The metalworking fluids market is $9-billion-plus segment of the specialty chemicals sector that includes cutting, forming, protecting and quenching fluids. One of the most important trends in cutting fluids is for products that work with multiple metals to reduce the number of fluids needed by a factory. Developing fluids suitable for multiple materials adds another layer of complexity to a design process that already involves selecting as many as 25 different additives and determining their optimal proportions. Houghton International used design of experiments (DoE) to develop — in half the time required by the company’s previous methods — a line of multi-metal cutting fluids that handle steel, cast iron, aluminum and cast aluminum.
The first stage in the development process for a new cutting fluid is to investigate individual lubricity additives that reduce friction between the cutting tool and workpiece. In this case, Houghton used one factor at a time (OFAT) studies to screen potential additives. We determined the performance of these additives by measuring the coefficient of friction between a test specimen and a pin with the additive applied. Comparing the measured coefficients of friction to a control value gave relative lubrication efficiency. Our chemists tested each selected lubricity additive on different metals and at different application conditions (Table 1). From the test results, we selected lubricity additives for the new formulations based on their ability to provide good lubricity on steel or aluminum or both.
Table 1. Comparison versus a control highlighted the dramatic differences in additives’ performance for steel and aluminum.
Typically, the lubricity package of a cutting fluid includes a number of additives. With many different possible additives to consider, Houghton chemists in the past relied on OFAT methods to identify the right formulation for a new product. This approach provides a quick way to get a rough comparative measurement of additive performance. The problem with OFAT at this stage is that it requires three or four repeats of tests or measurements to evaluate the effects of each additive. However, even after performing these tests, chemists are unsure if the difference in performance between various additives is statistically significant. They also have no way to understand possible interactions between the additives without running many more samples. As a result, this approach is time-consuming and leaves chemists uncertain about whether they have achieved the best formulation.
Houghton International has improved its abilities to develop cutting fluids and other new products over the past five years by applying DoE to its development process. In moving to DoE-based development, managers wanted to ensure that chemists would continue to spend nearly all their time on chemistry and not become bogged down or distracted dealing with statistical software. Houghton accomplished this goal by selecting Design-Expert from Stat-Ease, Inc. This statistical software is designed for people without any statistical background — yet still provides all the analytical power needed for designing specialty chemical products. We found it easy to train our chemists in its use and to convince them to use it. People in many of our global development centers now take advantage of the software. Stat-Ease provides all the different experimental designs, analytical methods and statistical reports that we need to develop successful products.
Figure 1. The lubrication efficiency of a component differs markedly depending upon the metal.
Choosing The Best Additives
After using the OFAT method to identify promising additives, we turned to mixture design and analysis tools to select the best four or five for more-comprehensive evaluation.
In developing the new multi-metal cutting fluid, Houghton chemists started by calling up the results of previous OFAT studies of ≈40 different additives. They selected the ten most promising ones and set up a mixture experiment in Design-Expert [1]. This DoE required 25 runs, far fewer than the OFAT method would need to evaluate the impact of each additive individually over a range of concentrations. The chemists performed lubricity tests of these mixtures on 1018 steel, 6061 aluminum and 319 cast aluminum. Multicomponent mixture design reduces the number of experiments required to develop formulations. All variables are included and simultaneously varied in the 25 runs. The analysis sorts out the individual effects of each variable as well as its statistical significance. The evaluation showed that some additives are good only on one type of metal while others are good on both (Figure 1).
Based on the mixture experiment, the chemists selected for further study four lubricity additives that provide the best overall performance. Determining this would have required at least twice as many samples using the OFAT method.
The chemists then performed a fractional factorial experiment on the four additives. They chose a template that would uncover interdependencies between variables. For example, it’s not uncommon for an additive to have only a minor impact on a property such as lubricity but for that impact to increase significantly when another additive is put into the mix. In-depth factorial experiments require additional runs but Houghton obtained the best of both worlds by using a mixture experiment to first narrow the field to just a few stellar additives; this enabled performing the follow-up factorial with only ten runs. The experimental results included a recommended optimized formulation that takes the direct impact of the additives as well as their interactions into account. It’s important to note that this optimized formulation was not one of the samples drawn up for the experiment but rather was calculated by Design-Expert based on the first- and second-order effects of the additives as determined over the course of the experiment (Figure 2). Houghton chemists created and tested the lubricity additive formulation recommended by statistical analysis; it delivered results close to those the DoE predicted. They made further modifications to the formulation to develop a final lubricity additive package that best meets the requirements of the market.
Figure 2. Fractional factorial analysis enabled determination of the best combination of four additives.
The chemists then used similar methods to select other additives, including emulsifiers, corrosion inhibitors and preservatives, for the multi-metal cutting fluid. Based on the results of these experiments, they developed a final formulation for the product platform that is optimized to deliver high levels of performance on steel, cast iron, aluminum and cast aluminum. In the fourth quarter of 2017, Houghton officially launched the HOCUT 8000 Series — state-of-the-art technology designed to improve customers’ operational efficiency, part quality and surface finish. These products have built-in detergency, run clean and provide excellent wetting that reduces consumption due to drag out, lowering customer fluid costs. In addition, the series has superior dispersibility, improving chip removal thereby minimizing downtime, excellent emulsion stability and low foaming characteristics, resulting in consistent lubrication and heat transfer performance. In addition, HOCUT 8000 products are extremely biostable, so they extend sump life and eliminate the need for costly sump-side additives. This new product line has been a success in the marketplace.
Houghton management estimates that developing the product line using conventional methods would have taken at least twice as long. Moreover, we would have had to be very lucky to attain the same levels of performance that DOE achieved.
DAVID SLINKMAN is vice president of global research for Houghton International, Valley Forge, Pa. YIXING (PHILIP) ZHAO is senior research scientist – innovation team leader for Houghton International in Valley Forge. Email them at [email protected] and [email protected].
REFERENCE
1. Mark J. Anderson, Patrick J. Whitcomb and Martin A. Bezener, “Formulation Simplified: Finding the Sweet Spot through Design and Analysis of Experiments with Mixtures,” Productivity Press, New York City (April 2018).
