<title>Characterization of Fatty Matter in Cold Process Soap by ¹³C NMR Spectroscopy</title>

<author><firstname>Kevin</firstname><surname>Dunn</surname></author> <author><firstname>Andrew</firstname><surname>McLeod</surname></author> <author><firstname>John</firstname><surname>Robbins</surname></author> <author><firstname>Robert</firstname><surname>O'Cain</surname></author>

The purpose of this project is to analyze the composition of fatty matter in cold process soap by extracting excess oil from soap and performing 13CNMR. The three oils which were used in the soap making process were olive oil for which oleic acid is the main fatty acid, castor oil for which ricinoleic acid is the main fatty acid, and grapeseed oil for which linoleic acid is the main fatty acid. Each soap contained 90% palm oil and 10% of the respective oil. It was found that composition of fatty matter in cold process soap does not differ depending on whether or not the minor oil was added early or late in the saponification process. In all three cases, the percent of unsaturated fatty acid is higher in the original oil from which each soap was made.

During the past thirty years, public interest in handcrafted-soap has led to a resurgence of the small-scale manufacturer: the possibility of creating truly unique bars of soap with various shapes, molds, fragrances, and appearances that attract both entrepreneurs and hobbyist alike <xref linkend="R1"/>. The growing interest in organic alternatives to petroleum based inductrial detergents has also contributed to the handcrafted soap craze. As a result, these soap makers tend to be average people working in small home-kitchen laboratories with tools and ingredients found in a typical American supermarket. Their procedures tend to be adaptations of recipes found in unscientific cookbooks, and their ingredients are typically natural vegetable oils and fragrances. Despite relatively crude tools and procedures however, handcrafted soaps tend to be of extremely high quality and often sell for four or five times that of ready-made soaps and detergents found in one's local grocery store <xref linkend="R1"/>.

<bibliomixed id="R1">Dunn, Kevin M., <citetitle>Scientific Soapmaking. Unpublished Manuscript</citetitle>, July 2006.</bibliomixed>

Discounting or superfatting creates soaps that consist of excess oil. Discounting involves reducing the stoichiometric amount of lye solution added to the saponification mixture, while superfatting involves increasing the stoichiometric amount of oil added to the saponification mixture. The main reason for discounting/superfatting is to have excess oil in the resulting soap which increases the moisturizing or emollient features the soap. One problem which may arise from discounting/superfatting is that it is thought to cause dreaded orange spots (DOS)<xref linkend="R2"/>. Studies investigating this hypothesis have been carried out by Hampden Sydney Students under the supervision of faculty member Dr. Kevin Dunn. The cause of the DOS was attributed to oxidation of the soap from oxygen in the air. This conclusion was reached by conducting several experiments in which various mixtures of the antioxidants Sodium Ascorbate (Vitamin C), TocoPherol, Ethylenediaminetetraacetic acid (EDTA), Butylated Hydroxtoluene (BHT), and Rosemary Extract (ROE) were added to soap samples prior to saponification <xref linkend="R3"/>. The resulting samples were then saponified using the Cold Process method and placed in either pure oxygen or pure nitrogen environments <xref linkend="R3"/>. Subsequent photographic scans were canmducted everyday on each sample and compared using the GNU Manipulation Program (GIMP) <xref linkend="R3"/>. The results concluded that EDTA and ROE were the most effective at preventing DOS <xref linkend="R3"/>.

<bibliomixed id="R2">Dunn, Kevin M., <citetitle>The Journal of the Hand Crafted Soapmakers Guild</citetitle>, Winter 2005/2006.</bibliomixed>

<bibliomixed id="R3">Dieglemann, Stephen R.; Dunn, Kevin M., <citetitle>An Investigation into the Cause and Prevention of Rancidity and Soda Ash in Cold Process Soap. Hampden Sydney College</citetitle>, Spring 2005.</bibliomixed>

The composition of soap also interests soap makers. The process of soap making has remained relatively unchanged since the mid-eighteen hundreds; for the most part, soap consists of a mixture of fatty acids, usually a vegetable oil, and caustic alkali suck as Sodium Hydroxide (Lye) or Potassium Hudroxide: Sodium Hydroxide produces hard or solid soaps while Potassium Hydroxie produces soft or liquid soaps <xref linkend="R4"/>. The saponification process or the process by which these chemicals react to form soap is a delicate process as it requires precise ratiosof the two reactants: excess fatty-acid reduces the soap's ability to produce lather <xref linkend="R4"/>. Nevertheless, soap makers often emply a method known as superfatting or discounting in order to ensure the excess of a particular oil.

