Tuesday, 14 September 2010

The Art and Science of Felix Replicata

Following a telephone enquiry I was sent a sample of a dearly loved but discontinued fragrance by a client wondering if I might be able to reproduce it. My first impression upon smelling it from the bottle was of a quality, predominately citrus, light, floral affair with some delicate woody and slightly balsamic notes.

Further sniffs using a strip revealed notes suggestive of bergamot, mandarin, grapefruit and lemon over rosewood and some very subtle warm, woody base notes, which I was unable to fathom even after dry down. The fragrance also contained synthetics. To compliment my olfactory impressions I sent the sample for analysis by GLC.

The analysis revealed a total of twenty nine constituents. This low number suggests the use of nature identical chemicals. The column headers on the analysis sheets have the following meanings: - PK is an abbreviation of peak and represents a grouping of like molecules selected by volatility. RT means retention time; this is the time taken by each group of molecules to travel the length of the column. Area% is the percentage of the sample volume comprised by each constituent. Library/ID, these are constituents identified in a database of approximately 50,000 volatile molecules, which are probable matches for each molecule isolated and grouped by retention time. Ref and CAS are ways of identifying each molecule by number in the database and international reference respectively. Qual is an indication of the accuracy of the match between each molecule identified by retention time and its suggested counterpart from the database. It is expressed by percentage probability.
We can see that the volume of identified separate constituents in the composition varies between 19.26% and 0.33%, whilst the probability of identifying each accurately ranged from 2% to 98%.

Essential oils are a highly complex mixture of volatile molecules and many have constituents in common. The latter fact means that many of their odour profiles will overlap too. Rational deduction supplements our olfactory enquiry, which is precisely the reason for analysis.
Taking it from the top - linalyl acetate is the constituent that comprises the greatest volume in the fragrance. However, it would be a mistake to assume that greatest volume was synonymous with characteristic odour as this ignores odour intensity. Some years ago I proposed essential oil of garlic to be added to a handbag spray product containing strong dyes for marking assailants or fitted under the dash to mark potential car thieves. It was too effective to be viable!

So with that proviso let’s see how far we can get with deduction to supplement our olfactory impressions. Linalyl acetate occurs in many essential including lavender and bergamot, but it is also available as a nature ‘identical’ chemical. Analysis will not disclose the source only the presence. If we knew the percentage of linalyl acetate present in any of the oils that contain it then we could determine how much of each contender may be present. In order to determine this we may get clues by application of theory to other constituents.

We do have parameters for linalyl acetate in certain oils but they can be quite diverse. If we looked at all the oils that contain linalyl acetate in relation to the overall volume in the fragrance sample analysed we may then be able to predict the presence of other key constituents from our knowledge of the composition of those essential oils.

To cut a long saga short we have linalol at 8.36% so lavender is immediately excluded as the source because the relative proportions are out. Other key constituents are limonene, phenyl ethyl alcohol (PEA) and citronellol. PEA can occur as highly as 38% in Rosa Damascena, but it also occurs in smaller percentages in geranium alongside citronellol and limonene. The lack of trace elements excludes rose Otto but the citrus oil which is implicated by this combination is bergamot.

We also know from citrus oil analysis that anthranilate methyl is the constituent that distinguishes the aroma of mandarin from that of tangerine and other citrus oils with similar constituents albeit in different proportions.

Peaks 18 and 20 identify guaiene and patchoulene as possible matches within the database and peak 20 identifies patchouli alcohol with 99% certainty. Guaiacwood oil is implicated by the presence of guaiene but the proportion of patchouli alcohol present without trace elements suggests it was added as a nature identical.

After those deductions it was back to working with the nose. As I don’t use nature identical chemicals I had to find oils which contained the principal constituents identified but were also low on other constituents.

The principal of nasal fatigue is an ally in the systematic and progressive attempted construction of any Felix Replicata fragrance. In other words when we enter an environment which has a certain odour and we stay there for even a short while our ability to detect that odour fades into the background.

I began by blending several oils adding amounts directed by experience. Mandarin, bergamot, lemon, grapefruit, lime, guaiacwood, patchouli and citronella were my first choices. The next step was to smell my blend carefully for several seconds before having a sniff of the target sample. Due to the principal of nasal fatigue the odours that stand out in the target sample are those lacking in the attempted replication.

The odour which stood out most as being different was a floral quality, which I know to be linalol (rosewood oil is about 80% linalol, so not as clever as it sounds). There was also a soapy note, which was missing in my sample. Rosewood oil is illegal but ho wood comes close and the soapy note I found by blending lime, lavender and cedarwood Virginian. Having added those and repeated the comparison geranium suggested itself. After adding geranium and repeating the process some of the citrus notes demanded adjustment. For the synthetic component I used a watery, fougere, citrus compound at 4.2% by volume.

This was fun and the first time that I have attempted to copy a fragrance. A mainstream fragrance house would have shelves full of nature identical synthetics. Each attempt at a match would be analysed to hone in until a precise match was attained.

The two most recent European court cases involving copyright of perfume were in Holland and France. Passing off by using similar packaging and fragrance name is easily upheld in court but the actual aroma of a fragrance is not currently protected legally. The reason is that so far courts have taken the view that aroma is a subjective sensory experience and nothing of the personality of the perfumer passes over into a fragrance that would enable the perfumer to be identified in contrast to books and music. The complexity of legally defining key odours and upholding their restrictive use would be a farce. I would expect such legal nonsense to emerge in the US if anywhere.