High Performance, Low Cost…

Motivation for a Low Cost Polarimeter

Recognition of dextro- and levo-rotatory compounds is carried out using a polarimeter. While fully automatic instruments have all but superceded earlier manual instruments, the principles underlying the instrumentation used for polarimetry have changed remarkably little over the long history of the technique.

Measurements require the use of monochromatic light, as the optical rotations displayed by the enantiomers of a chiral sample are wavelength-dependent. Almost all early instruments used a sodium vapour lamp emitting the Na D lines for this purpose.

In addition, it is also necessary for the light be polarized prior to entering the sample. In a common arrangement, an “optical null” position is sought by moving a second, analyzing polarizer located after the cell so as to bring the light intensity on a suitable detector to a minimum. Commercial instruments that adopt this principle typically use high quality Glan-Thompson polarizers that can achieve extinction ratios of 105:1, a photomultiplier tube for sensitive optical detection and a high precision rotation stage with optical feedback/encoding to measure optical rotations down to the millidegree level.

For measurements of really tiny optical rotations (typically when sample is extremely precious - perhaps only mg quantities are available) a completely different approach is adopted. In one such arrangement a special flow cell is inserted between crossed polarizers. Modulation of the polarization of the light traversing the cell using a magnetic field (Faraday effect) and phase sensitive detection of the break-through light has been used to measure extremely small optical rotations (at the microdegree level).

As with all analytical instrumentation there is a tradeoff between performance and cost. Commercial polarimeters tend to be extensive, many costing in the $5k - $20k range. In this work we are striving for the very best achievable performance in a microprocessor-controlled instrument with an LED source, a solid-state detector and a very low cost polarization analysis system employing inexpensive sheet polarizers.

The ready availability of LED’s emitting at a wide range of wavelengths opens up interesting possibilities. For example, the optical rotatory dispersion (ORD) of a sample measures the change in a sample’s optical rotation as a function of wavelength. For most optically active molecules the optical rotation increases as the wavelength decreases, resulting in potentially better detection limits with blue and violet light than with yellow light (e.g the Na D lines at 589 nm).

The instrument that I’ve developed and that’s described next has numerous applications – for example, in the determination of enantiomeric purity, for quantitative analysis (since the optical rotation of a chiral sample is proportional to its concentration), and in kinetic studies where chemical processes result in changes in optical rotation over time - for example, acid-catalyzed sucrose hydrolysis, or glucose mutarotation.

Aside : Sucrose Hydrolysis

Cleavage of the glycosidic linkage in sucrose results in the formation of two new sugars - fructose and glucose.

Each of these 3 sugars is optically active, but to differing extents. Also, while sucrose and glucose are dextrorotatory, fructose is levorotatory.

In the chemical transformation shown opposite, the optical rotation therefore changes as the reaction proceeds - see the Results page to see how this reaction can be monitored by polarimetry.
Alt image
Stacks Image 6438

PolaProp - a Propeller-Based Polarimeter

The Propeller polarimeter is implemented on a small PCB. Only two external connections are required - (i) a DC plugpack input to power a stepper motor and (ii) a USB connection to control and take data.

The instrument measures the optical rotation of an optically active sample by monitoring the light intensity when linearly polarized light undergoes sinusoidal modulation by passage through a rotating-disk analyzing polarizer.

A light-to-voltage converter measures this modulation and after digitisation, the observed phase shift between blank and sample waveforms is determined by advanced signal averaging and curve-fitting techniques.

The polarimeter PCB contains stepper motor drive electronics, optical detection electronics and a USB port that is used to download waveforms to a PC for subsequent processing. A LabVIEWTM vi logs the optical rotation over time, allowing the instrument to be used for kinetic studies (such as the classic hydrolysis of sucrose experiment).

Included with the system is a source block fitted with an LED (normally yellow - to closely match the wavelength of the Na D-lines –589 nm) and a linear polarizer.

The source wavelength can be easily changed to take advantage of higher analytical sensitivity when measurements are made at shorter wavelengths.

PolaProp in Operation

The polarimeter's electronics are mounted on a 55x55 mm printed circuit board (PCB) that sits inside a detector housing that forms one end of the instrument.

The stepper motor is mounted external to this housing so that when assembled, the motor’s shaft passes through the clearance hole in the center of the PCB.

