DE2014529A1 - Optical signal processing device - Google Patents

Description

20U529

Patent attorneys.

Dlpl.-In-. R .. no: E "T. :: son.
Dip: .- _{:; 3Ö} . u. .. '^.,. a. ht

Dr.-lng. Π. ^; ru γ; -; jr.
β Munich 22, Stoinodorfstr. 19 293-15-532P. 25-3-1970

National Research Development Corporation

LONDON, S.W.l. (Great Britain)

Optical signal processing device

The invention. Relates to a device for optical Signal processing.

Determining the molecular weight of large molecules such as protein molecules is difficult and time consuming. The Sedimentationsgewiclit can be determined in an ultracentrifuge, while the molecular weight from the equation · Svedberg can be calculated, which states that the molecular weight divided equal to the Sediinentatiousgewicht is through the diffusion coefficient, but it is difficult to ₅ to measure the diffusion coefficient. In general, the most convenient and accurate way of determining molecular weight is to allow the material to settle in equilibrium in the ultracentrifuge as the concentration changes

29'5- (JK 331 i / 0'l) -lId-r (7)

BADORlGtNAL

20U529

can be used with the depth in the gravitational field, to determine the diffusion coefficient so that on this way the molecular weight can be measured. Such an approach is, however, mostly avoided because it is extremely time consuming.

It is therefore an object of the invention, in particular one Specify device that can perform relatively fast measurements of diffusion coefficients.

It is already known from the Doppi rradar, the detected one Represent the Doppler shift by generating a function of time such as a Fourier transform.

The invention makes use of the fact that when a beam from a strong coherent light source such as a Laser is used to transmit radiation through a piece of material can be scattered, which has particles, which are in it with statistically disordered velocities move, with the scattered radiation having a large number of simultaneous different Doppler shifts is detected and converted into a function of time which is the Fourier transform of the Doppler shift can. This function can be used to calculate the mean value of the statistically disordered speeds in the material will. The invention therefore makes it possible, in particular, to determine the diffusion coefficients of many substances to eat. The invention allows the rapid measurement of the molecular weight of many substances as described in the Biology and chemistry occur.

The invention generally provides an apparatus for optical signal processing that has a

0098 ^ ^{1/1674 ßAD 0RIGlNAL}

2 0 U 5

Detector for detecting radiation with a frequency distribution which consists of a multitude of simultaneous, different Doppler shifts from the frequency of a beam of coherent radiation, and a device for generating a time function has j those representative of the frequency function of the detected radiation is.

The time function can be the Fourier transform of the Be the frequency function of the simultaneous, different Doppler shifts, and from this Fourier transform the values of e.g. the diffusion coefficients can be calculated.

"Optical" here also includes the infrared and ultraviolet Radiation area ..

The invention is explained in more detail with reference to the drawing. Show it:

1 shows the block diagram of an exemplary embodiment the device according to the invention. optical signal processing;

Fig. 2 several trains of waves over time to the theoretical To explain the basis of the invention; and ·

3 shows the block diagram of a detector and a digitally clipped! Correlators.

Fig. 1 shows the block diagram of an embodiment

009841/1674

BATH

20U529

example of the device for optical signal processing. A transparent container 1 contains suspended matter Material whose molecular weight is to be determined. The material in container 1 is treated with a strong Radiation from a laser 3 irradiated. Light scattered by the material in the Beliälters 1 is collected and used by a Detector 5 is detected, the axis of which is inclined at an angle θ to the axis of the laser and its output signal is fed into a digitally clipped autocorrelator 7.

The device works as follows:

The frequency of the light scattered by the material in the container 1 is shifted by Doppler shifts which are characteristic of the Brown ¹ see molecular movement of the material in the container 1 in the frequency of the incident light from the laser 3. It follows that the mean Doppler shift is characteristic of the diffusion coefficient of the material in the container 1, so that the diffusion coefficient of the material in the container 1 can be determined by evaluating the spectrum of the scattered light. The digitally clipped autocorrelator 7 is used * to gain enough information from the spectrum of the scattered light to calculate the diffusion coefficient. This is done by the so-called autocorrelation, which will now be explained with reference to FIG. 2, where several wave trains are plotted against time in order to provide the theoretical basis of the device according to the invention

Let us start from a wave train E (t) that has a complicated frequency spectrum, for example

009841/1674 bad ORIGINAL

20U529

a wave train (a) of Fig. 2. If such a wave train is shifted in time by an interval, gives the wave train E (t + T). The autocorrelation function G (T) is now defined as:

G (T) = Γ °° E (t) E (t + T) dt,

J-CO '

and its physical meaning is _f that it represents the area under the curve F (t, T), the value of which at any point in time t is given by:

F (t, T) = E (t) E (t + Τ "). G (T) therefore depends on the value of the interval T.

