TED2005
Luca Turin: The science of scent
Filmed:
Readability: 3.7
621,001 views
What's the science behind a sublime perfume? With charm and precision, biophysicist Luca Turin explains the molecular makeup -- and the art -- of a scent.
Luca Turin - Nose
Biophysicist Luca Turin studies the science of smell. He's the author of Perfumes: The Guide, and the subject of Chandler Burr's 2003 book The Emperor of Scent. His next project: developing an artificial nose. Full bio
Biophysicist Luca Turin studies the science of smell. He's the author of Perfumes: The Guide, and the subject of Chandler Burr's 2003 book The Emperor of Scent. His next project: developing an artificial nose. Full bio
Double-click the English transcript below to play the video.
00:12
The fragrance that you will smell, you will never be able to smell this way again.
0
0
4000
00:16
It’s a fragrance called Beyond Paradise,
1
4000
3000
00:19
which you can find in any store in the nation.
2
7000
3000
00:22
Except here it’s been split up in parts by Estée Lauder
3
10000
4000
00:26
and by the perfumer who did it, Calice Becker,
4
14000
3000
00:29
and I'm most grateful to them for this.
5
17000
1000
00:30
And it’s been split up in successive bits and a chord.
6
18000
4000
00:34
So what you’re smelling now is the top note.
7
22000
2000
00:36
And then will come what they call the heart, the lush heart note.
8
24000
4000
00:40
I will show it to you.
9
28000
3000
00:43
The Eden top note is named after the Eden Project in the U.K.
10
31000
3000
00:46
The lush heart note, Melaleuca bark note -- which does not contain any Melaleuca bark,
11
34000
5000
00:51
because it’s totally forbidden.
12
39000
3000
00:54
And after that, the complete fragrance.
13
42000
3000
00:57
Now what you are smelling is a combination of --
14
45000
3000
01:00
I asked how many molecules there were in there, and nobody would tell me.
15
48000
4000
01:04
So I put it through a G.C., a Gas Chromatograph that I have in my office,
16
52000
5000
01:09
and it’s about 400.
17
57000
3000
01:12
So what you’re smelling is several hundred molecules
18
60000
3000
01:15
floating through the air, hitting your nose.
19
63000
3000
01:18
And do not get the impression that this is very subjective.
20
66000
4000
01:22
You are all smelling pretty much the same thing, OK?
21
70000
4000
01:26
Smell has this reputation of being somewhat different for each person.
22
74000
4000
01:30
It’s not really true.
23
78000
1000
01:31
And perfumery shows you that can’t be true,
24
79000
2000
01:33
because if it were like that it wouldn’t be an art, OK?
25
81000
6000
01:39
Now, while the smell wafts over you, let me tell you the history of an idea.
26
87000
4000
01:43
Everything that you’re smelling in here
27
91000
4000
01:47
is made up of atoms that come from what I call
28
95000
2000
01:49
the Upper East Side of the periodic table -- a nice, safe neighborhood.
29
97000
3000
01:52
(Laughter)
30
100000
4000
01:56
You really don’t want to leave it if you want to have a career in perfumery.
31
104000
3000
01:59
Some people have tried in the 1920s
32
107000
2000
02:01
to add things from the bad parts, and it didn’t really work.
33
109000
4000
02:05
These are the five atoms from which just about everything
34
113000
2000
02:07
that you’re going to smell in real life, from coffee to fragrance, are made of.
35
115000
6000
02:13
The top note that you smelled at the very beginning,
36
121000
2000
02:15
the cut-grass green, what we call in perfumery -- they’re weird terms --
37
123000
4000
02:19
and this would be called a green note,
38
127000
2000
02:21
because it smells of something green, like cut grass.
39
129000
2000
02:23
This is cis-3-hexene-1-ol. And I had to learn chemistry on the fly
40
131000
7000
02:30
in the last three years. A very expensive high school chemistry education.
41
138000
6000
02:36
This has six carbon atoms, so "hexa," hexene-1-ol.
42
144000
4000
02:40
It has one double bond, it has an alcohol on the end,
43
148000
4000
02:44
so it’s "ol," and that’s why they call it cis-3-hexene-1-ol.
44
152000
3000
02:47
Once you figure this out, you can really impress people at parties.
45
155000
3000
02:50
This smells of cut grass. Now, this is the skeleton of the molecule.
46
158000
4000
02:54
If you dress it up with atoms, hydrogen atoms --
47
162000
3000
02:57
that’s what it looks like when you have it on your computer --
48
165000
2000
02:59
but actually it’s sort of more like this, in the sense that the atoms have a certain
49
167000
4000
03:03
sphere that you cannot penetrate. They repel.
