-
Experimental Design
I am currently reading about the advantages of experimental design. I found a couple of articles that discuss those methods and they seem highly efficient especially when there is a synergy or interaction between the ingredients (such as hair gels, and surfactants). Does anyone here had already used such designs, and could you please share with us your opinion and if it worth investing in such techniques.
-
14 Replies
-
I went through a three day course on DOE a number of years ago but the most useful of these types of experiments I found for formulating was the knockout experiment.
The problem with using formal DOE with cosmetic formulas is that there is rarely a good test to measure significant differences. For food products you can do taste tests which are decent. Most people can tell differences in taste and tell you what they like.
However, cosmetics are more difficult. Measuring things like moisturization is very difficult to be consistent. Same with other things like measuring foam quality, anti frizz, color fastness, etc.
It’s all very subjective and I found that not having an easy way to quantify subtle differences between prototypes is a real obstacle.
-
I did some work with DOE a few years ago when I was formulating floor polishes. As Perry mentioned it works better when you have an easy to quantify goal. In my case it was a gloss reading for the floor polish.
For cosmetics I don’t use the DOE methods formally but can see how a background in this helps with working out what to change in a formula when you are trying to optimise your formula.
-
Thank you for your fast replies, and I agree with you that from a sensorial perspective it is better to use the knockout experiment. But I was looking from another perspective when I was researching the DOE, I was hoping that with this method I can increase the stability of emulsions, or obtain the maximum viscosity for minimum viscosity agents when multiple viscosity agents are present in a formulation, which are very quantifiable parameters.
-
The problem with DoE apply to formulation, is that
response are not predictable, since the phase behavior is not lineal, and show
catastrophic change with very narrow change in composition. Phase diagram is
the only approach that can give you a full picture of a cosmetic system… unfortunately
it is very time consuming. -
I’m not trying to dissuade you and I believe it’s worth trying, especially for a characteristic like viscosity.
However, cosmetic formulations are very complicated (on a micro scale) such that you might create exactly the same formula and have a viscosity that ranges from 10,000 - 15,000. It may require you to run multiple samples of the same thing to get the proper variance and deviation. Even then you might not get a precise enough measure to be useful.
-
yet another reason doe is difficult with cosmetic formulas is that most cosmetic ingredients are multifunctional. although it might be possible to optimize viscosity, other properties such as skin feel, texture, color etc. need to be kept constant (or be measurable) in order to achieve a useful result.
-
Thank you @David for your insight. After more in-depth reading about the DOE, I find that applying it seems highly inefficient, and when it is useful it is the exception and not the norm, so I will be sticking to the Knockout experiments.
-
The Knock out experiment sounds interesting, however, knowing what each ingredients function is the easiest way to reverse engineer eg structural ingredients, solvent, humectant etc. You can quickly find out the ingredients that are there for the hell of it?
-
knockouts good for diagnosing problems (by removing an ingredient), this gives you the resulting formulation without the ingredient.
… but it doesn’t work, to quantify how pairs of ingredients, actually cause the problem; and what there “additive” effect does, thats beneficial.
for example:
take a formulation with ingredients A through E; where the underlying problem is a conflict between ingredients B, or C, and where B interacts beneficially with A, D and C interacts with beneficially with D, E.
knocking out B, or C, will clearly finger at least one of the culprits.
and if you do the full range of knockouts, it will (eventually) finger the other culprit.
… but it doesn’t quantify how removing B or C, actually affects the rest of the formulation. Perhaps the formulation, with the “problem resolved” is actually worse (B or C removed, or substituted), than if the formulation had B and C in it, with the addition of a third ingredient (F) to specifically resolve the conflict between B and C.
simply removing B or removing C, even with a substituted ingredient, may result in a worse formulation, simply because you fail to study the actual negative interaction between B and C vs the beneficial interactions of B or C.
this I think, is where the skill of the formulator comes into play. going through these forums, there are countless substitutions, and sometimes additions, to address specific types of problems.
if you like rigor, I think experimental design is a good way to go about developing this expertise, in a structured way, for the subset of formulations you may care about.
-
Hello @aperson,
Thank you for your input, in my readings about the subject, I was able to differentiate between the two main category of tasks required from a cosmetic formulator. The first category is the function of the formulation (moisturizing, anti-ageing, cleansing, …), this category can be quantified up to a certain point, and experimental designs can be useful in this category. The second category is the aesthetics of the formulation, the knockout experiments are better suited for this category. To the best of my knowledge, there is no paper about the use of DOE in improving the feel or eliminating the stickiness from a formula, but there is plenty about improving the stability of an emulsion for example. -
@aperson - all you say is correct. Knockout experiments have limitations. But the value they provide vs the amount of work required is unmatched by other experimental approaches. Especially, if you lack significant experience in a formulation type.
-
Good points @aperson
-
sure, the split between a formulations performance characteristics (formulators perspective), and the sensorials of the final product (customers perspective). the true black arts is linking one, to the other — with respect to an “uncertain” customer (oily skin, dry skin; curly hair, straight hair; young vs old etc) i.e. how to make tradeoffs in formulations so that your product has broader appeal or better performance with respect to a specific segment.
that is, the aesthetics *utilities* (not dimensions!), while linked to the formulation, are not necessarily the same utility depending on which customer is using your product.
There is no “perfect” formula, with respect to the general public, because the general public, has mutually exclusive needs. So this then depends on which segment of the population you are targeting with your cosmetics; in order to quantify the tradeoffs in sensorials.
This is where “niche” products come in; not to do everything ok, but to do one thing, particularly well. @perry I think has mentioned this several times. That the best place to start with introducing a new line is to determine who your customer will be (“formulate the story”) then formulate the formulation to match that customer segments needs.
> To the best of my knowledge, there is no paper about the use of DOE in improving the feel or eliminating the stickiness from a formula
There is nothing inherent in DOE that precludes it use from handling sensorials; I would look for a book (or a paper) that first sets up the dimensions (categories) for sensorials, then look for methods of quantifying that sensorial either scientifically, or by survey.In fact I think this is precisely where DOE would shine; as it forces you to consider, how your population target, affects the “realizability” of optimal design. Much like the color space model limits the reproduction of colors outside its gamut — some characteristics simply cannot be achieved simultaneously.
-
@Perry
> But the value they provide vs the amount of work required is unmatched by other experimental approaches.
I agree
> Especially, if you lack significant experience in a formulation type.
Indeed. As I lack significant experience in formulation, much less a formulation type, knockouts present a singularly useful method for refining formulas.
Log in to reply.
