Colloid & Polymer Science
Colloid Polym Sci 268:337-344 (1990)
Spectrophotometric behavior of polyvinylpyrrolidone in aqueous and nonaqueous media (I) L. TOrker*, A. GOner**, E Yigit**, and O. Gfiven** * Middle East Technical University, Department of Chemistry, Ankara, Turkey ** Hacettepe University, Faculty of Engineering, Chemistry Department, Beytepe, Ankara, Turkey
Abstract: Electronic spectral behavior of polyvinylpyrrolidonesolutions in various media has been determined by UV-VIS spectrophotometry. A theoretical approach has been developed to explain the experimentally observed concentration dependent spectral behavior of polyvinyl pyrrolidone in aqueous and nonaqueous solvents. Increase in the concentration of the polymer or the addition of guanidine salts caused bathochromic shift. A similar concentration effect has been observed in nonaqueous media in the absence of guanidine salts. Key words: Polyvinylpyrrolidone solutions, spectrophotometric behavior, UV-spectra, solvent effects, salt effect.
Introduction
Polymers inter- or intra-molecularly associated with aqueous or nonaqueous solvents have received much attention in the last few decades [1-4]. Of these polymers, polyvinyl pyrrolidone (PVP) has certain interesting aspects. It is capable of forming a high degree of hydrogen bonding with aqueous and nonaqueous solvent systems in the presence or absence of cosolutes. PVP, a water soluble synthetic polymer with a repeating unit, the structure shown below, has a number of interesting properties [5] and frequently, it has been the object of many investigations in order to understand the interactions between small molecules and a polymer. PVP has the ability to bind reversibly to small
H2N)C* H2C--CH 2 K676
molecules forming association complexes which are important from several points of view, in chemistry, biochemistr36 and pharmaceutical sciences. Especially, the existing similarities between the chemical structures of PVP and proteins have led to proposals for PVP as a synthetic-polymer model for proteins
[6, 7].
Molyneux et al. [8] have investigated the interactions of PVP with small cosolutes in aqueous solutions taking into account hydrodynamical aspects of the binding process with several kinds of cosolutes (ionic, non-ionic, aromatic, aliphatic, etc.). Rothschild used IR, viscosity, and spinecho measurements and suggested a water bridge between pyrrolidone rings [9]. The possibility of forming hydrogen bonding and association complexes with various protic and aprotic polar solvents should affect the many observables of systems containing PVP, depending on the extent of the interactions. In the present study, the electronic spectral behavior of aqueous and nonaqueous PVP solutions under the influence of various perturbing effects has been investigated.
338
Colloidand PolymerScience,Vol.268. No. 4 (1990)
Experimental
EK= EOK+ , ~ c ~ C2p + 2 ~ CKp CKrrS#p ~ , (2) p p-~
Poly (N-vinyl-2-pyrrolidone)sample used in this study was a commercialproduct of BDH,England.It has an average molecular weight of 5.48 105 g/tool, determined in chloroformat 25 ~ by using light-scatteringmeasurements.The analar grade guanidine salts used (guanidine carbonate, chloride, and sulfate) were obtained from BDH and Eastman Kodak Co. The nonaqueous solvents (methyl alcohol,ethyl alcohol, chloroform,and acetic acid) were spectroscopic grade. The spectra were recorded by using a Hitachi 100-60model UV-VIS,double-beam spectrophotometerat room temperattrre. The possible interfering effects of inpurities were eliminated by dialyzing the aqueous polymer solutions through cellulose nitrate membranes for 6 and 11 days at room temperature. The same spectral behavior was observed for the dialyzed and undialyzed samples in each set. The spectra of all these samples were also recorded by using a Hitachi 150-20 double-beam UV-VIS spectrophotometer at the same temperature. No spectral differences were observed between the spectra of dialyzed and undialyzed polymer samples recorded by both of the instruments in the range of 190-340 nm.
perturbed
where CKp, CKa are the molecular orbital coefficients (eigenvectors) of centers p and or, respectively, in the molecular orbital ~K- The variations in Coulomb and resonance integrals are denoted by S a p and 6tip a , etc. [11, 12], whereas the pertorbed and unperturbed molecular orbital energies are indicated by EK and E~ respecfivel3a Obviously, in the structure of PVP molecule, the effective chromophore is the lactam carbonyl and it is the resonance hybrid of structures I and II.
