Compositional dependence of calcium phosphate layer formation in fluoride BioglassesB Cheol Y. Kim Inha University, Inchon, Korea A. E. Clark* and L. L. Hench University of Florida, Gainesville, Florida 32610
BioglassesB form a double layer composed of apatite a n d a silica-rich layer when placed in a simulated physiological solution as well as in living tissue [A. E. Clark, C.G. Pantano, a n d L. L . Hench, ”Auger spectroscopic analysis of bioglass corrosion films,” 1. Am. Ceram. SOC., 59(1-2), 37-39 (1976).]. In the present work, the mechanisms of t h e calcium phosphate layer and the silica-rich layer formation of fluoride Bioglassesm i n Tris-buffer solution are studied as a function of the SiOz content. Fourier Transform Infrared Reflection Spectroscopy (FTIRS) is used to investigate the mechanism of formation of calcium phosphate and silica-rich layers on the glass surface. Ion concentration in reacted solution and elemental depth profiles are obtained by Induced Coupled Plasma Atomic Emission Spectrometry
IICP) nd Auger Elect on Spectroscopy (AES), respectGely. Si -0 bonds with one n o n b r i d g i n g o x y g e n a n d Si - 0- Si bonds form at the early stage of reaction. Strong p h o s p h o r u s ion u p t a k e occurs when a n amorphous calcium phosphate layer crystallizes. Glasses with high silica content (conventional glass) form t h e silica-rich layer first followed by a calcium phosphate layer on top. However, glasses w i t h low silica content (invert glass) form both layers simultaneously. The rate of apatite formation decreases with increasing SiOz content, especially in the region of conventional glass compositions. Ion release rates decreases as S i 0 2 content increases, with a significant change occurring at the compositional b o u n d a r y between invert a n d conventional glasses. 0 1992 John Wiley & Sons, Inc.
IN TRODUCTION
It has been known that Bioglassm forms a double layer on the surface, composed of calcium phosphate and silica-rich layers, when it is reacted in a simulated physiological solution’ as well as when it is implanted in living tissue.2It is believed that the formation of the calcium phosphate layer during the reaction promotes the chemical bonding between implanted Bioglassm and bone and/or soft tissue. Fluoride in Bioglassm may produce some advantages for dental implant applications. Fluoride released from implant materials may result in the formation of fl~orapatite,3.~ which reduces enamel solubility and also reduces acid production by bacteria which initiate ~ a r i e sThe . ~ addition of fluoride into glasses also decreases the viscosity of the glass melt, making “To whom correspondence should be addressed.
Journal of Biomedical Materials Research, Vol. 26, 1147-1161 (1992) 0 1992 John Wiley & Sons, Inc. CCC 0021-9304/92/091147-15$4.00
KIM, CLARK, A N D HENCH
1148
it easier to coat on substrates. Because of the poor mechanical properties of BioglassesB, coatings of BioglassesB on metal' or alumina7 substrates are being studied in order to enhance the mechanical properties of the resultant biomaterials. Spilman et al.' studied a systematic examination on various fluoride substitution levels in f Iuoride BiogIassesB and conduded that the optimal fluoride substitution level for a 45S5 BioglassB was 40%. Bioglassn is an example of an invert glass," which has a low amount of conventional network forming oxides (i.e., with silica contents of less than 50 mole%). In these basic glasses the formation of a three-dimensional network of SiO, tetrahedra is no longer possible. Ogino et al." studied the compositional dependence of the formation of calcium phosphate films on non-f luorine-containing BioglassesB, and reported that the calcium phosphate layer grew as a function of reaction time and the rate of formation was dependent on the SiOz content of bulk glass. No compositional dependence study has been done for f luoride-containing Bioglasseso. Therefore, the primary objectives of the present study are (a) to study the compositional dependence (Si02 content variation) of the calcium phosphate and silica-rich layer formations for fluoride BioglassesB, especially at the early stage of reaction, and (b) to study the calcium phosphate layer formation in terms of bulk Bioglassm structure.
E X PERIM E N TA L MET HOD
Table I shows the compositions of fluoride BioglassesB which were used in the present study. The compositions are based on 45S5, which has 46.1% SiO,, 2.6% PzOs,24.4% Na20, and 26.9% CaO in mole 5%. CaF, is substituted for 40% of CaO, and SiO, content is varied between 42 mole% and 60 mole%. The content of P20sand Na/Ca ratio are not changed. The glasses were prepared from reagent grades of sodium carbonate, calcium carbonate, phosphorous pentoxide, and silica sand. Raw materials were weighed and mixed in a dry glove box, and rolled for 5 h in a sealed polyethylene bottle. Premixed batches were melted in covered Pt-crucibles in a temperature range of 1150°C to 1400°C depending on compositions for 20 h. Samples were cast in a graphite mold (1 cm in diameter and 2 mm in thickness) and annealed at 450°C for 4 h. All glass discs were prepared by wet TABLE I Fluoride Bioglass@Batch Composition (in mole %) Sample
SiO,
PZOj
Na 2
CaO
CaFz
Y*
42SF 46SF 49SF 52SF 55SF 60SF
42.1 46.1 49.1 52.1 55.1 60.1
2.6 2.6 2.6 2.6 2.6 2.6
26.3 24.4 23.0 21.5 20.1 17.7
17.40 16.14 15.18 14.28 13.32 11.76
11.60 10.76 10.12 9.52 8.88 7.84
1.24 1.65 1.92 2.15 2.36 2.67
*Y = 6
~
20O/P, where P is mole % of SiOz
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
1149
grinding with 320-grit silicon carbide paper, followed by dry grinding with 600-grit paper. The polished samples were cleaned in an ultrasonic acetone bath and then air dried. Samples were suspended in sealed polyethylene bottle which contained 22 mL of trishydroxymethylaminomethane buffer solution, producing a surface area to volume ratio of 0.1 cm-'. The pH of the Tris-buffer solution was 7.20 and samples were reacted at 37°C for various times. The reacted surface films were then examined by FTIRS (Nicolet 10MX) with a diffuse reflection attachment to analyze the formation of surface layers, and scanning Auger Multiprobe (Perkin Elmer PML660) to analyze the surface composition. The vacuum pressure of Auger Electron Spectroscopy was 2.9 X lo-' mm Hg and a monoenergetic electron beam with a beam energy of 3 KV and beam current of 20 A were employed. Thirty seconds of milling by argon ion bombardment was done for all samples to remove any COz contamination from atmosphere. Ion concentrations in the reacted solution were analyzed by induced coupled plasma (IL plasma 200) for Na, Ca, P, and Si ions and by fluoride ion electrode (Orion) for F ion.
