Doped sol-gel glasses as pH sensors

Materials Letters 13 (1992) 293-298 Nosh-Holland

Doped sol-gel glasses as pI3 sensors C. Rottman M.L. Kagan

a, M. Ottolenghi a, R. Zusman ‘, 0. Lev ‘, M. Smith ‘, G. Gong ‘, ’ and D. Avnir a a Institute of Chemistry, The Hebrew University ofJerusalem, Jerusalem 91904, Israel b Department of Environmental Sciences, The Hebrew University @Jerusalem, Jerusalem 91904, Israel ’ Star Techno~ogies~S~dbur~l,MA, USA Received 24 February 1992

A series of pH indicators were trapped in sol-gel porous glasses by polymerization of tetramethoxysilane in the presence of a surface active agent. The properties of these novel sensing materials including spectral shifts, shifts in the pH-sensing range, cycle repeatability, leachability, rates of response and isosbestic points are described. A prototype of a pH meter based on a pH-sensing glass was constructed.

1. Introduction Room-temperature polymerization of metal alkoxides enables the trapping of organic and bio-organic molecules in inorganic porous glasses. An extensive review of this fast-growing field has recently appeared [ 11. In general, one can categorize the appli~ations of these novel ceramic materials into two families: photo-responsive materials used for a wide variety of optical and information-recording purposes [ 2-41; and reactive materials in which the dopant molecules are capable of responding chemically to the environment through the pore space [ 5 1. Aithough this latter use is quite recent, the following achievements in three different domains demonstrate its considerable potential: (a) Solar light energy conversion. We have recently shown that suitably doped silica sol-gel glasses can be used for the photogeneration of radical ion pairs, extending the pair lifetimes from seconds (the maximal current state of art) to several hours [ 71. (b) Bioactive materials. For the first time, it has become possible to trap enzymes inside glass [ 1,891 (in contradistinction to the currently used techniques of trapping in plastics and of adsorption or covalent bonding to glass surfaces) while retaining their activity. (c) Chemical sensors. The potential advantages of 0167-577x/92/$

glass matrices for the construction of chemically sensing materials need not be spelled out. Following a feasibility study of the idea to use sol-gel glasses in which suitable reagents were trapped for sensing purposes [ lo], we demonstrated recently an unprecendented sensitivity in detecting Fe (II) in water, down to a level of 100 ppt ] 111. Other cation-detecting sol-gel glass prepared in our laboratories are described in ref. [ 121. A nitrate-detecting glass has been reported in a recent conference [ 131 and the detection of Ni( II) and Fe( II) is described in a recent review [ 141. Following our preliminary report [ lo], we now wish to summarize progress which we have made in the development of sol-gel glasses as pH-sensing materials using several classical (ground-state) indicators. The use of excited-state glass entrapped pH indicators was described by Kaufman et al. [ 151, Dunn and co-workers [ 14,161 and Chernyak et al. [ 17 1. Finally, one should mention the extensive work on the immobilization of pH indicators in organicpolymer matrices with some examples collected in refs. [ 18-2 11. The advantages of the sol-gel ceramic matrix over organic polymers such as transparency, chemical durability and better protectability of the trapped molecules, were described in great detail in previous publications [ 1- 12 1.

05.00 0 1992 Elsevier Science Publishers B.V. All rights reserved.

293

Volume 13, number 6

MATERIALSLETTERS

2. Experimental details A typicalprocedure. A mixture of 2.3 ml water, 2.5 ml methanol, 2.5 ml tetramethoxysilane (TMOS, Petrarch), 0.1 ml of a 0.02-0.04% aqueous or methanolic solution of the indicator (conditions were changed according to the nature of the dye), 0.5 ml of 1 x 1Oe3 M NaOH and 8.2 x 10V2 M (3.8% w/w) methanolic solution of cetyltrimethylammonium bromide (CTAB, Fluka), was poured into a 60 mm diameter petri dish, covered and left to gel for 6 days. Final drying of the glass was carried out in a 37 k 2’ C over for 4 days. The list of indicators trapped is given in table 1. The resulting glasses are transparent, crackfree monoliths (fig. I), weigh 1.35 f 0.05 g, and are of typical dimensions of 26 mm in diameter and 2.5 mm thick. Absorption spectra were recorded on a HewletPackard 8452A diode-array spectrophotometer. Color-change point of the indicators was determined by immersing glass pieces in solutions of pre-determined pH. Leachability tests were performed by immersing the glass in either water or a pH=O.Ol solution of HCI or a pH=2.0 solution df HCI. Fufl titration curves were determined by using a series of 14 solutions of pre-determined pH. Repeated cycles of pH measurements were determined for methyl orange, by immersing the glass back and forth at two

May 1992

pHs around the PKi, recording each time the maximum absorbance. Isosbestic points were determined by recording the full spectrum at time intervals of 5IO s.

