Physics of Smell
Physics of Smell
Physics of Smell
Abstract:
Quantum mechanics in biology is becoming increasingly important. Only in the last few years
experimental equipment became powerful enough to allow observation and measurement of
quantum effects in biological systems. New and very promising theories started to arise
showing biological processes in a very different way that we were used to. Among three main
quantum theories in biology to this time (magnetic sensing, photosynthesis and olfaction)
olfaction theory will be presented in this seminar. Basic anatomical and biological properties
of olfaction system in human will be reviewed. Historically important theories in this area will
be mentioned along with pro et contra experimental facts. Physical model with all important
parameters and approximations will be presented. In the end, there will be discussion and
future perspective.
Table of contents
1 Introduction ........................................................................................................................ 2
2 Biological background ....................................................................................................... 3
2.1 Anatomy and general properties ................................................................................ 3
3 The process of smelling...................................................................................................... 4
4 A short history of olfaction theories................................................................................... 5
4.2 Molecular shape (“lock and key” model)................................................................... 5
4.3 Vibration theory ......................................................................................................... 6
4.4 “Swipe-card” model ................................................................................................... 6
5 Physical model of vibration theory .................................................................................... 7
5.5 Assumptions and parameters of the model ................................................................ 7
5.6 The electron tunneling rate equation.......................................................................... 8
6 Discussion ........................................................................................................................ 10
7 Future perspectives and conclusion.................................................................................. 10
8 Appendix .......................................................................................................................... 11
8.7 Derivation of probability for the quantum oscillator to be in excited state.............. 11
9 References ........................................................................................................................ 13
9.8 Figure references ...................................................................................................... 13
9.9 Other references ....................................................................................................... 13
1 Introduction
The physical phenomena behind the most important human sensory systems, seeing and
hearing, are generally well understood also among non physicists as they involve very basic
physical phenomena. When going further to tasting and touch, the broader understanding of
the processes starts to faint to the point, when there is no “official” understanding at all.
Smell is still the elusive sense. To this time there is no widely excepted theory about how
does it work, what responses are triggered by which stimuli and how and why does this
happen. From the literature it is clear, that process could not be “chemical” as there is no
irreversible chemical reaction involved in the process – it has to be something “physical” [1],
which is why supposedly chemistry theme appears in a physics seminar.
This seminar will focus on the very same question: what is it about the molecule that makes it
smell and how is this property detected. The theory that will be presented is not widely
accepted yet but it already has some very important affirmations in measurements and
successful applicative results.
2 Biological background
2.1 Anatomy and general properties
The molecular and genetic aspect of olfactory system has been already determined by Richard
Axel and Linda Buck for which they won Nobel Prize for medicine and physiology in 2004.
For illustrative purposes and better understanding we will take a look on basics of how
olfactory system works.
Figure 2: Place where the odorant meets cilia of olfactory sensory neuron is shown. Function of olfactory
receptors is to discriminate odor molecules. The signal is triggered only with certain types of attached
odor molecules. If triggered, the electric signal is projected onto the olfactory bulb (OB) and then further
in brain.
Despite the very detailed knowledge about biological processes involved in olfaction there is
still link missing about the very first moment when the odor molecule enters receptor: how
receptor discriminate the molecule and what exactly releases G-protein unit. There have been
several theories about that and some of them will be presented in the next chapter.
It is also important to stress some other facts. Smell, as all organoleptic senses, is believed to
be subjective which could impose a problem in scientific evaluation. Luckily this is only
partly the truth. On receptor level we all “smell” the same. It is the interpretation of the signal
in the brain that makes the smell subjective. There exist measurement methods that make
measuring signals on receptor level possible (cell depolarization that is triggered by receptor,
fluorescent magnetic resonance imaging of the brain [5]), solving the problem of “subjective”
measurements. Contrary to first impression the smell can be approached scientifically.
Olfaction theories can be divided based on supposed discrimination factor into three different
categories.
Main supporters of this theory were Amoore (1948), Boelens (1974), Ohloff (1991), Rossiter
(1996) and Kraft et al. (2000). The theory is based on viewpoint that whatever happens in
nose can’t be much different from what happens in other similar parts of human body. Most
of other receptor systems in biology works by shape: enzymes, antibodies, ... so it is not that
unreasonable to presume the same mechanism for olfactory system [11]. However, the shape
theory fails in case of molecules that are structurally very different and smell the same or are
structurally almost exactly the same but smell considerably different (Figure 5). In a research,
75 structurally different molecules have been documented to smell of bitter almond [12].