Since it is impossible to track every wave train during a To measure infinitely long time, G (T) is used as the mean value for practical reasons <F (t, T)> defined by, F (t, T):

<E (t) E (t

However, the calculation of such a function requires the computation of F (t _t%) f ^for a large number

of values t, T
ο ' ^v ο

It has already been proven that if E (t) is replaced by the statistical telegraph function E _k (t) :.

E _k (t) = 1 for JD (t)> k, E ₁ Xf) = 0 for E (t) "^ k

IC

for any value of k, it follows:

09841/167 « _BA0 ontSINAI.

20U529

G _k (T) = <E _k (t) E _k (t ₊ T)> = § sin ' ¹ G (I),

provided that k is the mean E of E (t). The wave train (c) of Fig. 2 shows the function E _k (t) for k = 0. The replacement of E (t) by E (t) is cut off.

it

called den (clipping).

In practice this means that when the concentration of the material in the container 1 in FIG. 1 is low is, little light is scattered by the material and the detector 5 must be very sensitive, namely must respond to the individual photons emitted in the direction of the detector.

This means that the wave train to be subjected to an autocorrelation only has integer values and n (t) can be written, where η is equal to the number of photons in the detector 5 in the interval between (t - - Τ?) and (t + - τ) occur, where T is a small one Time interval is. Under these conditions, a normalized autocorrelation function g * (T) can be defined which is given by:

g * ²⁾ (T) = <n (.t) n (t + T)> / n ² , with η = <n>.

The truncated function n ^ (t) can be defined as:

n, (t) = 1 for n (t) ^ k
n _k (t) = 0 for n (t) <k,

n, (t) can be done with a simpler setup than n (t) will measure.

009841/1674

- 7 - 20H529

Two truncated forms of the normalized autocorrelation functions appear useful The first form g ^ ² '(T) is defined as:

oo

e ^{ _o ² J (T) = < _no (t) _no (t ₊ T)> / <nX ² ,

where n (t) and (t + T) are both truncated at 0. It can be shown that

for η ^ 1 applies.

( 2 V tr
The second form g ^ ^; (X) is defined as i:

= <n _k (t) n (t +1
It can be shown that

1 + η

where g (X) is the Fourier transform of the function η (U)):

exp (-

and n (Cü) is the spectrum of the photon counts luktuatxon; the corresponding relation for g ^ '(C) is:

.-T ₊ " \ _S O (T) i ² .

009841/167 U

20U529

So if k = η let, i. If the signal is cut off at the mean count rate, it is cut off Autocorrelation function equals correct choice when clipping a channel at k * O is approaching the truncated autocorrelation function becomes the correct one as η Ο approaches

Figure k is the block diagram of a detector and a digitally clipped autocorrelator in which the normalized autocorrelation function gj_ (T) (as defined above) is generated. A photomultiplier 11 feeds a pulse shaper 13 · The photomultiplier 11 and the pulse shaper 13 together form the detector 5 of FIG. -State reached, as well as in eighteen parallel memories 1 <? _ To 19 ₁₇ ·

Siebzahn coicide gates 21. * to 21-_ are connected upstream of the storing gates 19- to 19 * ₇ .

A clock unit 23 controls the blocking input of the blocking gate 15 and the bistable circuit 17 so that it ^{goes into the H} O "state. The output signal of the bistable circuit 17 is fed into the first stage 25 of a shift register 25, which is constructed that it is switched on in stages by the output signal of the clock unit 23. The shift register 25 has seventeen stages 25 to 25 ₁₇ , the output signals of which are parallel Z1.

allele to the coicide gates 2I ₁ to 2I ₁ - forwarded

The circuit just described works as follows:

0QS841 / 1674

Photons received by the photomultiplier 11 are converted into electrical pulses which can be processed by the pulse shaper 13, which is used to generate pulses of equal height, duration, rise and fall time, one pulse corresponding to a photon detected by the photomultiplier 11 . The number of pulses generated by the pulse shaper 13 during successive time intervals of length T is subjected to the autocorrelation as follows. First the time interval T is chosen so that the probability of receiving a pulse from the pulse shaper 13 in the time interval is small compared to 1. For example, if an average of 200 pulses / sec are received, the time interval can be of the order of magnitude of about 500 / usec. The clock unit 23 is then switched in such a way that it emits a pulse at the end of each time interval T.

At the beginning of the measurements, it is ensured (by a conventional device not shown in FIG. 3) that the memories 19, -v to 19 * _{7 are} cleared and all stages of the shift register 25 are set to zero. The first clock pulse from the clock generator 23 sets the bistable circuit 21 in its "0" state _o Any pulses received by the pulse shaper 13 during the first time interval T are stored in the memory 19 _O and nowhere else, since all coincidence gates 2I ₁ to 2I ₁ - blocked are held because all stages of the shift register '25 ¹¹ O are "". If any pulses occur in the time interval T, the bistable circuit will 1? Brought into its "1" state. At the end of the first time interval T, the clock generator 23 causes a further switching. of the shift register 25, so that the respective state of the bistable circuit 1? at the end

009841/1874

20U529

of the first time interval T is stored. The bistable circuit 17 is also ^{reset to its M} O ^w state by the clock generator 23, and it is further prevented by it that pulses from the pulse shaper 13 arrive during the advancement of the shift register 25 by the locking gate 15 being closed.