<bibliomixed id="R4">Wagner, Rudolph.<citetitle>Soap. Wagner Chemical Technology. Appleton & Company. </citetitle>, 1872, 239-249.</bibliomixed>

Superfatting or discountiong is commonly practiced by soapmakers in an effort to ensure that a particular oil has been reacted completely by using less than the stoichiometric amount of lye that is necessary. However, there is little scientific evidence confirming the effectiveness of this process. Therefore, an analysis of the structure of fatty-acid soap crystals may serve to illustrate how a soap crystal is formed as well as if a particular fatty-acid reacts more readily over another <xref linkend="R5"/>. What is worth noting is that while soap (the salt of the fatty acid) may keep for a long time, oil does not; some chemical change occurs that yields the unsightly orange spots.

<bibliomixed id="R5">Lynch, Matthew L; Wireko, Fred; Tarek, Mournir; Klein, Michael.<citetitle>Intermolecular Interactions and the Structure of Fatty Acid-Soap Crystals. Journal of Physical Chemistry B. </citetitle>, 2000, 552-561.</bibliomixed>

All analytical solutions were prepared and measurements made using sunthetic methods with an analytic balance of capacity 200g and readability 0.01g (Acculab V-200 and JSCALE JS-XV Manual). All soaps were prepared using the semi-boiled process and the GE-168962 18-quart roaster oven. All vegatable oils used were grocery store name brand oils: Columbus Palm oil. Lye was purchased from Bayer Corporation. All other chemicals and solvents were purchased from Aldrich. Structural determination and analysis of oils was performed by nuclear magnetic resonance spectroscopy (H-NMR and ¹³C-NMR) analysis using the Peabody, MA JOEL Eclipse plus 400 MHz NMR spectrometer. IR spectroscopy was performed using the Excalibur HE Series FTS 3100 IR Spectrometer. GC/MS analysis was performed using the Varian CP-3800 Gas Chromatography with the Saturn 2200 GC/MS instrument. Data analysis and any subsequent calculations or graphing were performed using Microsoft Excel. In order to investigate the effects of superfatting/discounting on the composition of handcrafted soap, it was first necessary to master the soap making process. One-hundred gram (100g) samples of Palm oil soap were prepared using the appropriate saponification amount of standardized 333.3ppt NaOH lye solution: the soaps were made in accordance with the methods laid out in Scientific Soapmaking.¹ Individual procedures were observed by other lab partners and compared to further improve and perfect individual methods. The properties of the subsequent soaps were then compared.

Preparation of Experimental Palm/Grapeseed oil Soaps:Two soaps were prepared using standard soap making procedure drawn from Scientific Soapmaking.¹ Two mason jars were charged with 90g Palm oil. To one of these jars was added 10g grapeseed oil, and the jar labelled 10% lye discount added before trace. To this jar was added 38.10g(quantity calculated for 10% disount) lye solution. The jar was shaken, mixture brought to trace, tranferred to rubber mold and baked at 140-170 degrees F for 4 hours. The other jar was charged with 90g Palm oil, and labelled 10% lye discount added at trace. 38.10g lye was added to this jar, and mixture brought to trace at which point 10g grapeseed oil was added. The soap was transferred to rubber mold and baked at 140-170 degrees F for 4 hours. Both soaps were allowed to cool before experimentation was continued. The same procedure was used for a Palm Oil/Castor Oil soap with a 10% discount and also a Coconut oil/Olive oil soap with a 10% discount.

