal Congress, Cleveland, Ohio; ASM, Metals Engineering Quarterly May 1971. [5] Tom Croucher; “Polymer Quenching, Their Advantage For Quenching Aluminum Alloys, Heat Treating”; November 1982. Volume XIV, No.11, pages 18-19 [6] Tom Croucher; “Effectively Quenching Thick Sections of High Strength Aluminum Alloys Using Polyalkylene Glycol Quenchants” Presented to SAE Amec, October 2009. Available at www.croucher.us.

-22-