\ C f
/N'< I
Theory Absorption of a photon of suitable energy reshuffles the electron shell of the absorbing molecule, e. g., viewed within the H M O model, one of the "loosely b o u n d " 7r-electrons (or non-bonding electrons) is usually raised from one of the molecular orbitals ~K to an antibonding orbital ~'T" Consider an absorbing system, M, and let EK~ and ET~ be its ground state occupied and unoccupied molecular orbital energies, respectivel~ such that the energy gap between them corresponds to Zr~x value of the spectrum of M. Since, hv = EK~ - E ~ AE = EK0 - E ~ , then the longest wavelength excitation can be expressed as ;bnax = A(1/AE~ + B ,
(1)
where A and B are certain correlation constants specific for the given system [10]. In the case of ions or polar molecules in solution, they should also involve solvation energy differences between the ground state and the excited state.
C/O-
II
4-
II
Since the lactam carbonyl is a + E-type substituent [10], nitrogen should possess a partial positive charge. Thus, oxygen is partially negatively charged. Therefore, PVP should solvate the anions effectively by hydrogen bonding. However, substituents or solvent effects influence the energy of the system if the charge distribution changes on electronic excitation [111. Suppose ,t,i and ,~- are two Zmax values of the same substance at two different concentrations, respectively. Consider merely the perturbations having first order character. Furthermore, neglect all the perturbations occurring in resonance integrals which m a y arise from conformational changes, etc. about the bond in between a n y adjacent centers p and or. Then Eq. (2) simplifies to EK = EOK+ ~ C2r hp p
(3)
ET= E~ + ~C2php p
(4)
for molecular orbitals ~K and 'FT, where hp is the Coulomb integral parameter [11] for the perturbed center p. The energy gap (standing for the electronic transition) corresponding to 2ma~ of the system is then expressed by
339
Tarker et al., Spectrophotometric behavior of polyvinylpyrrolidone
AE= AE ~ + ~ , ( C 2 - C 2) hp .
(5)
P
the mixed type terms involving perturbations in oxygen and nitrogen centers mutually in the multiplied form, one gets
Inserting P for the terms in the parentheses yields AE = AE~ + ~ P p hp. P
(6)
A(Pozlh o + PNzlhN) )!,j- 3,i = (AE0)2+AEo(pN(Nhj + Nhi) + Po(ohj + ohi)) , (11)
Assuming that the number and kinds of the perturbed centers remain the same but the magnitudes of the perturbations change as the concentration of the system varies from C i to Cj, then one obtains Eq. (7) by combining Eqs. (1) and (6).
/ 1
zlE ~ + ( ~ Pp hp)j
zlE ~ + ( Z
P
/
(7)
P
Keeping the structure of PVP and the chromophoric subunit present in it, in mind, it is possible to realize that the main types of perturbations affecting the electronegativities of the centers arise from the possibility of hydrogen bond formation (inter- or intramolecular) via nitrogen and oxygen atoms. As long as the polymer concentration is kept below a certain critical value, the above assumption should be valid, namely the extent of the hydrogen bonding should be the primary factor playing a role on the variation of Xm~x values as the concentration of PVP changes. Hence, Eq. (7) can be put into a more explicit form,
I 1 ;Lj- 2,i = A AE0 + (PNhN + V~
1
LIE~ + (PNhN +
Poho)i
where ztho and zlhN stand for the differences in the Coulomb integral parameters (heteroatom parameters [12]) at concentrations C i and Cj of oxygen and nitrogen. Obviously, zlh = -kX and Ahn = -1X (Eqs. (9) and (10)). Then inserting these values into Eq. (11) and replacing 3,i with ~ (Xmax value at the initial concentration of the polymer) and 3y with Xm~ one gets
A1X + A~ 3,,~x - B1X + Bo "
(12)
where
A 1 = (3,oZlE~ - A)(lP N + kP o) A0 =/1"0((Z~E0)2 + 2zl~
B 1 = AE~
+ Po ohi))
+ kP o)
B0 = (/IE~ 2 + 2~E~
+ Po ohi) 9
Equation 12 represents a homographic function (one branch of a hyperbola [13]. Mathemathically, as x ~ oo, )t,maxasymptotically approaches to A1/B 1.
