1540
Anal. Chem. 1981. 53, 1546-1548
/
I
C
1
\ SE
Figure 1. Schematic of cooling system: (A) reservoir vent; (B) reservoir fill cap; (D) deionizer; (F) water filter; (M) water flow meter; (P) water pressure gauge; (T) water temperature gauge.
recirculating loop passes through the shell side of the unit. Filtered tap water flowing through the tube side of the heat exchanger removes the heat generated by the laser. The reservoir is constructed entirely of schedule 40 poly(vinyl chloride) (PVC) material which is readily obtainable from local plumbing supply houses. It consists of a 4 ft length of 6 in. diameter PVC tubing, two 6 X 6 X 1in PVC tees, and two 6 in. PVC end caps which are bonded together with heavy duty industrial grade PVC solvent cement. The reservoir has a capacity of about 35 L and is normally filled with 25 L of deionized water during use. The water lines from the heat exchanger to the reservoir and from the reservoir to the turbine pump, which carry cool water, are made from 1 in. diameter PVC tubing. Hard copper tubing (3/4 in. diameter) is used between the laser and the heat exchanger. The cooling system contains two water filters, two water flow meters, two pressure gauges, and four temperature gauges arranged as shown in Figure 1. Ameteck Plymouth Supreme Model PS-S1water filters (Ameteck, Plymouth Products Div., Sheboygan, WI) with type S1 replaceable cartridges are used to remove particulate matter from the water in the recirculating loop and from the tap water flowing through the heat exchanger. Flow rates are measured with Hershey-American Model 430 water meters. These flow meters are of bronze construction and are obtainable from local plumbing suppliers. The differential water pressure across the 171 laser system is measured with two Ashcroft 0-100 psi dial type pressure gauges. Water temperatures a t the inlets and outlets of the heat exchanger are monitored with Ashcroft 10-149 "C bimetal thermometers.
The deionizer consists of a series arrangement of two Barnstead type HN series 8900 mixed bed ion removal cartridges. When the cooling system is in use, the valve preceding the deionizer is adjusted to obtain the desired water pressure at the laser inlet (approximately 65 psi). This allows a fraction of the recirculating water to continuously pass through the deionizer thereby maintaining its purity.
COOLING SYSTEM PERFORMANCE The above cooling system has been used for the past 3 years to cool a Spectra-Physics 171-18 argon ion laser with excellent results. In operation, a flow rate of 4.2 gal/min is maintained in the recirculating loop, as is recommended by SpectraPhysics. To adequately cool the laser system when it is operating under full output power, a minimum tap water flow rate of 6 gal/min at 23 "C is required. In the summer months, when the tap water reaches a maximum of 30 "C, a flow rate of 8 gallmin is employed. These are modest requirements which are ordinarily satisfied by standard laboratory plumbing facilities. With the laser operating at full output power (-50 A tube current) the cooling system extracts approximately 34 kW of heat from the system with an efficiency (2)greater than 0.5. No problems with the cooling system have been encountered during the 3-year period it has been employed. The laser plasma tube has remained in a near virgin state with no visible evidence of scale buildup or other water contaminants. The Burkes turbine pump has performed flawlessly while providing a very stable water flow rate through the recirculating loop. Although complete cooling systems of similar design are available from commercial vendors, the present system can be assembled in a matter of hours and at a substantial savings. The total cost of the cooling system described here is less than two percent of the cost of the Model 171-18 ion laser and its performance to date suggests that it is well worth the time and effort needed to construct it. LITERATURE CITED Laser Instructlon Manual (C/171 9/ 77); Spectra-Physics, Sunnyvale, CA; p 4-3. Simonson, John R. "An Introduction to Engineering Heat Transfer";
(1) Spectra-Physics Model 171 Ion (2)
McGraw-Hill: New York, 1969; Chapter 2.
RECEIVED for review February 9, 1981. Accepted April 27, 1981. Work supported in part by NSF Grants DMR77-08443 and DMR79-11336.