RESULT A N D DISCUSSION
All glass compositions of the present study are tabulated in Table I, and the glass structure parameter Y Trap and Stevels' introduced this parameter, which represents the average number of bridging oxygen ions per SiO, tetrahedron. Y can be calculated from the chemical composition of glass (provided that no intermediate ions are present), i.e. y = 6 - - 200
P
where P is the mole % of SiOz. They suggested that if Y is greater than 2, the glass will have a continuous three-dimensional network (conventional glass), and if Y is less than 2, the glass will contain isolated SiO, chains (invert glass). The apatite crystal formation on the glass surface is strongly related to this glass structure change, and will be discussed later in detail. Figures 1 through 4 and Tables I1 through VII show the compositionaldependent FTIRS spectra of fluoride BioglassesB at the early stage of reaction. In Figure 1, in which samples are reacted in Tris-buffer solution for 2 min, the peak at 1095 cm-' is assigned as the Si-0-Si stretching vibration ( ~ 3 )The . next small peak at 1010 cm-' can be assigned as a Si-0 stretching vibration within Si04 units with one nonbridging oxygen." The presence of this peak indicates that Si -0 bonds with one nonbridging oxygen are immediately formed as Na and Ca ions are released from bulk glass. As reported by Kim et a1.,4 this peak disappears as reaction time increases when the Si -0 bond turns into Si -0 -Si bonds as silica gel forms due to a condensation reaction of surface silanol groups. The peaks at 920 cm-' and 482 cm-l, assigned as Si -0 stretching vibrations with two nonbridging oxygens and Si -0 -Si bending mode, respectively, remain unchanged for all compositions. At this very early stage of reaction, there is no significant
KIM, CLARK, AND HENCH
1150
Wave numbers (cm-I)
Figure 1. FTIRS spectra of fluoride BioglassesB after 2 min of reaction.
change in the peak at 598 cm-', which is due to the P-0 bending vibration (v4). However, Auger analysis has shown that a thin calcium phosphate film has already formed when 46SF is reacted for 2 min.4 When the reaction time increases to 1 h, no significant changes for 55SF and 60SF were noted (see Fig. 2). However, for 52SE the Si-0 (with two nonbridging oxygens) vibrational mode at 920 cm-' shifts to a lower wavenumber due to the release of modifying ions. For compositions 42SF 46SE and 49SE the Si -0 vibrational mode at 920 cm-' with two nonbridging oxygens has completely disappeared, and only a vibrational mode at 770 cm-' is shown. The peak at 770 cm-' is assigned as a Si -0 -Si vibration between two neighboring SiO, tetrahedra. This indicates that a silica-rich layer has formed on these samples, when they are reacted for 1 h. The peak at 566 cm-' ( P - 0 bending vibrational mode in a PO, unit; v4) increases sharply in intensity relative to the 2-min spectra (Fig. 1) for 42SF and 46SE which indicates the formation of a calcium phosphate film. However, Si -0 vibrational modes (1094 cm-' for stretching and 470 cm-' for bending) still dominate the overall spectra for the samples reacted for 1 h. When the samples of 46SF and 49SF are reacted for 2 h (Fig. 3), the peak at 566 cm-' (P- 0 bending vibration) splits into two peaks, at 610 and 566 cm-', which are the characteristic peaks of an apatite crystalline phase." This indi-
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
00
1151
,w,
1000
800 600 400 Wave numbers (cm-l)
Figure 2. FTIRS spectra of fluoride BioglassesB after 1 h of reaction.
cates that amorphous calcium phosphate films are crystallized at this reaction time. For 42SE the peak at 566 cm-’ stays as a single peak without splitting, which shows that it is easy for 42SF to form an amorphous calcium phosphate film but not to crystallize. As the calcium phosphate layer grows and crystallizes, COz is incorporated within the calcium phosphate layer producing the broad peak at 900 cm-l,13 and this peak shifts to lower wavenumbers as reaction time increases. At this stage of reaction, 60SF shows the Si-0 bond (with one nonbridging oxygen) formation, 55SF shows Si -0 bond (with two nonbridging oxygens) and Si-0-Si bond formation, and 52SF only shows Si -0 -Si bond formation. This indicates the progressive reaction which depends on the SiOz content in fluoride Bioglassesm. A significant change in overall spectra of Figure 3 is noted between glasses 49SF and 52SE which are the boundary compositions of invert and conventional glasses. The compositional boundary between an invert and conventional glass structure coincides with the crystallization of the calcium phosphate layer. When the fluoride Bioglassesm are reacted for 6 h in Tris-buffer solution, a fluorapatite crystalline phase develops on glasses 42SF up to 52SE The shoulder at 578 cm-’ indicates the fluorapatite formation on the surface of fluoride Bioglasses@.*A sharp C -0 vibrational peak at 866 cm-’ is seen for 42SE 46SF and 49SE which indicates that the C 0 , ” - ion becomes incorporated
KIM, CLARK, AND HENCH
1152
1
200
1000
800
600
&-O-&
400
Wave numbers (crn-I)
Figure 3. FTIRS spectra of fluoride BioglasseP after 2 h of reaction