50
171000
3000
03:06
OK, now. Why does this thing smell of cut grass, OK?
51
174000
6000
03:12
Why doesn’t it smell of potatoes or violets? Well, there are really two theories.
52
180000
6000
03:18
But the first theory is: it must be the shape.
53
186000
4000
03:22
And that’s a perfectly reasonable theory in the sense that
54
190000
2000
03:24
almost everything else in biology works by shape.
55
192000
2000
03:26
Enzymes that chew things up, antibodies, it’s all, you know,
56
194000
4000
03:30
the fit between a protein and whatever it is grabbing, in this case a smell.
57
198000
6000
03:36
And I will try and explain to you what’s wrong with this notion.
58
204000
3000
03:39
And the other theory is that we smell molecular vibrations.
59
207000
3000
03:42
Now, this is a totally insane idea.
60
210000
2000
03:44
And when I first came across it in the early '90s, I thought my predecessor,
61
212000
4000
03:48
Malcolm Dyson and Bob Wright, had really taken leave of their senses,
62
216000
3000
03:51
and I’ll explain to you why this was the case.
63
219000
3000
03:54
However, I came to realize gradually that they may be right --
64
222000
3000
03:57
and I have to convince all my colleagues that this is so, but I’m working on it.
65
225000
3000
04:00
Here’s how shape works in normal receptors.
66
228000
6000
04:06
You have a molecule coming in, it gets into the protein, which is schematic here,
67
234000
5000
04:11
and it causes this thing to switch, to turn, to move in some way
68
239000
5000
04:16
by binding in certain parts.
69
244000
4000
04:20
And the attraction, the forces, between the molecule and the protein
70
248000
2000
04:22
cause the motion. This is a shape-based idea.
71
250000
5000
04:27
Now, what’s wrong with shape is summarized in this slide.
72
255000
4000
04:31
The way --I expect everybody to memorize these compounds.
73
259000
5000
04:36
This is one page of work from a chemist’s workbook, OK?
74
264000
5000
04:41
Working for a fragrance company.
75
269000
2000
04:43
He’s making 45 molecules, and he’s looking for a sandalwood,
76
271000
3000
04:46
something that smells of sandalwood.
77
274000
1000
04:47
Because there’s a lot of money in sandalwoods.
78
275000
2000
04:49
And of these 45 molecules, only 4629 actually smells of sandalwood.
79
277000
7000
04:56
And he puts an exclamation mark, OK? This is an awful lot of work.
80
284000
3000
04:59
This actually is roughly, in man-years of work, 200,000 dollars roughly,
81
287000
5000
05:04
if you keep them on the low salaries with no benefits.
82
292000
4000
05:08
So this is a profoundly inefficient process.
83
296000
2000
05:10
And my definition of a theory is, it’s not just something
84
298000
3000
05:13
that you teach people; it’s labor saving.
85
301000
2000
05:15
A theory is something that enables you to do less work.
86
303000
3000
05:18
I love the idea of doing less work. So let me explain to you why -- a very simple fact
87
306000
6000
05:24
that tells you why this shape theory really does not work very well.
88
312000
5000
05:29
This is cis-3-hexene-1-ol. It smells of cut grass.
89
317000
5000
05:34
This is cis-3-hexene-1-thiol, and this smells of rotten eggs, OK?
90
322000
7000
05:41
Now, you will have noticed that vodka never smells of rotten eggs.
91
329000
5000
05:46
If it does, you put the glass down, you go to a different bar.
92
334000
4000
05:50
This is -- in other words, we never get the O-H --
93
338000
5000
05:55
we never mistake it for an S-H, OK?
94
343000
3000
05:58
Like, at no concentration, even pure, you know,
95
346000
3000
06:01
if you smelt pure ethanol, it doesn’t smell of rotten eggs.
96
349000
3000
06:04
Conversely, there is no concentration at which the sulfur compound will smell like vodka.
97
352000
7000
06:11
It’s very hard to explain this by molecular recognition.
98
359000
4000
06:15
Now, I showed this to a physicist friend of mine who has a profound distaste
99
363000
2000
06:17
for biology, and he says, "That’s easy! The things are a different color!"
100
365000
4000
06:21
(Laughter)
101
369000
5000
06:26
We have to go a little beyond that. Now let me explain why vibrational theory has
102
374000
5000
06:31
some sort of interest in it. These molecules, as you saw in the beginning,
103
379000
4000
06:35
the building blocks had springs connecting them to each other.