Silanized Carbon Microneedle Emitters for ChemicaVField Ionization David F. Fraley and Maurice M. Bursey" Wililam Rand Kenan, Jr., Laboratories of Chemistry, The University of North Carolina, Chapel Hiil, North Carolina 27514
I t is now possible to attach surface functional groups covalently to metal emitters used in some laboratories for field desorption (FD) mass spectrometry ( I , 2). These groups are not removed by the combined field, vacuum, and emitter heating used in FD. Further, the influence of the field on the acidity of attached carboxylic acid groups is profound: the carboxylic acid, in the strong electric field used in FD and field ionization (FI),is strong enough to protonate representative olefins, alcohols, and even alkyl halides. Because in FD untreated emitters may produce either M+. or (M f l)",this
protonating ability of the modified acid surface emitter is particularly useful. In the field the heated carboxylic acid emitter cleanly ionizes all molecules with functional groups as (M 1)". Further, although other methods for forming (M f I)+ions exclusively have been found, they involve the use of acids which are strong even in the absence of an electric field coating a surface with polyphosphoric acid (3) or mixing the sample with p-toluenesulfonic acid ( 4 ) runs the risk of acid-catalyzed decomposition of an especially fragile molecule to be analyzed before the ions can be obtained. In the absence
+
0003-2700/81/0353-1546$01.25/00 1981 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 53, NO. 9, AUGUST 1981
Table I. FI/FD Spectrrat Untreated and Acid-Surface Carbon Emitters untreated heating current, compound mA ( M - l]+ M+. (M + .,1 benzoni trile carbon disulfidle cyclohexane anthracene a
100
0 16 (365 K)'
pyridine
100 100 100
0 17 (373 K) 0 17 (373K) 0 20 (390K ) 0 13 (343K)
100 100 100 100 100
20 15
-CO
15 30 30
100 Temperature at emitter tips at the midpoint between the support posts ( 1 2 ) .
of a field, the carboxylic acid group, being weaker, is much less likely to destroy the sample before analysis than these others. The usual description of attachment of a silane to an oxidized metal surface is illustrated in eq 1. We have adopted ' I
1""
I I
+
(ROI~SIR'
OH
-
+
2 ROH
(1)
this model, with the understanding that polymerization of the silane with adsorbed surf'ace water (eq 2) may also occur. Both R
R'
I
( ~ S0I R) ' ~
n20 -__)
-0-SI-oI
SI-0I
? ? -0- S I - O - S I - O -
d
(2)
d'
processes develop a surface containing R' groups which cannot be removed by heat, vacuum, or electric fields. Since no special precautions beyond activation are taken to remove surface water and the emitters are then exposed to laboratory air prior to reaction with the silane, the possibility that eq 2 plays a significant role in fixing surface functional groups is great. This fixing is independent of the strength of any occasional bond to the underlying surface. The strength of this kind of covalent bond to the carbon surface would be weak. The structure of high-temperature microneedles deposited from benzonitrile vapor has been found to be less regular than graphite (5) and probably consists of very small domains of graphite, irregularlyjuxtaposed. Such a material would resemble to some extent the material from which glassy carbon electrodes are made. The surface modification of these electrodes has been examined; the functional groups at the edge of the basal plane of graphite are formed by exposure to oxygen and include the common oxygenated functions: phenolic, quinone, carboxylic acid, and aldehydic groups (6). Silanization of the surface would occur by reaction with phenolic and carboxylic acid groups (eq 3 and 4). Silyll
ethers (eq 3) and esters (eq 4) are well-known derivatives in gas chromatography (7). They are rapidly formed but notoriously unstable to hydrolysis, and the derivatives are used promptly. Hence the durability of a carbon surface modified by treatment with a silane would be questionable if the only mechanism for modification to provide involatile functional
acid surface (M - 1)'
70 70 15 15 30 25 20
1547
--
M'.
(M t l)+
100
20 100
30 100 30 100 50 100 60
100 80
70 100 45 100 25
100 10 15
40 100
groups was attachment of R' chains singly to surface functionalities. As we have seen, however, this will not necessarily be the case. The application of previously developed silanization techniques for emitters to carbon, rather than the development of totally new chemistry, was therefore pursued as a first step.