in the fluorapatite crystal structure. Also the shoulder at 962 cm-', which is assigned as a P - 0 stretching (vl) vibration is clearly shown. The Si-0-Si stretching vibration shows only for the glass 55SF and 60SF at this reaction time. Glass 55SF forms apatite crystals on its surface when it is reacted for 60 h, but no apatite formation was found on the 60SF glass reacted for up to 700 h. In Figure 5, the quantitative analysis of the rate of calcium phosphate layer formation for different composition of fluoride Bioglassesm is shown by comparing the peak intensity of FTIRS spectra. In order to isolate the apatite and silica contributions to the overall spectra, the region of 470-610 cm ' was selected. In this region the P -0 vibration at 566-610 cm-' and Si -0 vibration at 470-506 cm-' are individually resolvable. To enable quantitative comparisons between spectra, the intensities of each peak were measured relative to the minimum intensity between 470 cm-l and 610 cm-I. (Il = P- 0 in the range of 566-610 cm-l and I 2 = Si -0 in the range of 470-506 cm-I). A ratio was calculated of the apatite peak intensity divided by the silica peak intensity (Il/12).Apatite formation is delayed with increasing SiOz content in the glasses. In addition, there is a sharp decrease in the intensity ratio between 49SF and 52SE which is the boundary composition between invert glass and conventional glass. For 55SF and 60SE their apatite silica ratios are too small
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
1200
1000
1153
800 600 400 Wave numbers (cm-I)
Figure 4. FTIRS spectra of fluoride BioglassesB after 6 h of reaction. TABLE I1 Wavenumber (cm-') vs. Time for 42SF Peak Assignment
Wavenumber Range (cm-')
S i - 0 stretch Si - 0 2 N B 0 Si - 0- Si tetra C - 0 stretch P - 0 cryst. P -0 amorp. P - 0 cryst. Si -0 bend
1150-1000 1000-900 1175-71 0 890-800 6 10-600 600-570 530-515 540-415
Reaction Time 2 min
1h
2h
1094 910
1092 948 788
1092
484
728 848
556
550
470
464
6h
866 610 568 434
to compare at the early stage of reaction. As mentioned earlier, however, 55SF forms apatite crystal when it is reacted for 60 h. Figures 6 through 10 show the ion concentrations, of I?, Si, Ca, Na, and F ions, in the reacted Tris-buffer solution after the fluoride Bioglasses@are reacted for various times. Compositions 42SF through 52SF show maximum levels in phosphorus concentration. A maximum point has not developed for 55SF and 60SF during 6 hours of reaction. The drop in P-ion concentration indicates the strong uptake of I' ion from the solution as apatite crystals are
KIM, CLARK, AND HENCH
1154
TABLE I11 Wavenumber (cm-') vs. Time for 46SF Peak Assignment
Wavenumber Range (cm-')
Si -0 stretch Si-0 2NB0 Si- 0- Si tetra C -0 stretch P - 0 cryst. P -0 amorp. P - 0 cryst. Si - 0 bend
1150-1000 1000-900 1175-71 0 890-800 610-600 600-570 530-515 540-415
Reaction Time 2 min
l h
2h
6h
1096 910
1094
1082
786
720 848 608
866 608
566 470
566 444
566 474
476
TABLE IV Wavenumber (cm-') vs. Time for 49SF Peak Assignment
Wavenumber Range (cm-I)
Si - 0 stretch Si -0 2 N B 0 Si- 0- Si tetra C 0 stretch P - 0 cryst. "-0 amorp. P - 0 cryst. Si -0 bend
1150-1000 1000-900 1175-71 0 890-800 610-600 600-570 530-515 540-415
-
Reaction Time 2 min
l h
1098 926
1094 798
2h
6h
754 872 612
864 604
568 472
564 4 74
574 470
484
TABLE V Wavenumber (cm-') vs. Time for 52SF Peak Assignment
Wavenumber Range (cm-')
S i - 0 stretch Si-0 2NB0 Si -0 -Si tetra C - 0 stretch P - 0 cryst. P - 0 amorp. P - 0 cryst. Si -0 bend
1150-1000 1000-900 1175-710 890-800 610-600 600-570 530-515 540-415
Reaction Time
2 min
l h
Zh
1098 934
1096
1094
818
798
576
576
480
472
472
6h
728 842 608 564 474
forming. The peak reaction time varies depending on composition, and the reaction times are 2.5 h for 42SE 1.5 h for 46SE 2.0 h for 49SE and 4.0 h for 52% The phosphorus ion release for 42SF is even lower than that for 46SF and 49SE and its strong phosphorus ion uptake occurs later than that for 46SF and 49SE For all other species, 42SF always shows the highest ion release compared with other compositions. It is believed that the P-ion release rate for 42SF is also higher than that for other composition of glasses, but there may be a strong uptake of P ion occurring simultaneously. The phosphorus ion
1155
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
TABLE VI Wavenumber (cm-') vs. Time for 55SF Peak Assignment Si-0
Reaction Time
Wavenumber Range (cm-')
stretch
2 min
l h
2h
6h
1098 940
1090 920
1094
1092
806
798
570
562
472
472
1150-1000 1000-900 1175-710 890-800 610-600 600-570 530-515 540-415
Si- 0 2 N B 0
Si- 0- Si tetra C- 0 stretch P - 0 cryst. P- 0 amorp. P - 0 cryst. Si - 0 bend
492
480
TABLE VII Wavenumber (cm-') vs. Time for 60SF Peak Assignment
Wavenumber Range (cm-')
stretch Si-0 2 N B 0 Si-0-Si tetra C - 0 stretch P - 0 cryst. P -0 amorp. P - 0 cryst. Si -0 bend
1150-1000 1000-900 1175-710 890-800 610-600 600-570 530-5 15 540-415
Si-0
Reaction Time 2 min
l h
2h
6h
1094
1098 950 756
1100 964 762
1096 952 758
488
488
482
476
3.0
0 la LT Y
a
w 2.0
a
Q:
0
=!
$ 3 a k 1.0 t-
2 Q: 0.0
I
0
I
I
I
I
I
I
1
2
3
4
5
6
REACTION TIME (Hours) Figure 5. Apatite forming rate a t early stage of reaction for fluoride Bioglasses@.
KIM, CLARK, AND H E N C H
1156 52SF
I
1
I
1
2
I I I 3 4 5 REACTION TIME (Hours)
1 6
Figure 6. Phosphorus ion concentration in reacted solution for fluoride Rioglasses@.
1
2
3
4
5
6
REACTION TIME (Hours)
Figure 7. Silicon ion concentration in reacted solution vs. reacted time fa fluoride BioglassesB.
uptake time (peak points in P-ion concentration) corresponds to the reaction time that the amorphous calcium phosphate layers convert into crystalline phases. This fact indicates that strong P-ion uptake occurs when the amorphus calcium phosphate layer crystalIizes. When the sample 55SF is reacted
1157
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
--I
70
60
50-
z
1
2
3 4 5 REACTION TIME (Hours)
6
Figure 8. Calcium ion concentration in reacted solution vs. reaction time for fluoride Bioglasses@.
V"
70
-
60
h
n, 0
50
m
z
8
z 0
40
+
2
30
z W
0
5 0
20
10
1
2
3 4 5 REACTION TIME (Hours)
6
Figure 9. Sodium ion concentration in reacted solution vs. reaction time for fluoride Bioglasses@.
KIM, CLARK, AND HENCH
1158
t
30
1
2
3 4 5 REACTION TIME (Hours)
6
Figure 10. Fluoride ion concentration in reacted solution vs. reaction time for fluoride Rioglasses@.
