Photoswitchable Stationary Phase Based on Packed Spiropyran Functionalized Silica Microbeads∗

The spiropyrans series of molecular photoswitches exist in two forms, a colorless, non-ionic, spiropyran form (SP) that is relatively passive in terms of ion-binding behaviour, and a zwitterionic, strongly colored (purple) merocyanine form (MC) that exhibits some degree of metal ion binding behaviour through a phenolate anionic site. Switching of the SP to the MC form (SP⇒MC) is easily achieved by exposing the molecules to UV light, whereas the reverse process, MC⇒SP switching, is similarly easily achieved by exposing the molecules to visible or green light. In certain cases, binding of a metal-ion guest (commonly divalent, M) to form an ion-MC complex (MCM) causes further changes in the visible absorbance, with consequent colour change, and therefore the system is inherently self-indicating, in terms of which form is present (SP, MC, MC-M) [1–7] (Fig. 1). Recently, the covalent attachment of a carboxylate SP derivate (SPCOOH, Fig. 1) to the surface of polystyrene and silica microbeads has been reported in the literature. The resulting SP-functionalised microbeads can be switched back and forth between the two forms using a 375 nm UV-LED (SP⇒MC switching) and a 430-760 nm white LED (MC⇒SP switching) [8, 9]. Furthermore, when the ‘activated’ beads (i.e., in the MC form) come in contact with Cu or Zn solutions, the microbeads undergo further spectral and visible colour changes, due to the formation of MC-M complexes. Subsequent exposure of the beads to illumination with a white LED causes the metal-ion guest to be expelled and the ‘inactive’ SP form is restored, ready for another ion-binding event. In contrast, in similar experiments involving other metal ions like Ca, no appreciable colour and spectral changes were observed. This light-modulated ion retention and release behaviour, coupled with visible indication of the bead state, opens the possibility of developing photocontrolled stationary phases that can be activated and deactivated using light, allowing flexible and intelligent use of the column. For example, a column could be held in an inactive (SP) form until ion-retention is required, and then activated using UV light, either wholly or partially. Subse-


I. INTRODUCTION
The spiropyrans series of molecular photoswitches exist in two forms, a colorless, non-ionic, spiropyran form (SP) that is relatively passive in terms of ion-binding behaviour, and a zwitterionic, strongly colored (purple) merocyanine form (MC) that exhibits some degree of metal ion binding behaviour through a phenolate anionic site.Switching of the SP to the MC form (SP⇒MC) is easily achieved by exposing the molecules to UV light, whereas the reverse process, MC⇒SP switching, is similarly easily achieved by exposing the molecules to visible or green light.In certain cases, binding of a metal-ion guest (commonly divalent, M 2+ ) to form an ion-MC complex (MC-M 2+ ) causes further changes in the visible absorbance, with consequent colour change, and therefore the system is inherently self-indicating, in terms of which form is present (SP, MC, MC-M 2+ ) [1][2][3][4][5][6][7] (Fig. 1).
Recently, the covalent attachment of a carboxylate SP derivate (SPCOOH, Fig. 1) to the surface of polystyrene and silica microbeads has been reported in the literature.The resulting SP-functionalised microbeads can be switched back and forth between the two forms using a 375 nm UV-LED (SP⇒MC switching) and a 430-760 nm white LED (MC⇒SP switching) [8,9].Furthermore, when the 'activated' beads (i.e., in the MC form) come in contact with Cu 2+ or Zn 2+ solutions, the microbeads undergo further spectral and visible colour changes, due to the formation of MC-M 2+ complexes.Subsequent exposure of the beads to illumination with a white LED causes the metal-ion guest to be expelled and the 'inactive' SP form is restored, ready for another ion-binding event.In contrast, in similar experiments involving other metal ions like Ca 2+ , no appreciable colour and spectral changes were observed.
This light-modulated ion retention and release behaviour, coupled with visible indication of the bead state, opens the possibility of developing photocontrolled stationary phases that can be activated and deactivated using light, allowing flexible and intelligent use of the column.For example, a column could be held in an inactive (SP) form until ion-retention is required, and then activated using UV light, either wholly or partially.Subse-quenty, the column could be deactivated, and the retained ions released into the mobile phase using white light.The column could then be reactivated for further use using the same procedure.In addition, simple observation or reflectance colour measurements would indicate the status of the column − whether it is active or inactive, and to what degree it is populated with metal ions.In this paper we demonstrate this principle by comparing the behaviour of SP-functionalised microbeads in a packed microcapillary column when exposured to Zn 2+ and Ca 2+ ions.The SP-functionalised beads were packed in-situ using a novel approach which is described below.
Plain silica microbeads (5 ± 0.35 µm diameter, 5% solid contents), butyl methacrylate, ethylene dimethacrylate, decanol, 2,2-dimethyl-2-phenylacetophenone, 2,2-dimethyl-2-phenylacetophenone (DAP, UV photoinitiator, λ max = 255 nm), calcium nitrate hydrate and zinc chloride FIG.2: Scheme of the experimental set up.From the left: peristaltic pump with a solution reservoir, connected to the capillary positioned inside the holder (middle), in which is inserted, perpendicularly to the capillary, the fiber optic.The optic probe presents two exits, one to the detector, the other to the two sources, a deuterium lamp, emitting in the UV region and an halogen lamp, emitting in the visible region that can be separately regulated.
were purchased from Sigma Aldrich (Ireland).SPfunctionalised silica microbeads were prepared according to a previously reported procedure [9].
255 nm UV-LEDs were purchased from Sensor Electronic Technologies, Ltd., USA.Polymicro transparent polytetrafluoro ethylene (PTFE) coated fused silica capillaries (100 µm i.d.) were purchased from Composite Metal Services Ltd (United Kingdom).Sample spinning was carried out using a ROTOFIX 32 centrifuge (Global Medical Instrumentation, Inc., USA.).Capillary flushing after the monolith synthesis was performed using an HPLC pump (Shimadzu, Japan).Sample additions into the packed capillary were performed using an RS485 peristaltic pump from Lambda Laboratory Instuments (Switzerland, Europe).
The purposely designed capillary holder was fabricated in black ABS with a Stratasys 3D printer (Fig. 2).The two parts of the holder (the probe and the capillary were designed using a standard CAD/CAM software package (ProEngineer).In order to evaluate the colour changes happening on the SP-functionalised beads packed into the microcolumn, reflectance UV-vis spectra were recorded using an ocean optic spectrometer (S2000) combined with a reflection probe which was connected to deuterium (215-400 nm) and halogen (400-1700 nm) light sources (Ocean Optics Inc., Eerbeek, Netherlands ).The reflectance data was referenced to a reflectance standard calibrated for 100% reflectance.