EXPERIMENTAL SECTION Carbon Emitters. Carbon emitters were grown in a VG Micromass (Altrincham, Cheshire) emitter activating unit following standard procedures (8). The average dendrite length was 20 pm. Mass Spectrometer. A DuPont 21-492B instrument was used for field desorption. Since VG Micromass emitters have shorter support rods than DuPont emitters, they were mounted in the DuPont instrument by fabricating two 10.5-mm extenders from electronic connectors. Because of ion-source design, the distance from emitter to counterelectrode was 6 mm rather than the usual 3 mm in order to prevent arcs between the repeller plates and the emitter. The voltage between the emitter and the counterelectrode was established by settings of +3500 V and -7000 V, respectively. Surface Modification of the Carbon Emitter. Newly prepared emitters were heated to 190 "C for 15-60 min. A 5% (by volume) solution was prepared by the addition of 0.25 mL of 3-(2-aminoethylamino)propyltrimethoxysilane (PCR Research Chemicals, Inc., Gainesville, FL) to 5.0 mL of dry (Na) toluene. The emitter was suspended in this stirred solution for 3 min at ambient temperature. The silanized emitter was briefly washed in dry toluene and then suspended in a saturated solution of succinic anhydride in toluene at about 50 "C for 5 min. It was finally rinsed in toluene to remove excess succinic anhydride, examined under a microscope, and mounted in the mass spectrometer. Two precautions are noteworthy: the syringe used to transfer the silane must be flushed with acetone to avoid problems from the rapidly polymerizing silane, and the silane should be distilled and stored in an airtight vial before use. The surface silane may be removed by dipping in concentrated hydrofluoric acid for 10 s. This procedure does not significantly degrade the carbon emitter performance.
RESULTS AND DISCUSSION The reactions chosen to provide a protonating surface on carbon emitters are analogous to those recommended for cobalt emitters (2). In the context of involatilization through polymer formation, the chemistry may be formalized by reference t o eq 2, in which t h e R group is -CH2CH2CH2NHCH2CH2NH2.Most likely the terminal amino group then reacts with succinic anhydride. Following convention (6, 9, 10) eq 5 is offered as a paradigm of the acylation, ignoring the likely zwitterionic nature of the product or its further reactions. The attachment of the original silane did not significantly alter the sensitivity of the emitters. However, reaction with succinic anhydride is known (2) to alter the field ionization and field desorption spectra produced. Compounds chosen to represent a wide range of proton affinities (PA) and to
1548
Anal. Chem. 1981, 53, 1548-1551 0
II /c
H,C -CH,CH,NH,
\
I
+
G
-
-CH,CH2NHCGCH,CH,COGH
(5)
/
HC ,, C
II 0
compare behavior on modified cobalt emitters were pyridine (PA 218 kcal/mol) (11),benzonitrile (PA 195 kcal/mol) (II), carbon disulfide (PA 168 kcal/mol) (II), and cyclohexane (PA unlisted in ref 11 but probably C168 kcal/mol). All these compounds were analyzed by field ionization. For field desorption, the compound whose behavior on modified cobalt emitters was known was anthracene. Results for the behavior of these compounds a t untreated and acid-surface carbon emitters are summarized in Table I. In Table I the results from the silanized and untreated emitters at no heating current are seen to be usually the same; it is only on heating that a difference appears. The second heating current value listed for each compound is the threshold value for protonation by the acid-surface emitter. At this value the base peak in the spectra of compounds with an easily discerned site of protonation becomes the (M + 1)+ion, and in the spectra of the cycloalkane becomes the (M - 1)+ion, in accordance with the behavior of alkanes in ordinary chemical ionization (eq 6) (13). At the protonation threshold
silanizing solution can produce emitters whose ionization efficiency is greatly reduced. The existence of a threshold for protonation is concordant with the endothermicity of the protonation reaction at zero field. It also has been universally observed with acid-surface metal emitters ( I , 2 )and although the onset of protonation varies with the geometry of each emitter to some extent, the threshold is lower for carbon than for cobalt emitters. This is a manifestation of the lower temperature of a cobalt emitter than a carbon emitter when the same heating current is passed through two otherwise identical emitters (19). In summary, polymerization of silanes by surface-bound water allows the extension of results found for metal emitters to the much more common carbon emitters. While protonation is incomplete because of incomplete surface coverage by the polymer, a peak resulting from protonation becomes the base peak, and this observation clearly establishes the identity of the (M + 1)+peak in a functionalized unknown, a problem which has required much attention (20) in field desorption studies because of the problem of distinguishing whether the FD peak of highest mass is M+. or (M 1)’ in an unknown. Extensions to other silanes, and the development of other methods for modifying functional groups at graphite surface, have begun.