for 6 h, 2.8 ppm of phosphorus ions are released from glass. However, no apatite crystals are found on the surface (see Fig. 4). Compared wjth glass 46SE which releases 2.1 ppm of phosphorus ions into the solution and forms apatite crystal when it is reacted for 1 h, 55SF releases enough phosphorus ions into solution to form the crystalline calcium phosphate phase. Therefore, it can be said that the phosphorus ion concentration in the reacted solution is not the only factor affecting apatite crystallization on the Bioglassm surface. For the other ions, as expected, ion release rate (slope of each curve) and the amount of ions released at each reaction time decrease with the increase in SiOzcontent. For each element, the invert glass compositions (42,46,49SF) seem to be grouped together and the conventional glass compositions (52,55, 60SF) are associated together in a separate group. At the early stages of reaction, the Na-ion release rate is almost the same as that of Ca ion for all glasses. Sodium-ion concentration in the reacted solution becomes higher than Ca-ion concentration for longer reacted samples. This is probably due to the uptake of Ca ion from the reacted solution to form apatite crystals. Table VIII shows the ion concentrations in solutions for reaction times up to 700 h for samples of 46Sb 55SE and 60% According to the results from FTIRS spectra, all the 46SF and 55SF samples develop fluorapatite crystals on the surface, whereas none of the 60SF samples show any crystal formation. However, the increased intensity at 570 cm-l in the FTIRS spectra of a 200-h reacted 60SF sample indicates the formation of an amorphous calcium phosphate layer. The P-ion concentration is very low in the reacted soIution when
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
1159
TABLE VIII Ion Concentration in Solution for 46Se 55SF and 60SF Ion Concentration (PPM) Glass 45SF 55SF 60SF
Reaction time (h)
Na
Ca
Si
P
20 240 200 350 700 200 350 700
82.5 253 64.1 68.2 67.4 34.2 33.9 35.4
59.2 104 42.4 44.1 40.0 43.3 43.0 22.0
35.6 54.5 52.8 53.9 58.0 54.1 54.4 57.0
0.2 0.1 0.3 0.0 0.0 6.4 6.1 2.1
46SF and 55SF samples are reacted for longer times. This suggests that once apatite crystals form the release of P ions into solution ceases and instead immediate uptake of this ion by apatite crystal occurs as it is released from the bulk glass. A high concentration of P ion for 60SF occurs up to 350 h of reaction, and then the P- and Ca-ion concentration decreases, however the FTIRS spectra do not show any sign of crystal formation. The 46SF composition releases ions steadily (but not linearly) as reaction time increases. However, the ion concentrations for 55SF and 60SF are almost constant with reaction time. This suggests that the rigidity of bulk glass structure plays a more important role in hindering ion migration than do apatite crystals. Figure 11 shows the changes in surface composition as a function of Si02 content after 2 h of reaction. The concentration of each element is calculated from the peak-to-peak height of Auger electronic spectra. At 2 h, a calcium phosphate layer has formed for the glasses 42SE 46SE and 49% This calcium phosphate layer includes the fluoride ion, which indicates the formation of f luorapatite. Silica may also be incorporated into the apatite crystal. However, Ogino et aL9reported that almost no silicon is detected in apatite crystals formed on nonfluoride BioglassesB. Note the sharp change in the elemental concentration in the composition range of 49SF and 52% This is the boundary composition between invert and conventional glasses. For glass compositions 52SE 55SE and 60SE no Ca, I?, and F peaks are found, while a sharp increase in the Si peak occurs. This indicates that a silica-rich layer has formed on the surface, but no calcium phosphate film is present. Therefore, it can be said that fluoride BioglassesB which contain less than 50% SiO, (42% to 49% SiO,), which are invert glasses, form calcium phosphate films and Si02-rich layers simultaneously. However, for the glasses with higher Si02 content (52% to 60% SiOz), which have a conventional glass structure, the silica-rich layer forms first and the calcium phosphate layer develops later on the top of silica-rich layer. Apatite crystals are formed after 6 h of reaction for 52SF and after 60 h of reaction for 55SF according to the FTIRS results. The major factor affecting apatite crystal formation on fluoride BioglassesB is not the presence of a silica-rich layer but the nature of the silica-rich layer which depends on the composition of BioglassesB. Further studies of the ultrastructure of the silica-rich layer are needed.
KIM, CLARK, A N D H E N C H
1160 9
i
a
I
E7 z 3
> a: 4
III
6
t m
a: 5 5
+ I
(-7 u! 1 4 Y
4
w
a
53 ?4 2
1
42
46
49
52
55
60
3 0 2 CONTENT (mole %)
Figure 11. Changes in surface composition with variation of SiOz concentration after 2 h reaction.
CONCLUSION
Based on the results in this study, the following conclusions may be made. (1) For all compositions, at the early stages of reaction, FTIRS analysis shows the formation of Si -0 bonds with one nonbridging oxygen and Si -0 -Si bridging bonds to produce a silica-rich layer. (2) Strong P-ion uptake from solution occurs when the amorphous calcium phosphate layer converts into a crystalline phase. (3) There is a significant change in ion concentration and rate of apatite formation which occurs between the composition 49SF and 52SE which is the boundary between invert and conventional glasses. (4) The calcium phosphate and silica-rich layers form simultaneously for the invert glasses, but the silica-rich layer forms first and then a calcium phosphate layer forms on top of it later for the conventional glasses. The authors gratefully acknowledge the support of the Korean Science and Engineering Foundation and the University of Florida.
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
1161
References 1. A. E. Clark, C.G. Pantano, and L. L. Hench, 'Auger spectroscopic analysis of BioglassB corrosion films," 1.Am. Ceram. Soc., 59(1-2), 37-39 (19 76). 2. L. L. Hench and A . E. Clark, "Adhesion to bone," in Biocompatibzlity of Orthapaedic Implants, Vol. 11, D. F. Williams (ed.), CRC Press, Boca Raton, 1982, pp. 129-170. 3. T. Fujiu, M. Ogino, M. Kariy, and T. Ichimura, "New explanation for bonding behavior of fluorine containing Bioglasseso," J. Nun-Cryst. Solids, 56, 417-422 (1983). 4. C.Y. Kim, A. E. Clark, and L. L. Hench, "Early stage of calcium phosphate layer formation in Bioglasseso," J. Nun-Cryst. Solids, 112, 195-202 (1989). 5. K. Okuda and G. Ferstall, "The effect of fluoride on the acid production of Streptococcus mutans and other oral streptococci," Swedish Dent. J., 6, 29-36 (1982). 6. W. R. Lacefield and L. L. Hench, "The bonding of BioglassO to cobalt chromium surgical implant alloy," Biomaterials, 7(3), 104-108 (1986). 7. D.C. Greenspan and L. L. Hench, "Chemical and mechanical behavior of BioglassB-coated alumina," J. Biomed. Res. Symp., 7, 503-509 (1976). 8.
9. 10.
11. 12.
13.
D. Spilman, J. Wilson, and L. Hench, "In-vivo and in-vitro investigations into BioglassesQ which contain fluoride," Trans. Second World Congress on Biomaterials, 8, 287 (1984). H. J. L. Trap and J. M. Stevels, "Conventional and invert glasses containing titania. Part 1," Pkys. Chem. Glasses, 1(4), 107-118 (1960). M. Ogino, F. Ohuchi, and L. L. Hench, "Compositional dependence of the formation of calcium phosphate films on Bioglassm," I. Biomed. Muter. Res., 14, 55-64 (1980). W. L. Konijnendijk, "The structure of borosilicate glasses," Phillips Res. Rep. Suppl., 1 (1975). B.O. Fowler, "Infrared studies of apatite I. Vibrational assignments for calcium, strontium and barium hydroxyapatite utilizing isotopic substitution," Inorg. Chem., 13(1), 194-207 (1974). R.E. LeGeros, G. Bone, and R. LeGeros, "Type of HzO in human enamel and in precipitated apatites," Calcif. Tiss. Res., 26, 111-118 (1978).