B. Synthesis of Monolithic frits and packing of microcolumn with SP-functionalised microbeads
The capillary packing of the spiropyran functionalized silica microbeads was carried out synthesising a short plug of monolithic polymer, which functioned as a retaining frit [11].Before polymerisation of the monolithic frit the walls of the polytetrafluoroethylene (PTFE) coated fused silica capillaries were pre-treated with a silanising agent to ensure that the monolithic frit would be well anchored to the walls [12].
After this pre-treatment procedure was completed, the monolithic frits were synthesised within the mold using UV light initiated in-situ polymerisation, following a similar procedure similar to that described by Abele et al. [13].Butyl methacrylate (BuMA) and ethylene dimethacrylate (EDMA) were chosen as the monomer and cross-linker, respectively, as they do not exhibit any ion-exchange properties and so will not interact with the sample metal ions.Therefore, any metal-ion binding behaviour will be due to the SP-functionalised beads and not the monolithic retaining frit.
For the frit synthesis, a solution of total volume of 200 µL was made up, consisting of 48 µL BuMA, 32 µL EDMA and 120 µL of decanol as the porogenic solvent.Approximately 0.77 mg of DAP was added to initiate the polymerisation.The mixture was sonicated to dissolve the initiator, purged to remove dissolved oxygen and finally filled into pre-treated PTFE coated capillaries by capillary action.Using rubber septa as a photo-mask a length of capillary 3-5 mm long was exposed to UV light from a 255 nm LED (forward current = 20 mA) to initiate the insitu polymerisation.The LED was placed perpindicular to the capillary at a distance of 1 mm and the polymerisation allowed to proceed for 1 h.After this time, the light source was turned off and the capillary flushed with methanol using a LC-10 AD HPLC pump, in order to remove any unreacted components of the polymerisation mixture.The resulting polymeric monolith typically has a pore size of around 1-2 µm, which can efficiently retains the 5 µm diameter SP-functionalised microbeads during the packing process, while still presenting a relatively low back pressure.
For the packing stage, a slurry of the packing beads was made up using a small amount of ethanol.A length of polytetrafluoroethylene tubing was filled with the slurry and connected between the outlet to the HPLC pump and the capillary containing the monolithic frit.The pump was turned on to allow the eluent to flow at 10 µL/min (this was reduced if the back pressure exceeded around 10 Mpa).In short the eluent simply pushes the beads from the tubing into the capillary column to produce the packed capillary column.When ca. 1 cm of the capillary length was packed with beads, the loop was removed and the capillary column reattached directly to the pump.Methanol was then flushed through for 30 minutes, as this gives a tighter packing of the beads and therefore reduces the likelihood of voids forming.

C. Capillary holder set up and reflectance measurements
The capillary holder (Fig. 2) consists of a probe case in which the reflectance fiber can be inserted.The optical fibre is held effectively in place by its tight fit to the specially designed upper part of the holder.The lower part has a narrow hole of 0.5 mm diameter in the middle and an extrusion at each side of the hole for securing the attachment to the probe case.When the two parts are attached together the capillary can simply be inserted by threading it into the orifice of the capillary case, while the optical fiber is positioned perpendicularly to the capillary at a fixed distance.This refelctance probe has 6 illumination fibres concentrically arranged around a single collection fibre.
The capillary was evaluated according to the following procedure: 1. Flush the capillary with the desired solution for 30 min at a flow rate of 0.5 ml/ hour.
2. Irradiate with the halogen lamp (visible light) for 3 min to promote MC⇒SP switching (ensures the beads are predominantly in the inactive SP form).
3. Turn the halogen lamp off.
4. Record the visible spectrum by turning on the halogen lamp for 1 sec.
5. Irradiate with the deuterium lamp (UV-light) for 3 min to promote SP⇒MC switching (ensures the beads are predominantly in the active MC form which can bind metal ions).
6. Turn the deuterium lamp off.
7. Record the visible spectrum by turning on the halogen lamp for 1 sec.
This procedure was repeated in the presence of three different solutions, pure ethanol, 10 −3 M Zn 2+ and 10 −3 M Ca 2+ , all as ethanolic solutions, and washing the capillary for 30 minutes with pure ethanol between each ion sample solution to ensure there is no carryover.