Extensive research of this problem has been also made by C. Sell [13]. Apart from clear
difficulties with experimental affirmations, the shape theory also doesn’t propose any
mechanism that could release G-protein and start the detection process.
The idea about possible physical model that would include IR vibration modes of odor
molecules came from Lambe and Jaklevic paper about inelastic electron tunneling that made
inelastic tunneling microscopy (IETS) possible in nanoscale range [18]. For the model to be
feasible also with biological materials, some condition has to be considered [12]:
Contrary to continuous scanning in IETS, in biological system there would be pick up of
discrete energies as each type of GPCR receptor is tuned differently.
One electron may be enough to trigger the detection (contrary to electron flow in IETS).
Obviously, energy states of materials and temperature effect will be different from
normal IETS.
In IETS the inelastic electron tunneling flow is very faint in comparison with elastic
tunneling flow. In biological system the elastic tunneling should be suppressed
completely in order to avoid constant receptor “switch-on” and consequently sensory
overflow.
In order for the physical olfaction vibration
model to work also the following conditions
has to be fulfilled [9]:
While olfactory receptor is able to detect
many odor molecules, donor (D) and
acceptor (A) helices has to be able to return
elastically to initial state after detecting the
Figure 6: Schematic representation of GPCR odorant.
olfactory receptor in vibration mechanism To be able to detect smell within
physics model. Orientation is the same as in
Figure 3. X acts as electron source to donor (D) milliseconds (according to experimental
with Fermi level D. If odor molecule (M) has measurements) it is mandatory for A and D
the right vibration mode 0 the electrons can to replenish in the same time scale. The
tunell to the acceptor side (A) with Fermi level actual process of tunneling can be faster.
A where they release G-protein (G).
Energy splitting between D and A has to
correspond to 0 vibrational quantum in
odor molecule.
There are also some other important approximations and considerations that will be
introduced step-by-step in the text below.
The olfactory receptor has to be “resistant” to background noise, i.e. vibration modes of
molecules that are imbedded in the receptor. Certainly the most common is C-H stretch
vibration that occurs at 0,36eV (2911,3cm-1) [19]. Its coupling strength depends on root mean
square amplitude of vibration and effective charge. The first parameter we can estimate by:
1
x2 coth (1)
2 M 2 k BT
For M=1 for hydrogen (in atomic units) and energy 0,36eV the result is 0,076Å. The
educated guess for the distance between receptor helices (i.e. the distance between D and A)
is about 8Å [5]. The “noise” is small enough not to interfere with our model. The original
suggestion by Turin is that the frequencies around C-H stretches are “blind spot” (the
appropriate receptor type doesn’t exist) [20].
In the physical model we will use standard large body non-radiative theories (Huang-Rhys
theory and Marcus theory). We’ll assume that the process is slow enough to use perturbational
approaches (Fermi Golden Rule). Also very important observation is that environment of
olfactory receptor is hydrophobic which mean that it is very dry. We can safely assume that
the only moving charge is in the odorant molecule [5]. Although olfactory receptors can be
“tuned” to one or many odor molecule vibration we will simplify the number of “tuned”
vibrations to one per receptor only. We also assume that only one electron tunnels at a time.
We start the physical model building by approximating all molecules involved in the system
(D, A, M, environment) as set of quantum oscillators. The electron transfer from the initial
state D to the final state D is coupled to the vibrations of the odorant molecule M and of the
rest of the environment. Estimation of probability Pex for the odor molecule to be in excited
state can be done in the following way (see Appendix for derivation):
Pex exp (2)
k BT
With parameters for typical odorant molecule with vibrational quantum of 0 0, 2eV [5]
and normal human temperature ( k BT 0, 027eV ) we get probability of about 0,0006. We can
safely assume that all odorant molecules are in ground vibrational states. Based on what we
know about other similar biological systems we can assume that environment is a collection
of very low frequency oscillators that are only very weakly coupled to the transient electron
and impose no disturbance [5].
Next, we estimate key parameters for electron tunneling events. We have two types of
tunneling events: elastic and inelastic with timescales T 0 and T 1 respectively. As already
mentioned, the inelastic tunneling must surpass the elastic tunneling in order for our model to
be biologically feasible. We also have to estimate spectral resolution and at least crude
estimations of absolute rates for elastic and inelastic electron tunneling to be able to check the
model against data obtained experimentally.
