Neutral Oil Loss During Alkali Refining

Abstract

Dear Sir, While attending a Short Course held in the US last February, I met a refinery manager with whom I discussed the differences between the Long Mix neutralization process as commonly used in the US and the Short Mix process, which is more common in Europe. At one stage he mentioned that his refining loss factor was 1.15 and I asked him to explain that in more detail; his loss struck me as on the low side. He then told me that he first of all determined the total neutral oil content (expressed as a percentage) of his crude oil while using a laboratory method according to AOCS Official Method Ca 9f-57. This neutral oil content gave him the inevitable loss on alkali refining by subtracting the neutral oil content from 100%. Then in the plant, he determined the actual refining loss by subtracting the ‘‘Refined oil yield’’ (wt) from the ‘‘Crude oil input’’ (wt) and finally, he arrived at this loss factor by dividing the actual refining loss (expressed as a percentage) by the inevitable loss (also expressed as a percentage). What he did was standard procedure as described in the handbooks [1]. I then told him that in Europe, it is more common to calculate a ‘‘Theoretical loss’’ (TL) defined as (in percentage of crude oil): TL = FFA ? Phosphatides ? Moisture ? Impurities ? 0.3 This TL is used by suppliers of centrifugal separators in guarantees that limit the actual loss to 1.25 9 TL ? 0.3 (%) in which equation the factor of 1.25 is higher than the loss factor of 1.15 quoted by my refiner. However, I should point out that the definition of this TL is far from ideal because it is not clear how the phosphatide content is defined. It can be calculated from the phosphorus content of the oil but what conversion factor should be used? For the acetone insoluble (AI) of lecithin, a factor of 31 is used since the AI also includes carbohydrates and glycolipids. Looking at the phosphatides themselves (those that are left after water degumming), a conversion factor close to 24 would be more appropriate [2]. In addition, the equation to calculate the actual loss from the TL assumes the FFA and the phosphatides entrain the same relative amounts of neutral oil. It even assumes the moisture and impurities present in the crude oil to entrain a quarter of their weight in neutral oil. This equation is clearly an oversimplification but data on which to base a more accurate one are lacking. As pointed out to the author during the reviewing stage, such a more accurate equation might also have to take the partial glyceride content of the crude oil into account. To obtain a better understanding of these loss factors, I decided to study the literature and that turned out to be a fascinating and intriguing exercise indeed. So I came across a most informative article from 1955 [3] stating that ‘‘the oldest and best known method is that devised by David Wesson’’. The article gives a number of references from 1922 and 1926 the latter one of which [4] I only got hold of recently, but the method is so important that various authors [3, 5–7] have described the method in detail: