# Appendix

## A.1 Design aims

The SPM setup was designed and built with the following aims:

(i) The sample interface should be in the horizontal plane.

(ii) It should be possible to make angular scans.

(iii) The same part of the sample should be imaged during an angular scan.

(iv) It should be possible to use an AFM on the same sample simultaneously, also for measurements in water.

(v) It should be possible to use different wavelengths.

## A.2 Implementation

To have the sample interface in the horizontal plane, the optics are in the plane of incidence i.e. the vertical plane. They are mounted on two arms that can rotate around the same axis (see Fig. A.1(c)). These arms are driven by a stepping motor at the back side of the setup (see Fig. A.1(b)). The exact angle of incidence is determined at the axis by a high precision angle encoder. A half sphere was used as a prism to keep the imaged area at a fixed position while rotating the arms. Because the SPM uses internally reflecting light, the space above the sample on the prism is free. A platform was designed to allow an AFM to be placed over the sample, and do SPM and AFM measurements simultaneously. These measurements can even be done in an aqueous environment (e.g. to study protein / surface interactions). For most measurements a small HeNe-laser will be used as a light source, for measurements at other wavelengths a polarization preserving fiber can be used in combination with a bigger laser. A CCD video camera and a video digitizer were used to acquire and store SPM images.

Fig. A.1 (a) front view of the setup; (b) backside view; (c) front view without the AFM platform; (d) front view with AFM over the prism.

## A.2.1 Software

Software was developed such that all features of the setup can be controlled using an interactive computer program. The computer screen shows the live SPM video image with the menus and mouse cursor overlapping (see Fig. A.2). Via these menus the following functions can be performed:

(i) video control: images are continuously acquired but can also be frozen, saved to harddisk, etc.

(ii) polarization switching: the polarization of an external light source can be switched using a computer-controlled Pockels cell.

(iii) angle control: the angle of incidence is continuously measured with the angle encoder connected to the left upper arm. The stepping motor can be made to increment or decrement the angle or adjust it to a certain value.

(iv) cursor definition: a number of areas within the image can be defined for measuring their average intensity as a function of time or angle of incidence (e.g. for multisensor measurements).

(v) scanning: for measurements of the average intensity of the defined cursors as a function of time or angle of incidence

Fig. A.2 User interface for the software controlled operation of the setup.

## A.2.2 Hardware

Stepping motor: RS440-458 (Mulder-Hardenberg, Haarlem).
detent torque: 30 mNm. holding torque: 500 mNm.
step angle: 1.8 deg. ste angle accuracy: 5%.
rated voltage: 12 V. rated current: 0.6 A.
The parallel connection was used, and a corresponding driver was built.

ADC/DAC card: RTI 800 (Analog Devices, Norwood, MA, USA).
For conversions of analog and digital data, and digital communication.

Video digitizer: video blaster CT6000 (Creative Labs, Milpitas, CA, USA).
Digitizes the video signal in 8 bits per pixel. Uses the computer screen or any chosen part of it for displaying the live video images.

Angle encoder: ROC 417 (Heidenhain, Veenendaal).
Instead of using the commercially available read-out equipment for the encoder, the serial output signal with parity bit is read by the computer (using the ADC/DAC card for digital communication, and home-made software). The angle is encoded in 17 bits (resolution: 2.7 mdeg.) in gray code.

Spindle: diameter: M12, pitch: 1 (Jeveka, Amsterdam).

Rails with bearings: RSD3275 (Aalbers, Dedemsvaart).
This system is used to guide the part that connects the arms with the spindle.

## A.3 Angular accuracy

A simple formula can be derived for the angular movement as a function of the rotation of the vertical spindle:

$\cos\alpha =\frac{b^2-a^2}{2ab}\tag{A.1}$

where b is the distance along the spindle in between the hinge points in the middle, and a is the distance between the hinge points on the arms on which the optics are mounted (see Fig. A.3 (a)). The lower arms are a factor $$\sqrt{2}$$ longer than the upper ones with length a, and thus if b = a then $$\alpha$$ = 90°.

As can be seen in the calculated curve in Fig. A.3 (b) this construction makes it possible to rotate the arms in an almost linear way by rotating the spindle with the stepping motor. Since a movement of a hinge point of only 10 $$\mu m$$ corresponds to a change of the angle of about 10 mdeg., the accuracy of the hinges should be very high.

Experimentally, the measured angle was always found to be within 10 mdeg. of the calculated angle. Since the angle of incidence is measured directly at the incoupling (upper left) arm with an accuracy of 3 mdeg., this does not limit the accuracy of the determined angle.

Fig. A.3 Calculation of the rotation of the arms as function of the shift of the lower hinge point. (a) schematic of the arms; (b) calculation results, differing less than 10 mdeg. from the experimental results.