Received September 4, 1990 Accepted February 17,1992
BioglassesB form a double layer composed of apatite a n d a silica-rich layer when placed in a simulated physiological solution as well as in living tissue [A. E. Clark, C.G. Pantano, a n d L. L . Hench, ”Auger spectroscopic analysis of bioglass corrosion films,” 1. Am. Ceram. SOC., 59(1-2), 37-39 (1976).]. In the present work, the mechanisms of t h e calcium phosphate layer and the silica-rich layer formation of fluoride Bioglassesm i n Tris-buffer solution are studied as a function of the SiOz content. Fourier Transform Infrared Reflection Spectroscopy (FTIRS) is used to investigate the mechanism of formation of calcium phosphate and silica-rich layers on the glass surface. Ion concentration in reacted solution and elemental depth profiles are obtained by Induced Coupled Plasma Atomic Emission Spectrometry
IICP) nd Auger Elect on Spectroscopy (AES), respectGely. Si -0 bonds with one n o n b r i d g i n g o x y g e n a n d Si - 0- Si bonds form at the early stage of reaction. Strong p h o s p h o r u s ion u p t a k e occurs when a n amorphous calcium phosphate layer crystallizes. Glasses with high silica content (conventional glass) form t h e silica-rich layer first followed by a calcium phosphate layer on top. However, glasses w i t h low silica content (invert glass) form both layers simultaneously. The rate of apatite formation decreases with increasing SiOz content, especially in the region of conventional glass compositions. Ion release rates decreases as S i 0 2 content increases, with a significant change occurring at the compositional b o u n d a r y between invert a n d conventional glasses. 0 1992 John Wiley & Sons, Inc.
IN TRODUCTION
It has been known that Bioglassm forms a double layer on the surface, composed of calcium phosphate and silica-rich layers, when it is reacted in a simulated physiological solution’ as well as when it is implanted in living tissue.2It is believed that the formation of the calcium phosphate layer during the reaction promotes the chemical bonding between implanted Bioglassm and bone and/or soft tissue. Fluoride in Bioglassm may produce some advantages for dental implant applications. Fluoride released from implant materials may result in the formation of fl~orapatite,3.~ which reduces enamel solubility and also reduces acid production by bacteria which initiate ~ a r i e sThe . ~ addition of fluoride into glasses also decreases the viscosity of the glass melt, making “To whom correspondence should be addressed.
Journal of Biomedical Materials Research, Vol. 26, 1147-1161 (1992) 0 1992 John Wiley & Sons, Inc. CCC 0021-9304/92/091147-15$4.00
KIM, CLARK, A N D HENCH
1148
it easier to coat on substrates. Because of the poor mechanical properties of BioglassesB, coatings of BioglassesB on metal' or alumina7 substrates are being studied in order to enhance the mechanical properties of the resultant biomaterials. Spilman et al.' studied a systematic examination on various fluoride substitution levels in f Iuoride BiogIassesB and conduded that the optimal fluoride substitution level for a 45S5 BioglassB was 40%. Bioglassn is an example of an invert glass," which has a low amount of conventional network forming oxides (i.e., with silica contents of less than 50 mole%). In these basic glasses the formation of a three-dimensional network of SiO, tetrahedra is no longer possible. Ogino et al." studied the compositional dependence of the formation of calcium phosphate films on non-f luorine-containing BioglassesB, and reported that the calcium phosphate layer grew as a function of reaction time and the rate of formation was dependent on the SiOz content of bulk glass. No compositional dependence study has been done for f luoride-containing Bioglasseso. Therefore, the primary objectives of the present study are (a) to study the compositional dependence (Si02 content variation) of the calcium phosphate and silica-rich layer formations for fluoride BioglassesB, especially at the early stage of reaction, and (b) to study the calcium phosphate layer formation in terms of bulk Bioglassm structure.
E X PERIM E N TA L MET HOD
Table I shows the compositions of fluoride BioglassesB which were used in the present study. The compositions are based on 45S5, which has 46.1% SiO,, 2.6% PzOs,24.4% Na20, and 26.9% CaO in mole 5%. CaF, is substituted for 40% of CaO, and SiO, content is varied between 42 mole% and 60 mole%. The content of P20sand Na/Ca ratio are not changed. The glasses were prepared from reagent grades of sodium carbonate, calcium carbonate, phosphorous pentoxide, and silica sand. Raw materials were weighed and mixed in a dry glove box, and rolled for 5 h in a sealed polyethylene bottle. Premixed batches were melted in covered Pt-crucibles in a temperature range of 1150°C to 1400°C depending on compositions for 20 h. Samples were cast in a graphite mold (1 cm in diameter and 2 mm in thickness) and annealed at 450°C for 4 h. All glass discs were prepared by wet TABLE I Fluoride Bioglass@Batch Composition (in mole %) Sample
SiO,
PZOj
Na 2
CaO
CaFz
Y*
42SF 46SF 49SF 52SF 55SF 60SF
42.1 46.1 49.1 52.1 55.1 60.1
2.6 2.6 2.6 2.6 2.6 2.6
26.3 24.4 23.0 21.5 20.1 17.7
17.40 16.14 15.18 14.28 13.32 11.76
11.60 10.76 10.12 9.52 8.88 7.84
1.24 1.65 1.92 2.15 2.36 2.67
*Y = 6
~
20O/P, where P is mole % of SiOz
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
1149
grinding with 320-grit silicon carbide paper, followed by dry grinding with 600-grit paper. The polished samples were cleaned in an ultrasonic acetone bath and then air dried. Samples were suspended in sealed polyethylene bottle which contained 22 mL of trishydroxymethylaminomethane buffer solution, producing a surface area to volume ratio of 0.1 cm-'. The pH of the Tris-buffer solution was 7.20 and samples were reacted at 37°C for various times. The reacted surface films were then examined by FTIRS (Nicolet 10MX) with a diffuse reflection attachment to analyze the formation of surface layers, and scanning Auger Multiprobe (Perkin Elmer PML660) to analyze the surface composition. The vacuum pressure of Auger Electron Spectroscopy was 2.9 X lo-' mm Hg and a monoenergetic electron beam with a beam energy of 3 KV and beam current of 20 A were employed. Thirty seconds of milling by argon ion bombardment was done for all samples to remove any COz contamination from atmosphere. Ion concentrations in the reacted solution were analyzed by induced coupled plasma (IL plasma 200) for Na, Ca, P, and Si ions and by fluoride ion electrode (Orion) for F ion.