104
383000
2000
06:37
In fact, molecules are able to vibrate at a set of frequencies
105
385000
3000
06:40
which are very specific for each molecule and for the bonds connecting them.
106
388000
4000
06:44
So this is the sound of the O-H stretch, translated into the audible range.
107
392000
8000
06:53
S-H, quite a different frequency.
108
401000
2000
06:55
Now, this is kind of interesting, because it tells you
109
403000
4000
06:59
that you should be looking for a particular fact, which is this:
110
407000
3000
07:02
nothing in the world smells like rotten eggs except S-H, OK?
111
410000
4000
07:06
Now, Fact B: nothing in the world has that frequency except S-H.
112
414000
7000
07:13
If you look on this, imagine a piano keyboard.
113
421000
2000
07:15
The S-H stretch is in the middle of a part of the keyboard
114
423000
4000
07:19
that has been, so to speak, damaged,
115
427000
2000
07:21
and there are no neighboring notes, nothing is close to it.
116
429000
3000
07:24
You have a unique smell, a unique vibration.
117
432000
2000
07:26
So I went searching when I started in this game
118
434000
2000
07:28
to convince myself that there was any degree of plausibility
119
436000
3000
07:31
to this whole crazy story.
120
439000
2000
07:33
I went searching for a type of molecule, any molecule,
121
441000
3000
07:36
that would have that vibration and that -- the obvious prediction
122
444000
5000
07:41
was that it should absolutely smell of sulfur.
123
449000
2000
07:43
If it didn’t, the whole idea was toast, and I might as well move on to other things.
124
451000
6000
07:49
Now, after searching high and low for several months,
125
457000
4000
07:53
I discovered that there was a type of molecule called a Borane
126
461000
5000
07:58
which has exactly the same vibration.
127
466000
2000
08:00
Now the good news is, Boranes you can get hold of.
128
468000
3000
08:03
The bad news is they’re rocket fuels.
129
471000
3000
08:06
Most of them explode spontaneously in contact with air,
130
474000
3000
08:09
and when you call up the companies, they only give you minimum ten tons, OK?
131
477000
4000
08:13
(Laughter)
132
481000
3000
08:16
So this was not what they call a laboratory-scale experiment,
133
484000
3000
08:19
and they wouldn’t have liked it at my college.
134
487000
2000
08:21
However, I managed to get a hold of a Borane eventually, and here is the beast.
135
489000
8000
08:29
And it really does have the same -- if you calculate,
136
497000
2000
08:31
if you measure the vibrational frequencies, they are the same as S-H.
137
499000
4000
08:35
Now, does it smell of sulfur? Well, if you go back in the literature,
138
503000
5000
08:40
there’s a man who knew more about Boranes than anyone
139
508000
3000
08:43
alive then or since, Alfred Stock, he synthesized all of them.
140
511000
5000
08:48
And in an enormous 40-page paper in German he says, at one point --
141
516000
6000
08:54
my wife is German and she translated it for me --
142
522000
3000
08:57
and at one point he says, "ganz widerlich Geruch,"
143
525000
2000
08:59
an "absolutely repulsive smell," which is good. Reminiscent of hydrogen sulfide.
144
527000
6000
09:05
So this fact that Boranes smell of sulfur
145
533000
3000
09:08
had been known since 1910, and utterly forgotten until 1997, 1998.
146
536000
7000
09:15
Now, the slight fly in the ointment is this: that
147
543000
4000
09:19
if we smell molecular vibrations, we must have a spectroscope in our nose.
148
547000
5000
09:24
Now, this is a spectroscope, OK, on my laboratory bench.
149
552000
5000
09:29
And it’s fair to say that if you look up somebody’s nose,
150
557000
2000
09:31
you’re unlikely to see anything resembling this.
151
559000
3000
09:34
And this is the main objection to the theory.
152
562000
2000
09:36
OK, great, we smell vibrations. How? All right?
153
564000
5000
09:41
Now when people ask this kind of question, they neglect something,
154
569000
2000
09:43
which is that physicists are really clever, unlike biologists.
155
571000
4000
09:47
(Laughter)
156
575000
3000
09:50
This is a joke. I’m a biologist, OK?
157
578000
2000
09:52
So it’s a joke against myself.
158
580000
1000
09:53
Bob Jacklovich and John Lamb at Ford Motor Company,
159
581000
3000
09:56
in the days when Ford Motor was spending vast amounts of money
160
584000
2000
09:58
on fundamental research, discovered a way
161
586000
4000
10:02
to build a spectroscope that was intrinsically nano-scale.
