Aliphatic aldehyde production in shortbread cookies (extended shelf life, without butter)

Shortbread cookies were sent in for testing. The opened, returned product had a foul pungent odor. The customer thought it was contaminated with some type of acid, and reported a burning sensation in the mouth and etching of the teeth. This was a product intended for an extended shelf life, and was vacuum packed. The ingredients did not include butter; instead the label (shown below) lists margarine (with partially hydrogenated soybean oil and liquid soybean oil).

Below:  GC-MS peaks in normal cookies . Major peaks are heptadienal, the 2 decadienal isomers (normal breakdown products from linolenic acid with 2 double bonds), vanillin and ethyl vanillin (flavorings). The later peaks on the right starting around 30 minutes are the usual tocopherols, sterols and triterpenes (with no hydrocarbon waxes). The mass spectrum in the lower window is the ethyl vanillin peak (MW 166) at 19.2 minutes.

====================================

Below:  Peaks in the customer complaint sample:

The main peaks causing malodor in the bad sample were n-aldehydes from C4 to C11, and some unsaturated n-aldehydes up to C11:1.  i.e. "C6 ALD" = hexanaldehyde, "C8 ALD" = octanal, etc.

Free fatty acids (FA) were the next biggest abnormal peaks; C6 (hexanoic acid at 11 minutes) was the most abundant of these. The longer chain C16 & C18 fatty acids are to be expected in all foods; shorter chain FAs are less common.

Others of note: 1-pentanol at 5.5 minutes, and 2-heptanone at 8.1 minutes. Up to 100 other peaks are present and were examined after zooming in, but the tentative IDs will not fit on the printout without cluttering things. Authentic reference standard solutions were available for n-alkanes, n-aldehydes, n-alkanoic acids, 1-alkanols, alkan-2-ones and some others. Standards for unsaturated and branched compounds were generally not available. Kovats retention indices on a DB5 column were consulted when available.

===============================================

Comparison overlay. Complaint sample Top, normal sample way down at the Bottom barely visible,

followed by 2 zoomed-in views to show how much bigger the peaks are in the complaint sample.

==================================================================

Follow-up aging test a month later: looking for deteriorative changes in a normal cookie opened 1 month ago, compared to another normal cookie opened today.

There was no huge increase in breakdown products after 1 month. Things might be different for real shortbread cookies made with butter, rather than soybean oil based margarine. However, the soybean oil blend contained enough PUFAs to go very bad by 3 years in the bad sample.

=======================================================

Comment, additional info:

There are so many chemically aggressive, noxious breakdown products here, it's not surprising it isn't immediately recognized as plain old rancidity, at a very advanced stage.

All of the cookies were over 3 years old (they were intended to have extended shelf-life; probably why they were made with margarine instead of butter).

The package of the returned complaint sample was examined, and found to have a single pinhole leak. This most certainly has a huge effect on lipid oxidation over an extended time. It could not be determined when and where the damage (or defect) occurred.

The analyst (TA) looked at a couple of the normal unopened cookies (3 years old); they had a tolerable aroma, taste and texture. TA washed away the after-taste with a Vitamin C  containing juice (OJ, tomato).

Ingredient label showing oil source (precursor to the aldehydes):