RESULT A N D DISCUSSION
All glass compositions of the present study are tabulated in Table I, and the glass structure parameter Y Trap and Stevels' introduced this parameter, which represents the average number of bridging oxygen ions per SiO, tetrahedron. Y can be calculated from the chemical composition of glass (provided that no intermediate ions are present), i.e. y = 6 - - 200
P
where P is the mole % of SiOz. They suggested that if Y is greater than 2, the glass will have a continuous three-dimensional network (conventional glass), and if Y is less than 2, the glass will contain isolated SiO, chains (invert glass). The apatite crystal formation on the glass surface is strongly related to this glass structure change, and will be discussed later in detail. Figures 1 through 4 and Tables I1 through VII show the compositionaldependent FTIRS spectra of fluoride BioglassesB at the early stage of reaction. In Figure 1, in which samples are reacted in Tris-buffer solution for 2 min, the peak at 1095 cm-' is assigned as the Si-0-Si stretching vibration ( ~ 3 )The . next small peak at 1010 cm-' can be assigned as a Si-0 stretching vibration within Si04 units with one nonbridging oxygen." The presence of this peak indicates that Si -0 bonds with one nonbridging oxygen are immediately formed as Na and Ca ions are released from bulk glass. As reported by Kim et a1.,4 this peak disappears as reaction time increases when the Si -0 bond turns into Si -0 -Si bonds as silica gel forms due to a condensation reaction of surface silanol groups. The peaks at 920 cm-' and 482 cm-l, assigned as Si -0 stretching vibrations with two nonbridging oxygens and Si -0 -Si bending mode, respectively, remain unchanged for all compositions. At this very early stage of reaction, there is no significant
KIM, CLARK, AND HENCH
1150
Wave numbers (cm-I)
Figure 1. FTIRS spectra of fluoride BioglassesB after 2 min of reaction.
change in the peak at 598 cm-', which is due to the P-0 bending vibration (v4). However, Auger analysis has shown that a thin calcium phosphate film has already formed when 46SF is reacted for 2 min.4 When the reaction time increases to 1 h, no significant changes for 55SF and 60SF were noted (see Fig. 2). However, for 52SE the Si-0 (with two nonbridging oxygens) vibrational mode at 920 cm-' shifts to a lower wavenumber due to the release of modifying ions. For compositions 42SF 46SE and 49SE the Si -0 vibrational mode at 920 cm-' with two nonbridging oxygens has completely disappeared, and only a vibrational mode at 770 cm-' is shown. The peak at 770 cm-' is assigned as a Si -0 -Si vibration between two neighboring SiO, tetrahedra. This indicates that a silica-rich layer has formed on these samples, when they are reacted for 1 h. The peak at 566 cm-' ( P - 0 bending vibrational mode in a PO, unit; v4) increases sharply in intensity relative to the 2-min spectra (Fig. 1) for 42SF and 46SE which indicates the formation of a calcium phosphate film. However, Si -0 vibrational modes (1094 cm-' for stretching and 470 cm-' for bending) still dominate the overall spectra for the samples reacted for 1 h. When the samples of 46SF and 49SF are reacted for 2 h (Fig. 3), the peak at 566 cm-' (P- 0 bending vibration) splits into two peaks, at 610 and 566 cm-', which are the characteristic peaks of an apatite crystalline phase." This indi-
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
00
1151
,w,
1000
800 600 400 Wave numbers (cm-l)
Figure 2. FTIRS spectra of fluoride BioglassesB after 1 h of reaction.
cates that amorphous calcium phosphate films are crystallized at this reaction time. For 42SE the peak at 566 cm-’ stays as a single peak without splitting, which shows that it is easy for 42SF to form an amorphous calcium phosphate film but not to crystallize. As the calcium phosphate layer grows and crystallizes, COz is incorporated within the calcium phosphate layer producing the broad peak at 900 cm-l,13 and this peak shifts to lower wavenumbers as reaction time increases. At this stage of reaction, 60SF shows the Si-0 bond (with one nonbridging oxygen) formation, 55SF shows Si -0 bond (with two nonbridging oxygens) and Si-0-Si bond formation, and 52SF only shows Si -0 -Si bond formation. This indicates the progressive reaction which depends on the SiOz content in fluoride Bioglassesm. A significant change in overall spectra of Figure 3 is noted between glasses 49SF and 52SE which are the boundary compositions of invert and conventional glasses. The compositional boundary between an invert and conventional glass structure coincides with the crystallization of the calcium phosphate layer. When the fluoride Bioglassesm are reacted for 6 h in Tris-buffer solution, a fluorapatite crystalline phase develops on glasses 42SF up to 52SE The shoulder at 578 cm-’ indicates the fluorapatite formation on the surface of fluoride Bioglasses@.*A sharp C -0 vibrational peak at 866 cm-’ is seen for 42SE 46SF and 49SE which indicates that the C 0 , ” - ion becomes incorporated
KIM, CLARK, AND HENCH
1152
1
200
1000
800
600
&-O-&
400
Wave numbers (crn-I)
Figure 3. FTIRS spectra of fluoride BioglasseP after 2 h of reaction