After some mathematical formalism that exceeds the scope of this seminar we get the
equation for tunneling electron with or without vibrational excitation of the odorant
molecule [9]
1 2 2 n ( ) 2
t exp n (3)
D ,0 A,n 4 k BT 4 k BT
where q runs over all modes. Because the detailed atomic structure of olfactory receptor is not
yet known, can’t be accurately computed. Analogous to other similar biological systems it
is reasonable to assume that its value at room temperature will be below 30 meV [5]. Indeed,
this value is low enough for the physical model to work (Figure 9).
t in equation (3) is electronic matrix element for non-radiative transition and it is quite
possibly negligible when no odor molecule is present. It can be roughly estimated by:
v2
t (6)
M A
where v is hopping integral that will usually not exceed 0,1 meV. M corresponds to the
lowest unoccupied molecular orbit and A to the highest occupied molecular orbit and the
difference M A can be around 10 meV for most of the odor molecules [21]. Value of t can
then be approximated as 1 meV. We can cross-check our educated guess for t with
experimental data for C. vinosum that give us, when we put the data into the equation (3),
value of 2,4 meV for the matrix element t. Our estimation for t is close enough [22].
The plot with both values of n can be seen in Figure 9. For typical values (given in Figure 7)
we get T 0 87 ns and T 1 0,15ns for elastic and inelastic tunneling respectively. That
satisfy the condition T 0 T 1 and the physical model is biologically fissile, indeed.
Quantity 0 S t
Figure 7: Estimates values for computing characteristics times for elastic and inelastic electron
tunneling.
Figure 9:A plot time vs. reorganisation energy
( ). Thick line represents inelastic Figure 8: A Configuratio coordinate diagram
showing the initial state (D) on the left and the
(discriminating, T 1 , n 1 ) and dotted line final state (A) on the right. The final state has two
represents elastic (non-discriminating, T 0 , n 0 options: for elastic (n=0, thic line) and inelastic
(n=1, dotted line) tunneling.
) electron tunneling.
6 Discussion
Despite the vibrational theory of olfaction is still incomplete and it lacks systematic
experimental affirmations it already had great impact in industrial sphere.
Traditionally, fragrance companies are developing odors by organoleptic testing of large
amount of compounds. That is not only very time consuming but also very expensive and the
industry is constantly searching for improvements. To test the vibrational theory in its early
stage, the founder of the theory, Luca Turin, was asked by a fragrance company to synthesize
surrogate for the widely used odor compound called coumarine, which in its original form is
carcinogen. By using the theory results (and also by some lucky coincidence as he admits) he
managed to do it in just three weeks [11]!
There are still limit cases that vibrational theory can’t explain but it is clear that it has
potential that is worth exploring and is quantum based. For further physical modeling more
data are needed, especially regarding receptor detailed atomic structure.
G protein receptors are a family of receptors that are one of the most frequent targets of
various drugs. New theoretical findings about their discriminating mechanism might cause
revolution in pharmaceutical industry and new approaches in drugs healing [23]
Another very attractive property of smell is that very small quantities (1 part per billion in
human and much more in some animals) of unfamiliar molecules can be detected in
millisecond timescale without any previous training. With further theoretical understanding of
the mechanism revolutionary new methods of chemical analysis and sensory in general could
arise.
It appears that olfactory sensing really is, despite initial opposition of mainstream science,
based on quantum mechanics. In that sense it is similar to two other very important human
senses – hearing and sight – whose modus operandi is also known to be quantum mechanic.
Our interface to the outer world seems to be quantum mechanic! Despite very unintuitive (at
least to non physicists) we are intimately connected to its properties. In a way quantum
mechanics defines our awareness about ourselves and the world around us.
8 Appendix
8.7 Derivation of probability for the quantum oscillator to be in excited state
In this appendix we will derive the probability for the quantum oscillator to be in excited
state, i.e. equation (2). The presumption is that the oscillator is free and unperturbed. We start
with statistical mechanics according to which the probability for the system to be in the n-th
state equals to:
1
n exp ( F Wn ) , (A.1)
k BT
where k B is Boltzmann constant, T the temperature of the system and exp F partition
function, defined with statistical sum as
1
Wn (n ) , n 0,1, 2,... (A.3)
2
We derive the partition function with noticing that it represents infinite geometric series:
1 1
exp F exp (n ) exp exp n
n 0 2 2 n 0
(A.4)
1
exp (exp ) n
2 n 0
Infinite sum in (A.4) is geometric series for which we now that holds:
a0
a q
n 0
0
n
1 q
(A.5)
From (A.4) we see that a0 1 and q exp so the partition function is:
exp 12
exp F (A.6)
1 exp
To get the probability that the oscillator is in excited state, we have to sum equation (A.1)
over all such states i.e., indices n 1, 2,... Similar to (A.4) we get geometric series only that
time the starting index is 1.