III. RESULTS AND DISCUSSIONS A. Reflectance Spectra from capillary
The use of the photopolymer frit to form the packed capillary is much more convenient than the conventional approach, which typically involves using preformed silica frits which are held in position with large stainless steel ferrules.In contrast, using the in-situ photopolymerised frits [11] means that these ferrules are not needed, which facilitates the positioning of external detectors such as non-contact conductivity, or (as in this case) reflectance spectroscopy, as the capillary can be simply inserted through the purposely designed holder and the reflectance spectra from the packed capillary column recorded.
Despite the fact that the area under illumination is small (ca. 100 µm , the main spectral features and associated colour changes can be clearly distinguished, although the quality of the spectral data decreases sharply below about 450 nm as the intensity of the halogen lamp declines. Exposure of the SP-functionalised beads to the deuterium lamp for three minutes in the presence of ethanol mobile phase leads to a clear change in colour, indicating that effective SP⇒MC switching has occurred (Fig. 3).The before and after reflectance spectra (Fig. 4) confirm that the colour change is due to an increase in absorbance centred around 560 nm, which is characteristic for the presence of the active MC form [9]. Subsequent exposure of the packed capillary to the halogen lamp for 3 minutes reverses this process and the beads revert to the inactive SP form.
In order to evaluate the effect of metal ions on the beads, 10 −3 M ethanolic solutions containing Zn 2+ and Ca 2+ , respectively, were sequentially pumped into the capillary for 30 min using the procedure described above.Following that, the beads were activated (SP⇒MC) by illumination with the deuterium lamp for 3 minutes and the reflectance spectra obtained.The results show that in the presence of 10 −3 M Zn 2+ , the MC spectrum changes dramatically, with a large decrease of the 560 nm band being clearly evident, along with the appearance of a new band centred around 520 nm (Fig. 5), which is consistent with the formation of the MC-Zn 2+ complex.In contrast, in the presence of 10 −3 M Ca 2+ , no spectral changes are evident (Fig. 6).When the capillary is subsequently exposed for 3 minutes to white light using the tungsten lamp, the beads return to the colorless SP-form, and the Zn 2+ guest is expelled into the mobile phase.
These results suggest that the capillary column packed with SP-labeled microbeads can selectively retain certain ions under photonic control.When the beads are in the SP-form, there is no retention behaviour and metal ions pass through the column.However, upon exposure to UV light, the beads switch to the active MC-form, which is indicated by the development of the purple color, and characteristic absorbance band at 560 nm.In this activated form, the beads will retain certain metal ions such as Zn 2+ , further spectral and colour changes occurring that are associated with the formation of MC-M 2+ complexes− in the case of Zn 2+ , the 560 nm band decreases in intensity and a new band appears at 520 nm.In contrast, Ca 2+ ions produce no discernable colour change, and the absorbance spectrum is essentially unaffected.Illumination of the ion-loaded beads with white light releases the ions back in to the mobile phase and the cycle can then be repeated.
On the basis of previous work, we can predict that other ions like Cu 2+ and Co 2+ will behave similarly to Zn 2+ , whereas other ions, like group I and group II metal ions will have no effect [6][7][8].

IV. CONCLUSIONS
These results suggest that we have achieved the first steps towards the creation of a photoswitchable stationary phase.A column has been produced through the packing of spiropyran modified microbeads in a microcapillary, using a monolithic polymer as a frit which enables a low pressure approach to be adopted.The SP-functionalised microbead column can be reversibly switched between inactive (non-ion binding) SP and active (ion-binding) MC forms using UV and visible light, enabling cetain ions to be retained and released under photonic control.The column is inherently self-indicating in terms of whether the inactive SP, active MC or ion-complexed MC-M 2+ forms are present, suggesting that simple diagnostic measurements can be integrated into the system to determine how effectively it is functioning.
In future work, we will examine the behaviour of the column towards other ions using similar approaches, and apply the system to more complex sample mixtures to further characterise the interaction dynamics of the system.

FIG. 3 :
FIG.3: Picture of the packed capillary under UV light irradiation: the microbeads side, activated to the MC purple form can be clearly distinguished from the polymeric monolith, which acts as a retaining frit that facilitates bead packing and flushing without any microbead leakage.

FIG. 6 :
FIG.6: Effect of 10 −3 M Ca 2+ ions in the capillary on the reflectance spectrum compared to the initial MC reflectance spectrum.In contrast to Fig.5, no significant effect is observed.