in the fluorapatite crystal structure. Also the shoulder at 962 cm-', which is assigned as a P - 0 stretching (vl) vibration is clearly shown. The Si-0-Si stretching vibration shows only for the glass 55SF and 60SF at this reaction time. Glass 55SF forms apatite crystals on its surface when it is reacted for 60 h, but no apatite formation was found on the 60SF glass reacted for up to 700 h. In Figure 5, the quantitative analysis of the rate of calcium phosphate layer formation for different composition of fluoride Bioglassesm is shown by comparing the peak intensity of FTIRS spectra. In order to isolate the apatite and silica contributions to the overall spectra, the region of 470-610 cm ' was selected. In this region the P -0 vibration at 566-610 cm-' and Si -0 vibration at 470-506 cm-' are individually resolvable. To enable quantitative comparisons between spectra, the intensities of each peak were measured relative to the minimum intensity between 470 cm-l and 610 cm-I. (Il = P- 0 in the range of 566-610 cm-l and I 2 = Si -0 in the range of 470-506 cm-I). A ratio was calculated of the apatite peak intensity divided by the silica peak intensity (Il/12).Apatite formation is delayed with increasing SiOz content in the glasses. In addition, there is a sharp decrease in the intensity ratio between 49SF and 52SE which is the boundary composition between invert glass and conventional glass. For 55SF and 60SE their apatite silica ratios are too small
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
1200
1000
1153
800 600 400 Wave numbers (cm-I)
Figure 4. FTIRS spectra of fluoride BioglassesB after 6 h of reaction. TABLE I1 Wavenumber (cm-') vs. Time for 42SF Peak Assignment
Wavenumber Range (cm-')
S i - 0 stretch Si - 0 2 N B 0 Si - 0- Si tetra C - 0 stretch P - 0 cryst. P -0 amorp. P - 0 cryst. Si -0 bend
1150-1000 1000-900 1175-71 0 890-800 6 10-600 600-570 530-515 540-415
Reaction Time 2 min
1h
2h
1094 910
1092 948 788
1092
484
728 848
556
550
470
464
6h
866 610 568 434
to compare at the early stage of reaction. As mentioned earlier, however, 55SF forms apatite crystal when it is reacted for 60 h. Figures 6 through 10 show the ion concentrations, of I?, Si, Ca, Na, and F ions, in the reacted Tris-buffer solution after the fluoride Bioglasses@are reacted for various times. Compositions 42SF through 52SF show maximum levels in phosphorus concentration. A maximum point has not developed for 55SF and 60SF during 6 hours of reaction. The drop in P-ion concentration indicates the strong uptake of I' ion from the solution as apatite crystals are
KIM, CLARK, AND HENCH
1154
TABLE I11 Wavenumber (cm-') vs. Time for 46SF Peak Assignment
Wavenumber Range (cm-')
Si -0 stretch Si-0 2NB0 Si- 0- Si tetra C -0 stretch P - 0 cryst. P -0 amorp. P - 0 cryst. Si - 0 bend
1150-1000 1000-900 1175-71 0 890-800 610-600 600-570 530-515 540-415
Reaction Time 2 min
l h
2h
6h
1096 910
1094
1082
786
720 848 608
866 608
566 470
566 444
566 474
476
TABLE IV Wavenumber (cm-') vs. Time for 49SF Peak Assignment
Wavenumber Range (cm-I)
Si - 0 stretch Si -0 2 N B 0 Si- 0- Si tetra C 0 stretch P - 0 cryst. "-0 amorp. P - 0 cryst. Si -0 bend
1150-1000 1000-900 1175-71 0 890-800 610-600 600-570 530-515 540-415
-
Reaction Time 2 min
l h
1098 926
1094 798
2h
6h
754 872 612
864 604
568 472
564 4 74
574 470
484
TABLE V Wavenumber (cm-') vs. Time for 52SF Peak Assignment
Wavenumber Range (cm-')
S i - 0 stretch Si-0 2NB0 Si -0 -Si tetra C - 0 stretch P - 0 cryst. P - 0 amorp. P - 0 cryst. Si -0 bend
1150-1000 1000-900 1175-710 890-800 610-600 600-570 530-515 540-415
Reaction Time
2 min
l h
Zh
1098 934
1096
1094
818
798
576
576
480
472
472
6h
728 842 608 564 474
forming. The peak reaction time varies depending on composition, and the reaction times are 2.5 h for 42SE 1.5 h for 46SE 2.0 h for 49SE and 4.0 h for 52% The phosphorus ion release for 42SF is even lower than that for 46SF and 49SE and its strong phosphorus ion uptake occurs later than that for 46SF and 49SE For all other species, 42SF always shows the highest ion release compared with other compositions. It is believed that the P-ion release rate for 42SF is also higher than that for other composition of glasses, but there may be a strong uptake of P ion occurring simultaneously. The phosphorus ion
1155
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
TABLE VI Wavenumber (cm-') vs. Time for 55SF Peak Assignment Si-0
Reaction Time
Wavenumber Range (cm-')
stretch
2 min
l h
2h
6h
1098 940
1090 920
1094
1092
806
798
570
562
472
472
1150-1000 1000-900 1175-710 890-800 610-600 600-570 530-515 540-415
Si- 0 2 N B 0
Si- 0- Si tetra C- 0 stretch P - 0 cryst. P- 0 amorp. P - 0 cryst. Si - 0 bend
492
480
TABLE VII Wavenumber (cm-') vs. Time for 60SF Peak Assignment
Wavenumber Range (cm-')
stretch Si-0 2 N B 0 Si-0-Si tetra C - 0 stretch P - 0 cryst. P -0 amorp. P - 0 cryst. Si -0 bend
1150-1000 1000-900 1175-710 890-800 610-600 600-570 530-5 15 540-415
Si-0
Reaction Time 2 min
l h
2h
6h
1094
1098 950 756
1100 964 762
1096 952 758
488
488
482
476
3.0
0 la LT Y
a
w 2.0
a
Q:
0
=!
$ 3 a k 1.0 t-
2 Q: 0.0
I
0
I
I
I
I
I
I
1
2
3
4
5
6
REACTION TIME (Hours) Figure 5. Apatite forming rate a t early stage of reaction for fluoride Bioglasses@.
KIM, CLARK, AND H E N C H
1156 52SF
I
1
I
1
2
I I I 3 4 5 REACTION TIME (Hours)
1 6
Figure 6. Phosphorus ion concentration in reacted solution for fluoride Rioglasses@.
1
2
3
4
5
6
REACTION TIME (Hours)
Figure 7. Silicon ion concentration in reacted solution vs. reacted time fa fluoride BioglassesB.
uptake time (peak points in P-ion concentration) corresponds to the reaction time that the amorphous calcium phosphate layers convert into crystalline phases. This fact indicates that strong P-ion uptake occurs when the amorphus calcium phosphate layer crystalIizes. When the sample 55SF is reacted
1157
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
--I
70
60
50-
z
1
2
3 4 5 REACTION TIME (Hours)
6
Figure 8. Calcium ion concentration in reacted solution vs. reaction time for fluoride Bioglasses@.
V"
70
-
60
h
n, 0
50
m
z
8
z 0
40
+
2
30
z W
0
5 0
20
10
1
2
3 4 5 REACTION TIME (Hours)
6
Figure 9. Sodium ion concentration in reacted solution vs. reaction time for fluoride Bioglasses@.
KIM, CLARK, AND HENCH
1158
t
30
1
2
3 4 5 REACTION TIME (Hours)
6
Figure 10. Fluoride ion concentration in reacted solution vs. reaction time for fluoride Rioglasses@.