3. Results and discussion The addition of surface active reagents to pH indicators in solution is known to shift both their wavelength of maximum absorption (&,,,,) and the pH range of color change. This well documented effect [22,23] is due to specific interactions between the surface active agent and the indicator. A similar behaviour was observed in our present glasses. Table 1 summarizes both A,,,,, and pH-range shifts for the glass entrapped indicators, and compares them to the solution behaviour with and without CTAB. An example of a characteristic spectral shift is shown in fig. 2. Except for phenolphthalein, all indicators (In) shift their detecting pH range, i.e. their effective PKi value to more acidic environments. This is due to competition between the two reactions: In+H,O+-InH++H,O, In+CTA++(In...CTA)+

(1) .

(2)

Table 1 Properties of the pH indicators Indicator

methyl orange methyl red bromocresol purple bromothymol blue phenol red cresol red phenolphthalein thymolphthalein thymol blue (acidic range) thymol blue (basic range)

pH range solution ‘)

3.1- 4.4 4.4- 6.2 5.2- 6.8 6.2- 7.6 6.4- 8.0 7.2- 8.8 8.0-10.0 9.4-10.6 1.2- 2.8 8.0- 9.6

Colour change ‘)

R-Y R-Y Y-P Y-B Y-R Y-R C-Pi C-B R-Y Y-B

pH range glass

0.0 - 2.0 2.5 - 4.6 2.00- 3.1 2.9 - 4.5 2.9 - 4.6 4.0 - 5.3 11.8 -14.5 12.00- 13.00 0 - 1.05 4.30- 5.7

A,,, (nm) solution ‘)

sol. +CTAB d,

glass

522-464 530-427 433-591 433-617 433-558 434-572 - -553 - -598 544-430 430-596

510-426(507-426) 515-415(514-416) 414-598 416-624(415-621) 416-572(416-568) 418-584 - -558 - -592 556-440 440-606

510-426 513-413 414-598 416-614 430-570 420-580 - -588 - -592 556-440 440-602

‘) Ref. [24]. b, R-red, Y-yellow, P-purple, B-blue, C-colourless, Pi-pink. For all pairs of entries in the table, the left-right colour letters refer, respectively, to the left-right values in each pair. ‘) These values, obtained by us, are in full agreement with the values published in ref. [ 241. d, In parentheses are given the values from refs. [ 22,231.

294

Volume

13, number

MATERIALS

6

LETTERS

May 1992

Fig. 1. Examples of sol-gel glass matrices in which pH indicators were trapped. From top (red), clockwise: methyl red, methyl orange, tropoeline 0, bromothymol blue, phenol red, bromocresol purple. (The center piece contains methylene blue, a redox sensor.)

L

OL

400

500

Wavelength

600

(nm)

Fig. 2. Spectral shifts of thymol blue: (1) in glass; (2) in 0.01 HCl+CTAB; (3) in 0.01 M HCl.

This situation requires a higher concentration of Hz0 in order to reach the color transition range, or higher concentration of OH - for the case of phenolphthalein: InH+OH-+In-

+HzO .

(3)

When thymolphthalein [ lo], phenolphthalein [ 10 1, bromocresol and methyl orange are trapped in the glass in the absence of CTAB, no PKishifts are

observed. Moreover, the spectral and pK, shifts observed here are very similar to the reported spectral shifts of the indicators in various surface-active-agent solutions [22,23] (table 1). Thus, we attribute the PKishifts in the glass mainly to the co-existence of CTAB molecules in the vicinity of the indicator in the silica cage, rather than to direct interactions of the indicator with the surface silanols. Fig. 3 shows the response-rate curves for two indicators. In general, the glasses react faster (minutes) as the pH is farther from neutrality (to both directions), and very slowly ( z 1 h) at pH close to 7. This trend is a reflection of the several orders of magnitude decrease in the concentration of the analysate (H+ or OH-) from the former case to the latter. Depending on the trapped dye and on the pH of the solution, various rates of leaching have been observed, from essentially no leaching, through very slow leaching, to fast leaching. An example of the latter is phenolphthalein: At the very high pH values employed for that indicator, the glass itself begins to 295

Volume

13, number

6

MATERIALS

ii/----

0.150~ 0

50

100 Seconds

150

200

0.80

b

d 0.68

I

2 0.56 p 0.44 3

0.32 0.20 0

100

200

300

400

500

600

700

Seconds Fig. 3. The response rate of (a) phenolphthalein orange:(1)510nm;(2)426nm.