+
ACKNOWLEDGMENT We are grateful to J. R. Hass for access to a VG Micromass emitter activating unit. LITERATURE CITED
current for the acid-surface emitter, or even slightly above it, the untreated emitter is seen to yield the same spectrum as with no heating current. These results are in accord with those obtained with metal emitters. The greatly increased acidity of the carboxylic acid group in the intense electric field at the emitter tip is accommodated by theory, because of the effect of an electric field on ionization of weak electrolytes (14). The pK of an acid in an electric field ought to decrease by a term proportional to the square of the electric field (2). Emitters of all types have tip radii of disperse dimension, which obscures the dependence of signal on field strength, and variation of field was not studied; controversy surrounding field effects has arisen from variations in the degree of importance attached to this point (15-18). Protonation is not complete here, in contrast to the protonation at the acid-surface metal emitter. We believe this is the result of incomplete surface coverage, and indeed, after an arc has struck a working acid-surface emitter of any type, the usual effect is to decrease the intensity of the (M + 1)’ peak. This would be expected if part of the acid polymer is destroyed by the arc, so that some ionization occurs by the classical mechanism at the now untreated surface. We have not found conditions for coating carbon emitters so that only (M 1)+ions are formed. Excessive periods of time in the
+
Youngless, T. L.; Bursey, M. M.; Pedersen, L. G. J. Am. Chem. SOC. 1980, 102, 6881. Youngless, T. L.; Bursey, M. M.; Pedersen, L. G. Int. J . Mass Specfrom. Ion Phys. 1981, 38, 223. Ligon, W. V., Jr. Science 1979, 204, 198. Keough, T.; DeStefano, A. J. Anal. Chem. 1981, 53, 25. Neumann, G. M.;Rogers, D. E.; Derrick, P. J.; Paterson, P. J. K. J . Phys. D 1980, 13, 485. Murray, R. W. Acc. Chem. Res. 1980, 13, 135. Poole, C. F. “Handbook of Derivatives for Chromatography”; Blau, K., King, G. S., Eds.; Heyden: London, 1978; p 152. Beckey, H. D.; Hllt, E.; Schulten, H.-R. J. Phys. E 1973, 6 , 1043. Moses, P. R.; Wler, L. M.; Lennox, J. C.; Finklea, H. 0.; Lenhard, J. R.; Murray, R. W. Anal. Chem. 1978, 50, 576. Wlllman, K. W.; Greer, E.; Murray, R. W. Now. J. Chlm. 1979, 3, 455. Hartman, K. N.; Llas, S.; Ausloos, P.; Rosenstock, H. M.; Schroyer, S. S.; Schmidt, C.; Martinson, D.; Milne, G. W. A. “A Compendium of Gas Phase Baslclty and Proton Affinity Measurements, NBSIR 79-1777; U S . Department of Commerce: Washington, DC, 1979. Fraley, D. F. Ph.D. Dissertation, University of North Carolina at Chapel HIII, Chapel Hill, NC, 1981, Chapter 6. Munson, M. S. B. Anal. Chem. 1971, 43, 28A. Onsager, L. J . Chem. Phys. 1934, 2 , 599. Holland, J. F.; Sokmann, B.; Sweeley, C. C. Biomed. Mass Spectrom. 1976, 3 , 340. Beckey, H. D.; Rollgen, F. W. Org. Mass Spectrom. 1979, 14, 188. Holland, J. F. Org. Mass Spectrom. 1979, 14, 291. Beckey, H. D. Org. Mass Spectrom. 1979, 14, 292. Fraley, D. F.; Woodward, W. S.; Bursey, M. M. Anal. Chem. 1980, 52, 2290. Schulten, H A . ; Beckey, H. D. Org. Mass Spectrom. 1974, 9 , 1154.
RECEIVED for review March 2,1981. Accepted April 27,1981.
Elution of Adsorbed Organics from Graphltized Carbon Black Robert L. Petty Marine Science Institute, University o f California, Santa Barbara, Santa Barbara, California 93 106
In a recent article by Bacaloni et al. (I), the properties of graphitized carbon black (GCB)relating to adsorption of trace organics from water, and desorption with organic solvents, were examined. Several general classes of compounds were
studied. Most of the substances tested, although subject to eventual breakthrough, were shown to be completely retained by the GCB. For desorption of the compounds, several solvent systems were examined. A 1:l mixture of ether and hexane
0003-2700/81/0353-1548$01.25/00 1981 American Chemical Society