162
590000
4000
10:06
In other words, no mirrors, no lasers, no prisms, no nonsense,
163
594000
3000
10:09
just a tiny device, and he built this device. And this device uses electron tunneling.
164
597000
5000
10:14
Now, I could do the dance of electron tunneling,
165
602000
3000
10:17
but I’ve done a video instead, which is much more interesting. Here’s how it works.
166
605000
5000
10:22
Electrons are fuzzy creatures, and they can jump across gaps,
167
610000
4000
10:26
but only at equal energy. If the energy differs, they can’t jump.
168
614000
5000
10:31
Unlike us, they won’t fall off the cliff.
169
619000
2000
10:33
OK. Now. If something absorbs the energy, the electron can travel.
170
621000
7000
10:40
So here you have a system, you have something --
171
628000
3000
10:43
and there’s plenty of that stuff in biology --
172
631000
1000
10:44
some substance giving an electron, and the electron tries to jump,
173
632000
5000
10:49
and only when a molecule comes along that has the right vibration
174
637000
4000
10:53
does the reaction happen, OK?
175
641000
2000
10:55
This is the basis for the device that these two guys at Ford built.
176
643000
4000
10:59
And every single part of this mechanism is actually plausible in biology.
177
647000
7000
11:06
In other words, I’ve taken off-the-shelf components,
178
654000
1000
11:07
and I’ve made a spectroscope.
179
655000
2000
11:09
What’s nice about this idea, if you have a philosophical bent of mind,
180
657000
4000
11:13
is that then it tells you that the nose,
181
661000
2000
11:15
the ear and the eye are all vibrational senses.
182
663000
3000
11:18
Of course, it doesn’t matter, because it could also be that they’re not.
183
666000
4000
11:22
But it has a certain --
184
670000
1000
11:23
(Laughter)
185
671000
2000
11:25
-- it has a certain ring to it which is attractive to people
186
673000
4000
11:29
who read too much 19th-century German literature.
187
677000
4000
11:33
And then a magnificent thing happened:
188
681000
1000
11:34
I left academia and joined the real world of business,
189
682000
6000
11:40
and a company was created around my ideas
190
688000
3000
11:43
to make new molecules using my method,
191
691000
3000
11:46
along the lines of, let’s put someone else’s money where your mouth is.
192
694000
4000
11:50
And one of the first things that happened was
193
698000
4000
11:54
we started going around to fragrance companies
194
702000
2000
11:56
asking for what they needed, because, of course,
195
704000
4000
12:00
if you could calculate smell, you don’t need chemists.
196
708000
2000
12:02
You need a computer, a Mac will do it, if you know how to program the thing right,
197
710000
4000
12:06
OK? So you can try a thousand molecules,
198
714000
4000
12:10
you can try ten thousand molecules in a weekend,
199
718000
3000
12:13
and then you only tell the chemists to make the right one.
200
721000
3000
12:16
And so that’s a direct path to making new odorants.
201
724000
4000
12:20
And one of the first things that happened was
202
728000
1000
12:21
we went to see some perfumers in France --
203
729000
2000
12:23
and here’s where I do my Charles Fleischer impression --
204
731000
2000
12:25
and one of them says, "You cannot make a coumarin."
205
733000
4000
12:29
He says to me, "I bet you cannot make a coumarin."
206
737000
3000
12:32
Now, coumarin is a very common thing, a material,
207
740000
4000
12:36
in fragrance which is derived from a bean that comes from South America.
208
744000
3000
12:39
And it is the classic synthetic aroma chemical, OK?
209
747000
4000
12:43
It’s the molecule that has made men’s fragrances
210
751000
4000
12:47
smell the way they do since 1881, to be exact.
211
755000
3000
12:50
And the problem is it’s a carcinogen.
212
758000
3000
12:53
So nobody likes particularly to -- you know, aftershave with carcinogens.
213
761000
6000
12:59
(Laughter)
214
767000
2000
13:01
There are some reckless people, but it’s not worth it, OK?
215
769000
3000
13:04
So they asked us to make a new coumarin. And so we started doing calculations.
216
772000
5000
13:09
And the first thing you do is you calculate the vibrational spectrum
217
777000
3000
13:12
of coumarin, and you smooth it out,
218
780000
1000
13:13
so that you have a nice picture of what the sort of chord, so to speak, of coumarin is.