0.80,_

I

I

I

I

and (b) methyl

I

I

I 50

I 100

I 150

I 200

I 250

I 300

1 350

under

various

Hours of entrapped

methyl

orange

dissolve, so that beyond 15-30 min, this glass-trapped indicator cannot be used. On the other hand, negligible leaching over 350 h is observed for methylorange glass immersed in water (fig. 4). For the same glass, leaching was very slow at pH= 2.00 (fig. 4), and moderate at pH = 0.0 1. The leachings obey firstorder kinetics with k( pH = 2.0) = (6.0 ? 0.1) X 10P4 h-i and tl,Z(pH=2.00)= 1165 h, and k(pH=O.Ol) =(2.9?O.l)xlO-3 h-i and t,,,(pH=0.01)=239 h. Yet it should be noted that these rates mean that within typical measurement time of 10 min, only 0.0005 of the initial dye leaches out at the low pH. For many practical purposes this is acceptable. 296

Reversibility and durability were investigated with the methyl orange doped glass. It was subjected to 12 cycles of transition between pH = 0.0 1 (detection at A,,,=510 nm) and pH=2.05 (&,,,,=426 nm). Readings were taken after 15-20 min at the low pH and after 25-30 min at the high pH. The results are shown in fig. 5. While readings at pH=O.Ol remain constant, a slight decrease at pHc2.05 with cycle number may take place. It was also found that the trapped indicators behave similarly to homogeneous solutions in the sense of showing a transition between only two species, namely, in exhibiting an isosbestic point. Examples are methyl orange as a representative of the azo indicators (fig. 6a) and bromocresol purple as a representative of the (sulpho)phthaleins (fig. 6b). The same applies to the titration curves which retain their regular shape: fig. 7a shows it for methyl orange in the glass and in solution, and fig. 7b shows the results for phenolphthalein. The shifts in the transition range are evident, and were discussed above.

4. Prototype of pH sensor

G 0.35

Fig. 4. Leaching conditions.

May 1992

1

$ 0.65 t 1 5 0.50 0)

0.20 1 0

LETTERS

A prototype sensor was developed based on the principles discussed above. As shown in fig. 8, the sensor operated in transmission mode utilizing a white light source, optical fibers for conducting light, and a spectrally sensitive detector. The pH-sensitive material was fabricated by the standard method, except that Triton-X was used instead of CTAB. A standard universal indicator was incorporated in the 0.8 ”

0.41

0

”

0

7 I 2

n

0

n

’

I

4

s

Meosurment

0

”

I

6

0

”

8 1 1 1 8

IO

r

I -I

12

Number

Fig. 5. Reproducibility of color reading of entrapped methyl orange over 12 cycles of immersion in solutions of pH = 0.01 ( ( 0 ) recordedat510nm)andpH=2.05((0)426nm).

Volume

13, number

MATERIALS

6

a

May 1992

LETTERS

Incandescent

0.8 Optical

fiber

0.6 0.4

Sotgcl

containing

pHindicator 0.2 500

450

Wavelength

( nm) b

output Fig. 8. Schematic

400

of the prototype

pH sensor.

600

500

Wavelength

diagram

(nm)

Fig. 6. lsosbestic points obtained from entrapped Methyl orange and (b) bromocresol purple.

indicators.

(a)

11

12

ptl of solution Fig. 9. The response

PH

PH

PH

Fig. 7. Titration curves in (a) glass and (b) solution. Top: methyl orange. Bottom: phenolphthalein.

glass, as described above. The material exhibited a transmission edge, transmitting 70% at wavelengths above 600 nm and 10% below this cutoff. A I mm thick section of material was formed for inclusion in the prototype sensor, and was fixed between two pieces of 1 mm diameter optical fiber. An incandescent bulb served as the broadband source. The best sensitivity was found to be at the high pH range. For

of the sensor at the high-pH

zone

that range, a spectrally sensitive two-color detector, with broad overlapping detection bands peaking at 550 and 620 nm was used. The actual signals are convolutions of the transmission spectrum and the wavelength dependence of each detector. The ratio of the intensities at the two detectors is a quantitative measure of the color change. This ratio is also independent of source intensity, obviating the need for optical intensity regulation. An analog ratio circuit was used to perform the required computation in real time to within 1% accuracy. The prototype sol-gel pH sensor was tested in a stirred solution to which sodium hydroxide solution was added at 10 min intervals. The sensor was found to stabilize within 2 min. The results are shown in fig. 9. In the pH range from 7 to 10, the sensor output changed very little, remaining close to 6 V. The sensor responded most dramatically in the pH range 297

Volume 13, number 6

MATERIALS LETTERS

IO to I 1, showing again that a change in the range of sensitivity of an indicator occurs upon entrapment in the matrix with a surface active agent. At this range, the transmission markedly decreased at the longer wavelengths and somewhat increased at the shorter wavelengths. The total sensor output change was 1.4 V over the pH range 10 to 11.3. These and other chemically reactive sol-gel glasses are world-wide patent-pending.