t
(8) "
Now, suppose the Coulomb parameters for oxygen and nitrogen (oh and Nh) at two different concentrations can be expressable as
oh! = ohi + kX
(9)
nhj = Nhi + IX,
(10)
where k and 1 are certain constants, whereas X is a function of concentration of PVP in weights (Cp). Inserting Eqs. (9) and (10) into (8), then neglecting
Results and discussion a) Molecular orbital characteristics and structure of PVP PVP contains ~ N - - C = O group as the chromophoric moiety embedded in the lactam structure. Hiickel molecular orbital (HMO) calculations using the selected heteroatom parameters [11,12] have been carried out for the chromophoric group. The molecular orbital characteristics of the chromophore are presented below.
340
Colloid and Polymer Science, Vol. 268 9 No. 4 (1990)
- 0.8237 - -
~3
= 0.4129 r
Although Eq. (12) involves perturbations of oxygen and nitrogen centers in PVP structure, it is less likely that both of them form hydrogen bonds simultaneously, because of the positive charge development on nitrogen atom arising from the conjugation with the carbonyl group.
0.8611 r + 0.2964 r
+ 1.2570 ---~ ~2 = 0.7975 ~o + 0.1753~x- 0.5771~N + 2.0667 ---~ ~1 = 0.3544 Cx + 0.5405 Cx + 0.7631 qN.
b) Concentration dependence of the longest wavelength excitation in different media
The g-charge orders [11] in the ground state and the first excited state are q0: 1.5232, qc: 0.6457, qN:
As the concentration of the PVP decreases due to' the greater probability of a much better hydrogen bonding and solvation effects, the overall perturbation on the chromophore system should increase. Whereas at higher PVP concentrations due to the coiling process of the polymer chains, etc., the bulk of the chromophores should be less subject to solvent effects (Fig. 1). In other words, in dilute solutions, due to a better hydrogen bonding and solvation, core potential [17] of the carbonyl oxygen should be greater in comparison to higher concentrations of PVP. Thus, the difference in heteroatom parameter (see Eq. 11), zAh0, necessitates the existence of an inverse proportionality with the concentration. Hence, the variable X in Eq. (12) should be a function of 1/C Indeed, Xmax v s 1/Cpgraphs (Fig 2) plotted for the data recorded in aqueous, as well as nonaqueous solvents (ethyl alcohol, methyl alcohol, chloroform, and acetic acid) in the absence or presence of certain guanidine salts (guanidine sulfate, carbonate and chloride), verify the validity of Eq. (12) in, that the longest wavelength excitation energy of PVP in various solvents should display hypsocromic shift as its concentration lessens and asymptotically approaches to a limiting value. Most
1.8307 and q0: 1.0576, qc: 1.356 5 q~: 1.5855, respectivel)a Also taking into account the lone pairs of oxygen and nitrogen, one can visualize that in the excited state, electron population on oxygen atom should be less as compared to the ground state. Then the longest wavelength excitation should correspond to n ~ U 3 and is associated with a charge transfer from oxygen to carbon. The vacancy left in the nonbonding orbital is equivalent to a positive charge and the electron in I-I* - orbital places excess negative charge between the carbon and oxygen [14]. It can be formulated as
c=o
" - * G 4=0-.
p.
However, the ground state of the chromophore in PVP can be represented by a canonical set (structures I and II) having a greater bias to boundary structure II. All this information implies that the chromophoric group will be hydrogen bonded to solvent protons and this will lower the energy of the ground state. However, hydrogen bonding will be less important in the excited state because of the lessened electron population on oxygen [15, 16].
0
)
,;o
i
i
0
2 o' "~ (nrn)
i
[
i
|
ZOO Z20 Z40 "~ (nm)
(C)
i
i
i
i
i
_
~
t
t
200 220 240
230 250
"~(nm)
"~ (nrn)
Fig. 1 UV-spectra of PVP in water (A), Guanidine chloride (5.10-4 M) (B), CH3OH (C), and Chloroform (D) solutions. PVP concentrations: 0.005, 0.02, 0.1, and 0.5 g/dl.