During the second time interval, the number of pulses received by the pulse shaper 13 is added into the memory 19q and, when the first stage 25- of the shift register 25 is in the "1" state, the number of pulses received by the pulse shaper 13 during the second time interval T is Pulses are stored in the memory 19 ₁ via the coincidence gate 21. This operational sequence lasts a total of eighteen time intervals T.

The contents of the memories 19 _Q to 19 ₁₇ # generated by this operating sequence can be seen from the following table, in which the columns denote the quantities that are stored in the individual memories at equivalent times in accordance with the individual rows. Therefore, the sum of all the values in each column gives the total number stored in each individual memory. In the table, η denotes the number of pulses that are recorded in the time interval T, and B the Bode function, which is defined by:

B = O for η = 0
χ χ

B = 1 for η = 1.
χ χ

00984 1/1 674

20 U 5

Time- ¹⁹ O 19, ' Storage ¹⁹ 2 ¹⁹ 3 • · ■ - · ¹⁹ 17 ⁿ 1 ■ ""; ■ ** - ■■■ '-. »3 ⁿ 3 ^B 2 n ₃ bi % ⁿ * ^B 3 ^a if ^B 2 n * ^B 1 % . * *. · ⁿ x ⁿ x ^B x-1 ⁿ x ^B x-2 ⁿ x ^B x-3 V ♦ "♦ ♦ ■. ⁿ 17 ⁿ 17 ^B i6 ⁿ 17 ^B i5 ⁿ 17 ^B i4 ⁿ 17 ^B 1 ^T 17

Each memory thus represents the difference T between two time intervals, ie the contents of the memory T iet equal to S ^η _χ Β _χ ^. which is proportional to the autocorrelation function g _o ² '(t) iet. Therefore, the quantities stored in the memories 19q to 19 _t ~ represent the autocorrelation function gv ² ) (T), which is plotted over time in steps of T, since η is very close to the cutoff level 0. koffcmt. The number of repetitions of the sequence consisting of the seventeen steps determines the accuracy of the autocorrelation function.

The contents of memories 19 _Q through 19 ₁₇ can be used to make a graphical display as shown in the lower half of FIG. 3. The autocorrelation function generated will generally be an exponential decay which is actually the Fourier transform

009841/1674

20U529

of the intensity fluctuation spectrum of the light that is scattered by the material in the container 1 in Fig, 1 » The fall time of the exponential function is related to the diffusion coefficient of the material in container 1 of FIG. 1 is linked in the following way: The observed autocorrelation function has the shape

<n (t) n (t + Γ)> - <η ή [i ₊ ^ exp (-2D | κ ² | Γ) J,

1 + n

1 + n

provided that the scattering is a statistically disordered process. This autocorrelation function is characterized by a decay time 1 / D _T , where D_ is the diffusion coefficient. Therefore, the diffusion coefficient can be calculated by plugging the data into the above equation. The mean value η n (t) is calculated from the total counted pulses stored in the memory 19 _O. K is then calculated as:

K = (4lT N sin ^ 0) / λ _o ,

provided that the particle speed of the material in container 1 (Fig. 1) is small compared to the speed of light, where N is the refractive index of the solution in which the material is suspended, and θ the scattering angle (ie the angle at which the laser beam is is deflected) as well as \ the wavelength of the laser radiation in a vacuum.

For low η, ^ n > n approaches; for higher values η a correction can be carried out, which depends on the statistics of the scattered light.

009841/1674

By linking the diffusion coefficient thus determined with the sedimentation rate, which can be determined in an ultracentrifuge in about 1 hour, the molecular weight of the material in container 1 of Figure 1 using the Svedberg equation given above be calculated. ,

009841/1674

Claims (2)

20H529 Claims (y device for optical signal processing with a detector for detecting radiation that is Doppler-shifted with respect to the frequency of an output light beam
Has frequencies, and with a device for generating
of time functions which are representative of the Doppler shifted frequencies, characterized in that the detector (5) responds to
Doppler-shifted radiation caused by the irradiation
a sample (i) by a strong source (3) more coherent
optical radiation is generated, and radiation which has a number of simultaneous different Doppler shifts (Fig. 1). 2. Apparatus according to claim 1, characterized in that the sample (i) is a piece of material having particles that move at statistically random speeds, and that the means (7) for generating time functions representative of the
Doppler shifted frequencies are a means of generating a function of time that is equal to or approximately
is equal to the Fourier transform of the statistically unordered Doppler shifts and has a device for obtaining data from the time function,
which are representative of the mean of the statistically disordered velocities of the particles in the material
are (Fig. 1, 3). 3 · Device according to claim 2, characterized in that the device for generating time functions, which are representative of the Doppler-shifted frequencies, a digitally cut off; its autocorrelator (7) is (Fig. 1). 009841/1674 Blank page