219
781000
6000
13:19
And then you start cranking the computer to find other molecules,
220
787000
6000
13:25
related or unrelated, that have the same vibrations.
221
793000
3000
13:28
And we actually, in this case, I’m sorry to say,
222
796000
4000
13:32
it happened -- it was serendipitous.
223
800000
2000
13:34
Because I got a phone call from our chief chemist
224
802000
3000
13:37
and he said, look, I’ve just found this such a beautiful reaction,
225
805000
5000
13:42
that even if this compound doesn’t smell of coumarin,
226
810000
1000
13:43
I want to do it, it’s just such a nifty,
227
811000
3000
13:46
one step -- I mean, chemists have weird minds --
228
814000
3000
13:49
one step, 90 percent yield, you know, and you get this lovely
229
817000
4000
13:53
crystalline compound. Let us try it.
230
821000
2000
13:55
And I said, first of all, let me do the calculation on that compound, bottom right,
231
823000
4000
13:59
which is related to coumarin, but has an extra pentagon inserted into the molecule.
232
827000
6000
14:05
Calculate the vibrations, the purple spectrum is that new fellow,
233
833000
5000
14:10
the white one is the old one.
234
838000
1000
14:11
And the prediction is it should smell of coumarin.
235
839000
4000
14:15
They made it ... and it smelled exactly like coumarin.
236
843000
8000
14:23
And this is our new baby, called tonkene.
237
851000
5000
14:28
You see, when you’re a scientist, you’re always selling ideas.
238
856000
2000
14:30
And people are very resistant to ideas, and rightly so.
239
858000
4000
14:34
Why should new ideas be accepted?
240
862000
2000
14:36
But when you put a little 10-gram vial on the table in front of perfumers
241
864000
6000
14:42
and it smells like coumarin, and it isn’t coumarin,
242
870000
2000
14:44
and you’ve found it in three weeks,
243
872000
2000
14:46
this focuses everybody’s mind wonderfully.
244
874000
3000
14:49
(Laughter)
245
877000
1000
14:50
(Applause)
246
878000
6000
14:56
And people often ask me, is your theory accepted?
247
884000
4000
15:00
And I said, well, by whom? I mean most, you know -- there’s three attitudes:
248
888000
5000
15:05
You’re right, and I don’t know why, which is the most rational one at this point.
249
893000
4000
15:09
You’re right, and I don’t care how you do it, in a sense;
250
897000
4000
15:13
you bring me the molecules, you know.
251
901000
1000
15:14
And: You’re completely wrong, and I’m sure you’re completely wrong.
252
902000
3000
15:17
OK? Now, we’re dealing with people who only want results,
253
905000
3000
15:20
and this is the commercial world.
254
908000
2000
15:22
And they tell us that even if we do it by astrology, they’re happy.
255
910000
5000
15:27
But we’re not actually doing it by astrology.
256
915000
3000
15:30
But for the last three years, I’ve had what I consider to be
257
918000
2000
15:32
the best job in the entire universe, which is to put my hobby --
258
920000
5000
15:37
which is, you know, fragrance and all the magnificent things --
259
925000
2000
15:39
plus a little bit of biophysics, a small amount of self-taught chemistry
260
927000
5000
15:44
at the service of something that actually works.
261
932000
2000
15:46
Thank you very much.
262
934000
1000
15:47
(Applause)
263
935000
1000
ABOUT THE SPEAKER
Luca Turin - NoseBiophysicist Luca Turin studies the science of smell. He's the author of Perfumes: The Guide, and the subject of Chandler Burr's 2003 book The Emperor of Scent. His next project: developing an artificial nose.
Why you should listen
Biophysicist Luca Turin studies the science of smell. He's a leading proponent of the vibrational theory of smell -- the idea that when our noses pick up a scent, we're reacting to the vibrational properties of the molecule we're smelling. (This is in opposition to the shape theory of smell, which imagines smelly molecules fitting into conveniently shaped receptors in our noses.)
His work on olfactory reception and the prediction of what a particular molecule will smell like has led to an interesting line of work: designing new fragrances and flavor molecules, as part of the firm Flexitral, where he is CTO. Coming up, he'll be part of a team in Biomedical Engineering at MIT, headed by Professor Shuguang Zhang, to develop an artificial nose made with natural receptors as part of DARPA's just-launched RealNose program.
He's the author of several books on scent, including Perfumes: The Guide and The Secret of Scent, and is the subject of Chandler Burr's 2003 book The Emperor of Scent.
Luca Turin | Speaker | TED.com