The analyzing polarizer is fastened to a bushing at the end of this shaft so that it spins approximately five millimeters clear of the detector. A 1 mm diameter hole in the front of the detector housing admits light into the detection system.

The LED source and fixed analyzing polarizer are located at the opposite end of the instrument, with the sample cell located in-between.

Angle-Tuned Interference Filter

In this version of the polarimeter a yellow interference filter is mounted in a precision rotation stage in front of the source LED. This allows the wavelength of light passing through the sample cell and onto the detector to be adjusted.

Using quartz calibration plates with accurately known optical rotations at the Na D line wavelength, the filter’s angle can be set to ensure that the polarimeter operates at the same wavelength and therefore gives the same results as an instrument with a Na lamp.


The instrument performs a single optical rotation measurement every 200 msec. 32 point signal averaging results in a minimum detectable optical rotation of 0.0050 at the 3σ level. Using an interference filter to carefully adjust the source wavelength by tilt-tuning, the absolute accuracy of such measurements can be improved to ~ 0.010, while long-term drift over an 8 hour period is typically less than 0.010.

Polarimeter Block Diagram

A schematic diagram of the Propeller-based polarimeter instrument - refer to the following section for details of operation.

Technical Description

The polarimeter is controlled by a P8X32A-Q44 Propeller™ chip (Parallax), a 32 bit processor with a novel architecture that supports true multi-processing via 8 internal “cogs”. Conveniently for instrument control applications, these cogs each have access to a shared set of 32 I/O pins and also to a 32kB internal memory space.

In this instrument one of the Propeller’s cogs is used for data acquisition and for receiving/transmitting data from/to a host personal computer (PC) running a LabVIEWTM vi. A second cog drives a stepper motor that has an analyzing polarizer (AP) disk affixed to its spindle. These two cogs communicate via a digital I/O pin that allows precise synchronization of these two operations.

A schematic diagram of the polarimeter is shown above. The polarimeter’s optical bench consists of a light emitting diode (LED), a fixed defining polarizer disk (DP), a collimating lens (L), an analysing polarizer (AP) disk and a cell holder that sits in the light beam path between the two polarizers. The analyzing polarizer disk is rotated at a constant rate by a stepper motor drive IC (ULN2003).

A TSL250 light-to-voltage converter (Texas Advanced Optoelectronic Solutions, Inc) monitors the light intensity emerging from the analyzing polarizer (AP) and the voltage readings are digitised by an analog-to-digital converter (ADC). As the analyzer disk spins it produces a sinusoidally modulated optical signal on the detector that is first stored into Propeller memory and then uploaded to a LabVIEWTM vi for post-processing to recover the optical rotation (see below).

The stepper motor is a 7.50 per step, 48 steps per revolution type (Changzhou Delilai Electrics type 42BY48B04). It is operated in half-stepping mode by driving the motor’s four coils in the appropriate sequence using four digital lines on the Propeller chip. A ULN2003 motor driver IC supplies the necessary drive current.

Software running in the motor control cog causes the motor to make one full revolution after 96 half steps that are precisely timed to each take 2.000 msec, for a total of 192 msec per revolution. The motor control cog generates a synchronization pulse on one of the Propeller’s I/O lines that signals the data acquisition cog (which polls this pin) to start acquiring data at precisely the same point in each rotational cycle of the stepper motor.

On receipt of this synchronization signal the data acquisition cog takes 2048 readings at a 10 kHz sampling rate from a 16 bit ADC (LTC1865) and co-adds these readings into a portion of the Propeller chip’s internal memory. Each scan thus takes 204.8 msec. After summation averaging over a series of (typically) 4 scans, the data acquisition cog then transfers the block of 4096 data bytes over a USB interface to a LabVIEWTM vi running on a host PC. The USB interface is implemented using an FT245RL chip (FTDI).

Optical Rotation Algorithm

The data record received by the LabVIEWTM vi after each data upload consists of slightly more than two full cycles of a sine wave since each 900 rotation of the stepper motor corresponds to a half cycle of optical modulation. The vi then fits the dataset to the functional form I(t) = Asin(ωt+φ)+B using LabVIEWTM’s non-linear least squares curve fitting algorithm. There are four parameters in this function, but since the stepper motor rotates the analysing polarizer at a constant speed - one (ω) is treated as a constant - leaving the other three (A, φ and B) to be reported by LabVIEWTM.