Tiirker et al., Spectrophotometric behavior of polyvinylpyrrolidone
X
341
mox(nm)I
Table 1. The regression statistics of ~ma• vs 1/Cp plots of various PVP solutions. Water
220 rn 1 m~ R2 ryx FCalc" SY Sbx ts
210
""-
200
......
a,--
o
GSO, ,-i
GC03
190 250
500 1000 1/Cp(dUgr)
Fig. 2 2maxvalues of aqueous PVP solutions vs
-0.1031 216.1780 0.7471 -0.8643 17.7283 4.5371 0.0244 4.2105
Guanidine carbonate -0.1114 218.1040 0.8664 -0.9308 38.9262 3.3024 0.0178 6.2390
Guanidine chloride -0.1067 221.5630 0.8333 -0.9128 29.9954 3.6033 0.0194 5.4768
----GC L
. . . . . . .
. . . . .
-0.1121 216.1380 0.9293 -0.9640 78.9089 2.3338 0.0126 8.8830
Guanidine sulfate
1/Cp.
probably, as the heteroatom parameter Aoh increases due to the increased solvation through hydrogen bonding, its energy lowering effect on the n-orbital should be much more pronounced as compared to lowering on the H orbital, thus causing a much larger interfrontier energy gap [18] that results in a hypsochromic effect as the concentration decreases. Table I tabulates the regression statistics [19] of Am~x vs 1/Cp plots of PVP in aqueous systems for concentrations within the range of 5 10-3 -0.6 g/dl. A linear regression of type Y = m i X + m exists. The comparision of the Fc~ic. and Ftheo"values indicates that the regression is statistically significant. Table I (together with Fig. 2) reveals that each ;tm~x vs 1/Cp plots can be approximated as two intersecting lines, one for the high concentrations and the other for the low concentrations of PVP solutions. Then one may define a hypothetical critical concentration which coincides with the concentration value corresponding to the intersection point of these lines. Below the critical concentration, solvent-solute interactions should prevail and above it solute-solute interactions should start to grow substantially.
R2: coefficient of determination; ryx: correlation coefficient; SY: unexplained standard deviation; SbX: unbiased estimate of the variance of the regression coefficient; t s and F: tests of significance of the regession with 1,6 degrees of freedom. Theoretical F and t s values are 13.74 and 3.143, respectively, at 1% level of significance.
c) Salt effects on ;~ax values The extent of ion pairing may be greatly influenced by the unevenness of charge distribution which so commonly exists in the ions of organic chemistry. These lead to localized regions of high potential gradient in the neighborhood of the ion which can result in strong attractive interactions between cation and anion. Hydrogen bonding is a case in point. Conditions favorable to such bonding between cation and anion lead to increase of several orders of magnitude in the association constant (20). Water and alcohols can serve as hydrogen bond donors to anions and the intensity of the anion-donor interaction increases with decreasing ionic radius [21, 22]. On the other hand, it is a general rule that the strength of a given hydrogen bond is weakened by the formation of an additional hydrogen bond to the same acceptor molecule [20]. Now, consider a PVP solution in water at constant temperature. Depending on the concentration, the strength of hydrogen bonding between the polymer and the solvent should vary as discussed earlier. If certain salts are added without changing the polymer concentration, the anions and cations introduced disturb the evenness of charge distribution and should affect the strength of the hydrogen bonding already existing between the polymer and the solvent. In the case of a series of salts having a common cation the observed perturbational effects can be attributed to the nature of the anions being introduced. Of course, guanidine salts used in the present study may pro-
342