1 exp
Pex exp (exp ) n
1
(A.7)
exp
1
2
2
n 1
We can transform equation (A.7) to be similar to equation (A.5) by introducing new, prime
index:
n n 1 n n 1 (A.8)
Pex 1 exp (exp ) n1 (A.9)
n1
We see that this time the parameters for infinite sum of geometric series are a0 exp
and q exp and the final result is as expected:
exp
Pex 1 exp exp
1 exp
(A.10)
exp
k BT
9 References
9.8 Figure references
Figure Reference
[1] R. H. Wright, The Science of Smell. George Allen and Unwin, London, 1964.
[2] D. T. Moran, J. C. Rowley, B. W. Jafek, and M. A. Lovell, “The fine structure of the
olfactory mucosa in man,” J. Neurocytol., vol. 11, no. 5, pp. 721–746, Oct. 1982.
[3] F. M. S. de Souza and G. Antunes, “Biophysics of olfaction,” Reports Prog. Phys., vol.
70, no. 3, pp. 451–491, Mar. 2007.
[4] P. Graziadei, Handbook of sensory physiology, chemical senses. Part 1. Olfaction.
Springer, New York, NY, 1971.
[5] J. Brookes, A. Horsfield, and A. Stoneham, “The Swipe Card Model of Odorant
Recognition,” Sensors, vol. 12, no. 12, pp. 15709–15749, Nov. 2012.
[6] H. Breer, “Olfactory receptors: molecular basis for recognition and discrimination of
odors,” Anal. Bioanal. Chem., vol. 377, no. 3, pp. 427–433, Oct. 2003.
[7] B. Malnic, P. A. Godfrey, and L. B. Buck, “The human olfactory receptor gene family,”
Proc. Natl. Acad. Sci. U. S. A., vol. 101, no. 8, pp. 2584–2589, 2004.
[8] B. Malnic, J. Hirono, T. Sato, and L. B. Buck, “Combinatorial Receptor Codes for
Odors,” Cell, vol. 96, no. 5, pp. 713–723, Mar. 1999.
[9] J. C. Brookes, “Olfaction: the physics of how smell works?,” Contemp. Phys., vol. 52, no.
5, pp. 385–402, Sep. 2011.
[10] J. C. Brookes, “Science is perception: what can our sense of smell tell us about
ourselves and the world around us?,” Philos. Trans. R. Soc. Math. Phys. Eng. Sci., vol.
368, no. 1924, pp. 3491–3502, Jul. 2010.
[11] “Luca Turin on the science of scent | Video on TED.com.” [Online]. Available:
http://www.ted.com/talks/luca_turin_on_the_science_of_scent.html. [Accessed: 12-May-
2013].
[12] L. Turin and F. Yoshii, “Structure-odor relations: a modern perspective,” Handb.
Olfaction Gustation, pp. 275–294, 2003.
[13] C. S. Sell, “On the unpredictability of odor,” Angew. Chem. Int. Ed Engl., vol. 45, no.
38, pp. 6254–6261, Sep. 2006.
[14] “BBC Horizon: A Code in the Nose,” Horizon, BBC, 1995.
[15] C. Roberts, “Psychophysical tests of the vibration theory of olfaction,” Chem. Senses,
no. 31, p. E74, 2006.
[16] A. Keller and L. B. Vosshall, “A psychophysical test of the vibration theory of
olfaction,” Nat. Neurosci., vol. 7, no. 4, pp. 337–338, Apr. 2004.
[17] S. Gane, D. Georganakis, K. Maniati, M. Vamvakias, N. Ragoussis, E. M. C.
Skoulakis, and L. Turin, “Molecular Vibration-Sensing Component in Human Olfaction,”
Plos One, vol. 8, no. 1, p. e55780, Jan. 2013.
[18] J. Lambe and R. Jaklevic, “Molecular Vibration Spectra by Inelastic Electron
Tunneling,” Phys. Rev., vol. 165, no. 3, pp. 821–832, Jan. 1968.
[19] T. W. G. Solomons, Organic chemistry, 6th ed. New York: Wiley, 1996.
[20] L. Turin, “A spectroscopic mechanism for primary olfactory reception,” Chem. Senses,
vol. 21, no. 6, pp. 773–791, 1996.
[21] J. Brookes, F. Hartoutsiou, A. Horsfield, and A. Stoneham, “Could Humans
Recognize Odor by Phonon Assisted Tunneling?,” Phys. Rev. Lett., vol. 98, no. 3, Jan.
2007.
[22] D. DeVault, Quantum-mechanical tunnelling in biological systems, 2nd ed.
Cambridge [Cambridgeshire] ; New York: Cambridge University Press, 1984.
[23] “BBC - Discovery: Quantum Biology.”