Extraction of Excess Oils from Experimental Soaps: 20g of each Coconut/olive oil soap was shaved off and placed in separate 250ml round-bottomed flasks. Each sample was charged with 250ml ethyl ether (Aldrich), fitted with a column condensor, and refluxed for two hours in a hot water bath at 70 degrees C. The ether was then extracted from the solid residue through vacuum filtration. The ether was then washed with 250ml brine solution in a separatory funnel, and then evaporated using Rotovap. The residue was dried using a vacuum pump, weighed and characterized using ¹³C-NMR Spectroscopy. Palm/castor oil and palm/grapeseed oil soap samples were prepared with 40g of each soap in order to get more extracted oil for characterization purposes.

Through the use of the ¹³CNMR, the extracted oils were able to be characterized and compared against each other. The ¹³CNMR spectra of the oil from a soap which had the minor oil added at trace had no meaningful difference between the oil from a soap where the minor oil was added before trace. The fact that there is no difference between the extracted oils suggests that it does not matter whether or not the oil is added before trace or at trace because saponification occurs during the cooking phase rather than the mixing phase. The important range for deciding the composition of the fatty acid in the extracted oil was the 125ppm to 135ppm range.<xref linkend="F6"/> shows the spectrum of 90% Palm Oil and 10% Castor Oil. That range corresponds to the double bonds present in the oil. For Castor Oil, for which ricenoleic acid is the major fatty acid, there are two peaks in the region at 125ppm and 133ppm. In Palm Oil, which has two double bonds, there are peaks which can be found at 127ppm, 129ppm, and 130ppm. Each of the spectra acquired from the extracted oils showed a larger 127-130ppm range and two much smaller peaks at 125ppm and 133ppm respectively which suggests that, in the case of both palm/castor and palm/grapeseed oil soaps, most of the oil extracted is palm oil.<xref linkend="T1"/> Shows an NMR spectrum for 90% palm / 10% castor oil.

<figure id="F6"><title>90% Palm Oil 10% Castor Oil CNMR Spectrum</title> <mediaobject><imageobject><imagedata fileref="90palmoil10castoroilCNMR.jpg" format="JPG" scale="20"/></imageobject></mediaobject> </figure>

In order to determine the amount of ricenoleic acid in a sample, ¹³CNMR was run on series of known mixtures to determine the ratio between the peak at 125ppm(double bond in ricenoleic acid) and 14ppm(the terminal methyl group present in all fatty matter). First a peak ratio for 99% ricenoleic acid, ordered from Aldritch, was obtained. After the peak ratio was obtained for 99% ricenoleic acid, a series of known samples were prepared, 100% palm oil, 100% castor oil, 75% palm oil 25% castor oil, 50% palm oil 50% castor oil, and 25% palm oil 75% castor oil. The peak ratios were plotted using Excel and the calibration curve had an R2 value of .9938. <xref linkend="F4"/> Shows the calibration plot of percent ricenoleic acid vs. percent castor oil. The data suggests that it is appropriate to be able to compare the peak at 125ppm vs. the peak at 14ppm and multiply it by the constant obtained in the 99% ricenoleic acid sample, 120.83, to determine the total ricenoleic acid content of a sample. This was done for each of the extractions and it was found to be that there was an average of 2-4% ricenoleic acid leftover after the extractions. The extractions which yielded higher % ricenoleic acid can be attributed the increased soapy matter in the sample. The extraction was not as good and therefore more of the extraction was soap than the others.