vide chloride, sulfate, bisulfate, bicarbonate, carbonate, and guanidinium ions into the media. Of these ions, bisulfate and bicarbonate are hydrogen bond donor and acceptor in character. Moreover, the charges possessed by them are stabilized through resonance [16, 20]. Hence, their overall anionic effect on the hydrogen bonding should not be very pronounced. Although sulfate and carbonate ions are doubly charged, they have broadly distributed charge and interact much less strongly with hydrogen bond donors: [20]. Wheh chloride, sulfate and carbonate ions are compared, the chloride ion, being smaller than the others and having no resonance stabilized charge, should form much better hydrogen bonding with the solvent molecules at the expense of PVP-solvent interactions through hydrogen bonding. Hence, the addition of guanidine salts, especially guanidine chloride, should be accompanied by a redshift in the longest wavelength excitation of PVP (Figs. 1, 2). One should also keep in mind that guanidine carbonate and sulfate, at least ideally, provide two guanidinium cations per mole of salt as compared to guanidine chloride. Therefore, in the solutions of the first two salts, it is more likely to have an increased number of intimate ion pairs [14] such that the anions and the cations somewhat buffer their actions. Indeed, at comparatively high polymer concentrations (0.6-0.005 g/dl.) guanidine sulfate and carbonate do not have much perturbing effect on Xmaxvalues as compared with the values recorded in aqueous solutions (Fig. 2). It is noteworthy that the addition of guanidine sulfate (5 10-4 M) into the concentrated PVP solutions are accompanied by blue-shifts (Fig. 2), whereas, at the low concentrations red-shifts are observed. The cause of this effect might be the pronounced interaction of PVP with the guanidinium ion through hydrogen bonding. It is a kind of interaction which may arise between hydrogen-bond acceptors and cations [23-25]. It has to be noticed (Fig. 2) that evidently, additon of guanidine salts shifts the critical concentration to lower values. It is also worth mentioning that the behavior of ,t~x values displayed in the presence of various salts (Fig. 2) should actually be the overall effect of the various factors discussed above. d) Solvent effects on ~ x values
The effects of various solvents on the longest wavelength excitation of PVP are shown in Fig. 3
Colloid and Polymer Science, Vol. 268 9 No. 4 (1990)
rnax(nrn)
260_ ........
CH3COOH
Z4o
.,..
.
.
.
.
.
.
.
.
.
.
CHCL3
220
Z00
9
'--- Et.OH
~ . ~
~
........ ,
250
H20
#
500
1000
I/cp(dllgr) Fig. 3. ~
valuesof PVPsolutions vs I/C. in various solvents.
and the following order of bathochromic shifts are obtained. However, there exists no trend of correlation between the observed CH3COOH > CHCI 3 > CH3OH > C2HsOH > H20 order and the solvent parameters [14], Z, ET, Y, and Hildebrand solubility parameter [26]. Moreover, the spectra of PVP in acetic acid and chloroform are almost independent of the concentration of the polymer. In the case of acetic acid, most probably a cage-pair or an extended cage-pair [14] is formed between the lactam group protonated by CH3COOH and acetate ion. The chromophore may exhibit a tautomeric equilibrium (tautomers III and IV) and which tautomer predominates highly depends on the solvent characteristics.
XNC-OH ill /
\C=O Iv /
The molecular orbital energy levels of the lactam, hydrogen bonded lactam, protonated lactam, and lactim forms are shown in Fig. 4. Heteroatom parameters suggested by Kuder (26) were used for the
Tiirker et al., Spectrophotometric behavior of polyvinylpyrrolidone
"'~'-0.t~237 #
~, 0.78/,1#
,~-0.6245 p
-
~-0.6219p
[3
343
One should also note that d u e to the protonation of PVP in acetic acid, the transition m o m e n t [11] for n ~ ~r should decrease. Indeed, absorbances in acetic acid (at the same PVP concentration) are m u c h less than the corresponding values obtained in some other protic solvents.
f'l n
n
~,+1.2570# ~+1.3643l~
,§ ~+2.0660~,
1~ ~+2.1197#
r
Conclusion
ll
1~ ,,,,+2.9076 ~,
1L 1"~+2-0000 # ~§
/c =CyH
Fig. 4 Molecular orbital energies of lactam, hydrogen bonded lactam, protonated lactam, and protonated lactim groups.