for 6 h, 2.8 ppm of phosphorus ions are released from glass. However, no apatite crystals are found on the surface (see Fig. 4). Compared wjth glass 46SE which releases 2.1 ppm of phosphorus ions into the solution and forms apatite crystal when it is reacted for 1 h, 55SF releases enough phosphorus ions into solution to form the crystalline calcium phosphate phase. Therefore, it can be said that the phosphorus ion concentration in the reacted solution is not the only factor affecting apatite crystallization on the Bioglassm surface. For the other ions, as expected, ion release rate (slope of each curve) and the amount of ions released at each reaction time decrease with the increase in SiOzcontent. For each element, the invert glass compositions (42,46,49SF) seem to be grouped together and the conventional glass compositions (52,55, 60SF) are associated together in a separate group. At the early stages of reaction, the Na-ion release rate is almost the same as that of Ca ion for all glasses. Sodium-ion concentration in the reacted solution becomes higher than Ca-ion concentration for longer reacted samples. This is probably due to the uptake of Ca ion from the reacted solution to form apatite crystals. Table VIII shows the ion concentrations in solutions for reaction times up to 700 h for samples of 46Sb 55SE and 60% According to the results from FTIRS spectra, all the 46SF and 55SF samples develop fluorapatite crystals on the surface, whereas none of the 60SF samples show any crystal formation. However, the increased intensity at 570 cm-l in the FTIRS spectra of a 200-h reacted 60SF sample indicates the formation of an amorphous calcium phosphate layer. The P-ion concentration is very low in the reacted soIution when
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
1159
TABLE VIII Ion Concentration in Solution for 46Se 55SF and 60SF Ion Concentration (PPM) Glass 45SF 55SF 60SF
Reaction time (h)
Na
Ca
Si
P
20 240 200 350 700 200 350 700
82.5 253 64.1 68.2 67.4 34.2 33.9 35.4
59.2 104 42.4 44.1 40.0 43.3 43.0 22.0
35.6 54.5 52.8 53.9 58.0 54.1 54.4 57.0
0.2 0.1 0.3 0.0 0.0 6.4 6.1 2.1
46SF and 55SF samples are reacted for longer times. This suggests that once apatite crystals form the release of P ions into solution ceases and instead immediate uptake of this ion by apatite crystal occurs as it is released from the bulk glass. A high concentration of P ion for 60SF occurs up to 350 h of reaction, and then the P- and Ca-ion concentration decreases, however the FTIRS spectra do not show any sign of crystal formation. The 46SF composition releases ions steadily (but not linearly) as reaction time increases. However, the ion concentrations for 55SF and 60SF are almost constant with reaction time. This suggests that the rigidity of bulk glass structure plays a more important role in hindering ion migration than do apatite crystals. Figure 11 shows the changes in surface composition as a function of Si02 content after 2 h of reaction. The concentration of each element is calculated from the peak-to-peak height of Auger electronic spectra. At 2 h, a calcium phosphate layer has formed for the glasses 42SE 46SE and 49% This calcium phosphate layer includes the fluoride ion, which indicates the formation of f luorapatite. Silica may also be incorporated into the apatite crystal. However, Ogino et aL9reported that almost no silicon is detected in apatite crystals formed on nonfluoride BioglassesB. Note the sharp change in the elemental concentration in the composition range of 49SF and 52% This is the boundary composition between invert and conventional glasses. For glass compositions 52SE 55SE and 60SE no Ca, I?, and F peaks are found, while a sharp increase in the Si peak occurs. This indicates that a silica-rich layer has formed on the surface, but no calcium phosphate film is present. Therefore, it can be said that fluoride BioglassesB which contain less than 50% SiO, (42% to 49% SiO,), which are invert glasses, form calcium phosphate films and Si02-rich layers simultaneously. However, for the glasses with higher Si02 content (52% to 60% SiOz), which have a conventional glass structure, the silica-rich layer forms first and the calcium phosphate layer develops later on the top of silica-rich layer. Apatite crystals are formed after 6 h of reaction for 52SF and after 60 h of reaction for 55SF according to the FTIRS results. The major factor affecting apatite crystal formation on fluoride BioglassesB is not the presence of a silica-rich layer but the nature of the silica-rich layer which depends on the composition of BioglassesB. Further studies of the ultrastructure of the silica-rich layer are needed.
KIM, CLARK, A N D H E N C H
1160 9
i
a
I
E7 z 3
> a: 4
III
6
t m
a: 5 5
+ I
(-7 u! 1 4 Y
4
w
a
53 ?4 2
1
42
46
49
52
55
60
3 0 2 CONTENT (mole %)
Figure 11. Changes in surface composition with variation of SiOz concentration after 2 h reaction.
CONCLUSION
Based on the results in this study, the following conclusions may be made. (1) For all compositions, at the early stages of reaction, FTIRS analysis shows the formation of Si -0 bonds with one nonbridging oxygen and Si -0 -Si bridging bonds to produce a silica-rich layer. (2) Strong P-ion uptake from solution occurs when the amorphous calcium phosphate layer converts into a crystalline phase. (3) There is a significant change in ion concentration and rate of apatite formation which occurs between the composition 49SF and 52SE which is the boundary between invert and conventional glasses. (4) The calcium phosphate and silica-rich layers form simultaneously for the invert glasses, but the silica-rich layer forms first and then a calcium phosphate layer forms on top of it later for the conventional glasses. The authors gratefully acknowledge the support of the Korean Science and Engineering Foundation and the University of Florida.
COMPOSITIONAL DEPENDENCE OF Ca-P LAYER
1161
References 1. A. E. Clark, C.G. Pantano, and L. L. Hench, 'Auger spectroscopic analysis of BioglassB corrosion films," 1.Am. Ceram. Soc., 59(1-2), 37-39 (19 76). 2. L. L. Hench and A . E. Clark, "Adhesion to bone," in Biocompatibzlity of Orthapaedic Implants, Vol. 11, D. F. Williams (ed.), CRC Press, Boca Raton, 1982, pp. 129-170. 3. T. Fujiu, M. Ogino, M. Kariy, and T. Ichimura, "New explanation for bonding behavior of fluorine containing Bioglasseso," J. Nun-Cryst. Solids, 56, 417-422 (1983). 4. C.Y. Kim, A. E. Clark, and L. L. Hench, "Early stage of calcium phosphate layer formation in Bioglasseso," J. Nun-Cryst. Solids, 112, 195-202 (1989). 5. K. Okuda and G. Ferstall, "The effect of fluoride on the acid production of Streptococcus mutans and other oral streptococci," Swedish Dent. J., 6, 29-36 (1982). 6. W. R. Lacefield and L. L. Hench, "The bonding of BioglassO to cobalt chromium surgical implant alloy," Biomaterials, 7(3), 104-108 (1986). 7. D.C. Greenspan and L. L. Hench, "Chemical and mechanical behavior of BioglassB-coated alumina," J. Biomed. Res. Symp., 7, 503-509 (1976). 8.
9. 10.
11. 12.
13.
D. Spilman, J. Wilson, and L. Hench, "In-vivo and in-vitro investigations into BioglassesQ which contain fluoride," Trans. Second World Congress on Biomaterials, 8, 287 (1984). H. J. L. Trap and J. M. Stevels, "Conventional and invert glasses containing titania. Part 1," Pkys. Chem. Glasses, 1(4), 107-118 (1960). M. Ogino, F. Ohuchi, and L. L. Hench, "Compositional dependence of the formation of calcium phosphate films on Bioglassm," I. Biomed. Muter. Res., 14, 55-64 (1980). W. L. Konijnendijk, "The structure of borosilicate glasses," Phillips Res. Rep. Suppl., 1 (1975). B.O. Fowler, "Infrared studies of apatite I. Vibrational assignments for calcium, strontium and barium hydroxyapatite utilizing isotopic substitution," Inorg. Chem., 13(1), 194-207 (1974). R.E. LeGeros, G. Bone, and R. LeGeros, "Type of HzO in human enamel and in precipitated apatites," Calcif. Tiss. Res., 26, 111-118 (1978).
Received September 4, 1990 Accepted February 17,1992