Acknowledgement

Supported by grants from the Israel National Council for R&D and KFK, Karlsruhe, Germany; the US Army Research, Development and Standardization Group (UK); the Krupp Foundation; the Szald Foundation. MO and DA are members of the L. Farkas (Minerva) Center for Light-Energy Conversion and acknowledge its support. DA is a member of the F. Haber Research Center for Molecular Dynamics.

References [ I] D. Avnir, S. Braun and M. Ottolenghi, in: ACS Symposium Series. Supramolecular architecture in two and three dimensions, ed. T. Bein (American Chemical Society, Washington, 1992) in press. [2] D. Levy,C.J. Semaand J.M. Gton, Mater. Letters 10 (1990) 470; D. Levy, S. Einhom and D. Avnir, J. Non-Cryst. Solids 113 f 1989) 137. (31 M. Lecomte, B. Viana and C. Sanchez, J. Chim. Phys. 88 (1991)39.

298

May 1992

[4] E.T. Knobbe, B. Dunn, P.D. Fukqua, F. Nisbida, R.B. Kaner and B.M. Pierce, Ceram. Trans. 14 (1990) 137. [ 51 A. Slama-Schwok, D. Avnir and M. Ottolenghi, J. Phys. Chem. 93 (1989) 7544. [6] A. Slama-Schwok, D. Avnir and M. Ottolenghi, J. Am. Chem. Sot. 113 ( 199 1) 3984. [7] A. Slama-Schwok, M. Ottolenghi and D. Avnir, Nature 355 ( i 992) 240. [8 ] S. Braun, S. Rappoport, R. Zusman, D. Avnir and M. Ottolenghi, Mater. Letters 10 ( 1990) 1. 191 S. Braun, S. Rappoport, S. Stelzer, R. Zusman, S. Druckman, D. Avnir and M. Ottolenghi, in: Biotechnology: bridging research and appii~tions, eds. D. Kamely et al. (Kluwer, Dordrecht, 1991) pp. 205-2 18. [IO] R. Zusman, C. Rottman, M. Gttolenghi and D. Avnir, J. Non-Cryst. Solids 122 ( 1990) 107. 1110. Lev, B. Iosefson-Kuyavskaya, I. Gigozin, M. Ottolenghi and D. Avnir, Fresenius J. Anal. Chem. ( 1992) in press. 121 B. Iosefkon-Kuyavskaya, I. Gigozin, M. Gttolenghi, D. Avnir and 0. Lev, J. Non-Cryst. Solids ( 1992), in press. 131 V. Raman, O.P. Bahl and V.K. Parashar, in: Abstracts of the 6th International Workshop on Glasses and Ceramics from Gels, Sevillia, September 199 1,abstract BP26. 141 B. Dunn and J.I. Zink, J. Mater. Chem. 1 ( 199 1) 903. 15 ] V.R. Kaufman, D. Avnir, D. Pines-Rojanski and D. Huppert, J. Non-Cryst. Solids 99 ( 1988) 379. [ 161 J.C. Pouxviel, B. Dunn and J.I. Zink, J. Phys. Chem. 93 (1989) 7544. [ 171 V. Chemyak, R. Reisfeld, R. Gvishi and D. Venezky, Sens. Mat.2 (1990) 117. [ 181 M. Cruz-Monero, M. Jimenez, C. Perez-Conde and C. Camara, Anal. Chim. Acta 230 (1990) 35. [ 191 S.H. Alabbas, D.C. Ashworth and R. Narayanaswamy, Anal. Proc. 26 f 1989) 373. [20] T.P. Jones and M.D. Porter, Anal. Chem. 60 ( 1988) 404. [ 2 I] G. Serra, A. Schirone and R. Boniforti, Anal. Chim. Acta 232 ( 1990) 377. [22] F.J. Drummond, F. Grieser and T.W. Healy, Chem. Sot. Faraday Trans. 185 (1989) 537,561. 1231 A.B. Zade and KN. Munshi, in: Surfactants in solution, Vol. 5, eds. E.K.L. Mittal and P. Bothorel (Plenum Press, New York, 1984 ) pp. 7 13-724. [24] E. Bishop, Indicators (Pergamon Press, Oxford, 1972).