The algorithm for extracting the optical rotation is as follows :

For every measurement the sign of the amplitude term A is first examined and if this is negative, π is added to φ. Next, φ is normalized into the range – π … π. If φ is negative we continually add 2π until φ is in range. Similarly, if φ is positive we continue subtracting 2π until φ is in range. A phase measurement determined in this manner is required for both a blank (φb) and a sample (φs).

The optical rotation (in degrees) is then computed via θ = 90*(φsb)/π 0. There is one further consideration when using this method to ensure that the instrument returns the correct sign for measurements. A sample having a small positive rotation must have a more positive φs than φb. Experimentally, if this turns out to not be the case, the sign of θ should be reversed. By presenting a sucrose solution (which has a positive optical rotation) one can quickly check that the reported optical rotation has the correct sign.

In the event that a reading is expected to lie outside the range –900 <= θ <= 900 the measurement should be repeated at a different concentration - as is the usual practice in polarimetry.

Polarimeter LabVIEWTM Front Panel

Clicking on the right arrow near the top left of the panel starts data acquisition. The uploaded raw data is displayed as a blue trace (the sinusoidal trace on a yellow background) onto which the fitted curve is superimposed as a red trace, while the computed residuals are displayed as a yellow trace on a black background in a separate graph immediately below.

Optical rotations are calculated by the vi from the raw data and the results displayed in a text field, on an analog meter, and as a green trace on a black background where the readings are plotted against elapsed time. The exponential change in the optical rotation values over time as measured in an actual acid hydrolysis of sucrose experiment can be seen in the graph window at upper right.

A file “ON / OFF” button and Filename box allow the user to log the optical rotation values and elapsed time into a data file if desired.

The instrument is zeroed at the start of a new experiment by turning the ZERO SET switch to “ON”, waiting a few seconds until the instrument shows an Optical Rotation of 0.000 degrees, then turning the ZERO SET switch to “OFF”. This ZERO SET operation is carried out with a polarimeter cell containing a blank in-place and makes the phase measurement for B (ie φb) as discussed earlier.

Optical Rotation of Two Enantiomers : Carvones

The classic demonstration of optical rotation uses a clear container interposed between two sheets of polaroid material that is placed on an overhead projector. It is easily shown that when the container contains water, optical extinction is achieved when the polaroid sheets are placed at right angles; ie the polarizers are “crossed”.

If the liquid in the container is now changed to a concentrated sugar solution (keeping the polarizers exactly as they just were), some light is now observed to penetrate; and finding the point of beam extinction requires rotating the top polaroid sheet through an angle that measures the sugar solution’s optical rotation θ.

It is easy to relate this simple, yet rather perplexing demonstration (both are clear solutions !) to the instrument I’ve described above - since it employs exactly the same operating principle using polarized light to measure a sample’s optical rotation.

One can demonstrate that pairs of enantiomers rotate plane polarized light in opposite directions by filling two 10 cm polarimeter cells with samples of two naturally occurring enantiomers - one containing R-(-)-carvone (spearmint oil) and the other S-(+)-carvone (extracted from the caraway plant). The chemical structures of these molecules are shown below.
Carvones are naturally occurring compounds extracted from plants.

Carvone exists as two optical isomers – R and S – these compounds are mirror images of each other that are non-superimposable.

The two isomers rotate linearly polarized light in opposite directions.

The traces below show the averaged detector signal recorded by our polarimeter (fitted with a yellow LED) over a 2048 point = 204.8 msec interval for a blank (green), for S-(+)-carvone (purple) and for R-(-)–carvone (blue).
The traces opposite confirm that the constant rotational motion of the analyzing polarizer results in sinusoidal modulation of the light intensity as observed by a detector situated immediately behind the analyzer, as shown in the instrument simulation described here.

Superimposed on each of these traces is the best fit to the functional form I(t) = Asin(ωt+φ)+B ; these fits are in each case virtually indistinguishable from the raw data.

When the traces for each of the two samples are compared with that for the blank we see a sizeable phase shift on account of the sample’s optical rotation, as well as some significant optical absorption.

From the fitted phase values the optical rotations of the neat liquids are determined to be S-(+)-, +55.760 and R-(-)-, -56.910. The stated values for the neat liquids as provided are +540 and -610, respectively.

These results are consistent with the stated purities of the samples used in the experiment - 98% and 96%.