molecular orbital calculations of the h y d r o g e n b o n d e d system. In the light of Fig. 4 one m a y postulate that the protonation lowers the l a c t a m / 7 energy m u c h more than its possible lowering effect on the n-level. Even in the case of p r o t o n a t e d lactim form, n-level energy should be closer to the corresponding level of the u n p r o t o n a t e d lactam tautomer. The overall effect of the unequal extent of lowering i n / f and n-levels should cause ;tmax values to appear at longer wavelengths in acetic acid. However, although some w e a k h y d r o g e n b o n d i n g m a y occur b e t w e e n chloroform molecules [16, 20], most probably, there is not a n y appreciable extent of h y d r o g e n bonding between the p o l y m e r and chloroform molecules, so that ;tm~x values in it appear to b e i n d e p e n d e n t of the p o l y m e r concentration. Thus, the spectra of PVP solutions in acetic acid and chloroform have some c o m m o n features. Then mathematically, Eq. (12) yields ( A1/B1 ) = ( A0/B0 ),
In the present s t u d y the effects of certain salts and various solvents on the spectral behavior of PVP solutions have been tried to be explained, mainly on the basis of h y d r o g e n bonding. Dipole-dipole, iondipole, and some other types of interactions which m a y exist in a solution have all been neglected. In addition to that, any second order perturbation effects of solute-solvent interactions have been completely ignored. In spite of all these shortcomings of the approaches presented above, the parallelism f o u n d between the t h e o r y and the observed facts implies that h y d r o g e n b o n d i n g is the main responsible factor for the spectral behavior of PVP solutions.
References
1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
as ( d t / d Z )
x _. O = O.
Gfiner A, Giiven O (1978)Macromol Chem 179, p 2789 MolyneuxP,Ahmed G S (1973)Kolloid-Zu Z Polymere251:310 GargaUoL, Radic D (1983) Polymer 24:91 Bandyopadhyay P, Rodrguez F (1972)Polymer 13:119 Haaf A, Sanner A, Straub F (1985)Polymer Journal 1:143 Jirgenson B (1951)Macromol Chemie 6, p 30 Molyneux P, (1966) Monograph Society of Chemical Industry (London) 24, p 91 Molyneux P, Frank H P (1961) J Am Chem Soc 83, p 3169 Rothschild W G (1972) J Am Chem Soc 94, p 8676 Dewar M J S,Dougherty R C (1975) The PMO Theory of Organic Chemistry, Plenum-Rosetta Ed., New York Heilbronner E, BockH (1976)The HMO-Modeland its Application, Wiley-VerlagEd., London Streitwieser A Jr. (1961)Molecular Orbital Theory For Organic Chemists, John Wiley and Sons, New York VygodskyM (1975)Higher Mathematics,Mir Publishers,Moscow Kosower E M (1968) Physical Organic Chemistry,John Wiley and Sons, New York Wiberg K B (1963) Physical Organic Chemistry, John Wiley and Sons, New York Ferguson L N (1969)The Modern Structural Theory of Organic Chemistry, Prentice-Hall of India, New Delhi Yates K (1978) Hiickel Molecular Orbital Theory, Academic press, London
344 18. Fleming I (1976) Frontier Orbitals and Organic Chemical Reactions, John Wiley and Sons, New York 19. Salvatore D (1982) Statistics and Econometrics, McGraw Hill Co New York 20. Hammett LP (1970)Physical Organic Chemistry, McGraw Hill Kogakusha, Tokyo 21. Lund H (1958) Acta Chem Scand 12:298 22. Ailerhand A, Scleyer P yon R (1963)J Am Chem Soc 85, p 1233 23. Witschonke C R, Kraus C A (1947) J Am Chem Soc 69, p 2472 24. Ralph E K, Gilkerson W R (1964) J Am Chem Soc 86, p 4783 25. Coetzee J F, Podmanabhan G R (1965) J Am Chem Soc 87, p 50O5 26. TagerA (1978) Physical Chemistry Of Polymers, Mir Publishers, Moscow 27. Kuder J E (1972) Tetrahedron 28, p 1973
Colloid and Polymer Science, Vol. 268. No. 4 (1990) Received April 25, 1989; accepted September 9, 1989
Authors' address: Lemi Tfirker Middle East Technical University Department of Chemistry Ankara, Turkey