<figure id="F4"><title><superscript>13</superscript>C-NMR Spectral Analysis Data for Castor Oil</title> <mediaobject><imageobject><imagedata fileref="castoroilpercent.pdf" format="PDF" scale="50"/></imageobject></mediaobject> </figure>

<table id="T1"><title>% Ricenoleic Acid From Each Extraction</title>

</table>

Another section of the ¹³CNMR spectrum was analyzed to determine the content of the glyceride molecules present. The range from 170ppm to 180ppm gives properties of the glyceride molecules and the relative area of the peaks can be used to determine the amount of each glyceride present. The peak at 174ppm represents the carboxylic acid group from a monoglyceride, the peaks in the 173ppm represents the carboxylic acid group from a diglyceride and the peaks in the 172 range represents the carboxylic acid group from a triglyceride. The peak at 178ppm is the free fatty acid. From the spectra acquired from the extractions, most of the glycerides are mono- and di-glycerides. As expected, the spectrum of the pure fatty acid(ricenoleic acid) shows only one peak at 178ppm.<xref linkend="R6"/>

<bibliomixed id="R6">Hamilton, R.J., <citetitle>Lipid Analysis in Oils and Fats</citetitle>, 1998. 93-117</bibliomixed>

In order to determine the amount of linoleic acid in a sample, ¹³CNMR was run on series of known mixtures to determine the ratio between the peak at 129ppm(double bond in linoleic acid) and 14ppm(the terminal methyl group present in all fatty matter). First a peak ratio for 99% linoleic acid, ordered from Aldritch, was obtained. After the peak ratio was obtained for 99% linoleic acid, a series of known samples were prepared, 100% palm oil, 100% grapeseed oil, 75% palm oil 25% grapseed oil, 50% palm oil 50% grapeseed oil, and 25% palm oil 75% grapeseed oil. The peak ratios were plotted using Excel and the calibration curve had an R2 value of .9758. The data suggests that it is appropriate to be able to compare the peak at 129ppm vs. the peak at 14ppm and multiply it by the constant obtained in the 99% ricenoleic acid sample, 80.775, to determine the total ricenoleic acid content of a sample. This was done for each of the extractions and it was found to be that there was an average of 16-19% linoleic acid leftover after the extractions. This procedure was repeated for coconut/olive oil soap samples, where it was found to be that there was an average of 21% oleic acid leftover after the extractions.

<xref linkend="T2"/> summarizes the results from the palm/grapeseed oil soap studies.

<xref linkend="F5"/> shows the calibration curves generated from ¹³C-NMR Spectral Analysis for the three different set of oil combinations. Such calibration curves would enable soap makers to quickly determine relative fatty matter composition of finished soaps.

The results from this experiment can be very useful for soap makers who waste a lot of time and effort and soap by adding the minor oil at trace. The percent (%) yield of the soap is always higher when the minor oil is added before trace because more of the soap is poured into the mold when it reaches trace, instead solidifying on the walls of the mixing apparatus. The research could prove to save soap makers money because they would get more out of their materials and would have less waste.

<table id="T2"><title>Summary of Results (Palm/grapeseed oil soaps)</title>

</table>

<xref linkend="F5"/> shows the calibration curves generated from ¹³C-NMR Spectral Analysis for the three different set of oil combinations. Such calibration curves would enable soap makers to quickly determine relative fatty matter composition of finished soaps.

<figure id="F5"><title>Combined <superscript>13</superscript>C-NMR Spectral Analysis Data</title> <mediaobject><imageobject><imagedata fileref="combinedoilvsacidpercent.pdf" format="PDF" scale="50"/></imageobject></mediaobject> </figure>

Figures are uploaded via the attachment interface in the same line as "edit." Figure files are to be uploaded in two formats:

1. in the original file format (e.g. Titration.xls for an excel file)
2. as a postscript image of the figure

To get a postscript image in Excel, you have to install a postscript printer on your computer. Don't worry that one isn't attached, just install it as if it were a local printer. I have used an HP Laserjet 5/5M Postscript printer driver with excellent results. Once the printer is installed, select the chart with your mouse and go to "File:Print." When the dialog comes up, make sure you are printing the chart and not the whole file. Select the postscript printer you installed and then "print to file." You will then enter filename, e.g. Titration.ps. Do not use a WikiName for the filename. Once the file is saved on your computer you can upload it as an attachment to this wiki page.

One more thing, when you update your attachments you have to delete the old attachments before uploading the new ones. The wiki will not overwrite existing files.